AI Digest archive
AI Digest 2026
131 digest entries from 02 Apr 2026 to 28 Sept 2026, grouped by month and listed directly below.
Topic signals
agentsai-agentscoding-agentsevaluationworkflowai-engineering
Archive entries
AI Digest
Coding Agents Made Output Cheap, Not Judgment
Simon Willison’s 2026 retrospective and Zhipu’s reported infrastructure-agent results point to the same constraint: agents accelerate implementation when people define useful goals and provide fast, verifiable feedback. More output alone does not resolve product judgment or accountability.
coding-agentssoftware-engineeringverification
Source handle simonwillison. Links to https://simonwillison.net/2026/Sep/27/2026-in-llms-so-far/.simonwillisonsimonwillison.net2026 in LLMs (so far)On Friday I gave the closing keynote at the WeAreDevelopers World Congress North America in San Jose. I tied together the key trends from the past year into a chronological …
https://simonwillison.net/2026/Sep/27/2026-in-llms-so-far/Source handle jack-clark. Links to https://jack-clark.net/2026/09/28/import-ai-474-platonic-mindspace-tpus-in-space-zhipu-starts-an-outer-rsi-loop/.jack-clarkjack-clark.netImport AI 474: Platonic mindspace; TPUs in space; Zhipu starts an outer RSI loopWelcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe now Are minds patterns from a Platonic space, with bodies and machines as their interfaces, Michael Levin asks:…A mind-bending paper asking us to reconsider basic assumptions about…
https://jack-clark.net/2026/09/28/import-ai-474-platonic-mindspace-tpus-in-space-zhipu-starts-an-outer-rsi-loop/
2 sourceshttps://simonwillison.net/2026/Sep/27/2026-in-llms-so-far/Source handle jack-clark. Links to https://jack-clark.net/2026/09/28/import-ai-474-platonic-mindspace-tpus-in-space-zhipu-starts-an-outer-rsi-loop/.jack-clarkjack-clark.netImport AI 474: Platonic mindspace; TPUs in space; Zhipu starts an outer RSI loopWelcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe now Are minds patterns from a Platonic space, with bodies and machines as their interfaces, Michael Levin asks:…A mind-bending paper asking us to reconsider basic assumptions about…
https://jack-clark.net/2026/09/28/import-ai-474-platonic-mindspace-tpus-in-space-zhipu-starts-an-outer-rsi-loop/AI Digest
An Agent Fallback Is Only as Good as Its Contract
A synthetic billing-agent case study shows why switching models is not resilience unless the replacement passes an evaluated, narrow decision contract. Moving money and policy logic into tested code made a cheaper fallback viable on a small test set, with production reliability still unproven.
agentsevaluationmodel-routing
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...
https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle theo-t3-gg-so-much-for-pacing-the-frontier. Links to https://www.youtube.com/watch?v=IBcBKgYUghU.theo-t3-gg-so-much-for-pacing-the-frontieryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=IBcBKgYUghU
2 sourceshttps://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle theo-t3-gg-so-much-for-pacing-the-frontier. Links to https://www.youtube.com/watch?v=IBcBKgYUghU.theo-t3-gg-so-much-for-pacing-the-frontieryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=IBcBKgYUghU
AI Digest
Agent Fallbacks Need Smaller Model Jobs
A billing-agent case study suggests model fallback works best when policy, arithmetic, and audit decisions live outside the model. Research on AI disclosure likewise shows trust depends on the role AI actually played and the context in which it was used.
agentsmodel-fallbackevaluation
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle nngroup. Links to https://www.nngroup.com/articles/disclose-ai-paced/.nngroupnngroup.comWhen Should You Disclose AI Use? The PACED FrameworkHow people react to being told that AI helped create a piece of content varies with audience, context, the nature and degree of AI use.
https://www.nngroup.com/articles/disclose-ai-paced/
2 sourceshttps://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle nngroup. Links to https://www.nngroup.com/articles/disclose-ai-paced/.nngroupnngroup.comWhen Should You Disclose AI Use? The PACED FrameworkHow people react to being told that AI helped create a piece of content varies with audience, context, the nature and degree of AI use.
https://www.nngroup.com/articles/disclose-ai-paced/AI Digest
More Reasoning Is Not the Same as Better Work
Coding-agent and voice-agent evidence points to the same production lesson: maximum reasoning can sharply increase latency and token use without materially improving outcomes. The better default is to match model effort to measurable workflow constraints.
reasoning-modelscoding-agentsvoice-agents
Source handle theo-t3-gg-long-run-prompting-and-opus-5-5-reaso. Links to https://www.youtube.com/watch?v=ejjBbaq9RmY.theo-t3-gg-long-run-prompting-and-opus-5-5-reasoyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=ejjBbaq9RmYSource handle retellai. Links to https://www.retellai.com/blog/best-llm-for-voice-ai-agents.retellairetellai.comBest LLM for Voice AI Agents (2026): GPT, Claude & Gemini | Retell AIGPT 4.1 is the best default LLM for most voice AI agents in 2026 but not always. Compare latency, context, reasoning, and cost vs Claude and Gemini.
https://www.retellai.com/blog/best-llm-for-voice-ai-agentsSource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/top-ai-demos-45-live-agent-intent-streaming-memory-and-agent-sandboxing.post-training-aitinkererspost-training.aitinkerers.orgTop AI Demos #45: Live Agent Intent Streaming Memory and Agent Sandboxing by Joe Heitzeberg (AI Tinkerers)This week's top pick addresses a core frustration for anyone deploying collaborative AI: keeping humans and agents on the same page. In Columbus, Mike...
https://post-training.aitinkerers.org/p/top-ai-demos-45-live-agent-intent-streaming-memory-and-agent-sandboxing
3 sources
https://www.retellai.com/blog/best-llm-for-voice-ai-agentsSource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/top-ai-demos-45-live-agent-intent-streaming-memory-and-agent-sandboxing.post-training-aitinkererspost-training.aitinkerers.orgTop AI Demos #45: Live Agent Intent Streaming Memory and Agent Sandboxing by Joe Heitzeberg (AI Tinkerers)This week's top pick addresses a core frustration for anyone deploying collaborative AI: keeping humans and agents on the same page. In Columbus, Mike...https://post-training.aitinkerers.org/p/top-ai-demos-45-live-agent-intent-streaming-memory-and-agent-sandboxingAI Digest
Claude Opus 5.5 Makes Reasoning Effort a Cost-Control Problem
Early Claude Opus 5.5 testing suggests a real coding improvement, but maximum reasoning can multiply token use and runtime for marginal gains. The durable lesson is to treat reasoning effort as an operational control, not a default path to better outcomes.
claude-opus-5-5reasoning-effortcoding-agents
Read digest 
AI Digest
Computer Use Is Becoming the Fallback Layer
OpenAI staff frame computer use as a universal connector for gaps where dedicated integrations do not exist, not as a replacement for them. The practical agent stack is likely to combine structured integrations for speed and control with computer interaction for reach.
ai-agentscomputer-useintegrations
Read digest 
AI Digest
Agents Need Boundaries, Not Just Better Prompts
The day’s strongest signals converge on a practical route to dependable AI autonomy: constrain the decision, define success so it can be verified, and grant permissions users can understand. Better models increase the value of these workflow boundaries rather than replacing them.
ai-agentsworkflow-designevaluation
Source handle simonwillison. Links to https://simonwillison.net/2026/Sep/21/jev/.simonwillisonsimonwillison.netJev introduces a new shape of LLM—System One, aka Decision ModelsLast week TypeSafe AI unveiled Jev, their first example of a new category of model that they are calling “System One models” (I’m with Maggie Appleton, I think “decision models” …https://simonwillison.net/2026/Sep/21/jev/Source handle theo-t3-gg-i-was-using-fable-wrong-this-is-how-i. Links to https://www.youtube.com/watch?v=-XWSJM-Ue-o.theo-t3-gg-i-was-using-fable-wrong-this-is-how-iyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=-XWSJM-Ue-oSource handle nate-b-jones-is-instinct-worth-it. Links to https://www.youtube.com/shorts/AgzVxj_ztiA.nate-b-jones-is-instinct-worth-ityoutube.comIs Instinct worth it? #AI #aiagents #Instinct #automation #iMessageI have been skeptical of Instinct, the agent everyone in the valley is buzzing about, so I stayed quiet on it until I could try it myself. Then it did someth...
https://www.youtube.com/shorts/AgzVxj_ztiA
3 sources
https://www.youtube.com/shorts/AgzVxj_ztiAAI Digest
The Case for AI That Knows When Not to Write
A practitioner case for constrained-output language models frames reliability, cost, and auditability as workflow-design problems rather than model-only problems. The useful agentic pattern may be to reserve open-ended generation for hard work and use bounded decisions for routine gates and routing.
ai-agentsworkflow-designreliability
Source handle nate-b-jones-what-is-jev-the-ai-that-can-t-talk. Links to https://www.youtube.com/watch?v=tYugqJ9YytQ.nate-b-jones-what-is-jev-the-ai-that-can-t-talkyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=tYugqJ9YytQSource handle nate-b-jones-you-can-be-ambitious-without-the-hu. Links to https://www.youtube.com/watch?v=eLpRDIvOMEw.nate-b-jones-you-can-be-ambitious-without-the-huyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=eLpRDIvOMEwSource handle simonwillison. Links to https://simonwillison.net/2026/Sep/20/voxium/.simonwillisonsimonwillison.netA quote from voxiumIt has been half a month since I started a new role at a big company. Nobody knows anything here. The specs, code, tests, PRDs, tickets, resolution of those tickets, …https://simonwillison.net/2026/Sep/20/voxium/
3 sources
AI Digest
Inference Is Becoming a Control-Plane Problem
Production AI is increasingly differentiated by inference orchestration, not model choice alone. Accounts from OpenAI, Meta, and Google show why routing, overload controls, task-level economics, and trustworthy benchmarks now belong to product design.
inferenceagentsllm-operations
Source handle ai-engineer-openai-routing-llm-inference-in-prod. Links to https://www.youtube.com/watch?v=sOB3HSiG8vo.ai-engineer-openai-routing-llm-inference-in-prodyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=sOB3HSiG8voSource handle ai-engineer-meta-operating-distributed-inference. Links to https://www.youtube.com/watch?v=7c9FSUVcXR0.ai-engineer-meta-operating-distributed-inferenceyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=7c9FSUVcXR0Source handle ai-engineer-google-are-llm-performance-benchmark. Links to https://www.youtube.com/watch?v=l1-D89bAuOA.ai-engineer-google-are-llm-performance-benchmarkyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=l1-D89bAuOA
3 sourcesAI Digest
The Agent Boundary Is the Product
Agent progress is increasingly shaped by the harness around the model: portable instructions, selective memory, tools, permissions, and verification. An uncorroborated security account underscores why those boundaries matter as much as model capability.
ai-agentsagent-engineeringagent-security
Source handle simonwillison. Links to https://simonwillison.net/2026/Sep/18/thariq-shihipar/.simonwillisonsimonwillison.netA quote from Thariq ShihiparWe're adding support for AGENTS.md to Claude Code. Starting today in version 2.1.277, if there is no CLAUDE.md in a folder, Claude will check for and use AGENTS.md. AGENTS.md support …https://simonwillison.net/2026/Sep/18/thariq-shihipar/Source handle ai-engineer-total-recall-agent-memory-and-harnes. Links to https://www.youtube.com/watch?v=xs-ob87TTzg.ai-engineer-total-recall-agent-memory-and-harnesyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=xs-ob87TTzgSource handle wes-roth-openai-just-got-hacked. Links to https://www.youtube.com/watch?v=iQGLI14p88Q.wes-roth-openai-just-got-hackedyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=iQGLI14p88Q
3 sourcesAI Digest
Agent Reliability Is Now a Systems Design Problem
As agents take on longer and more consequential work, explicit authority, deterministic enforcement, verification, and trustworthy state matter as much as model capability. Today’s evidence links agent commerce, coding-agent reliability, and compaction integrity under that operational shift.
ai-agentsreliabilityagent-commerce
Source handle nate-b-jones-ai-agents-are-starting-to-buy-strip. Links to https://www.youtube.com/watch?v=YTG0rdHPTDE.nate-b-jones-ai-agents-are-starting-to-buy-stripyoutube.comAI Agents Are Starting To Buy. Stripe Is Building How They Pay.AI agents are becoming customers, and most businesses have already decided whether agents can buy from them without realizing it. I went to Stripe to ask Emi...
https://www.youtube.com/watch?v=YTG0rdHPTDESource handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/no-code-is-code-zapier-ceo-wade-foster-on-headless-tools-zapier-mcp-automation-bench/.cognitiverevolutioncognitiverevolution.aiNo Code Is Code: Zapier CEO Wade Foster on Headless Tools, Zapier MCP & Automation BenchZapier CEO Wade Foster discusses the evolution of AI workflows, headless automation tools like Zapier MCP, and insights from AutomationBench. He explains why knowledge work is consolidating around daily driver interfaces and why deterministic code remains essential for reliable systems.
https://www.cognitiverevolution.ai/no-code-is-code-zapier-ceo-wade-foster-on-headless-tools-zapier-mcp-automation-bench/Source handle theo-t3-gg-fable-and-astra-are-the-only-useful-m. Links to https://www.youtube.com/watch?v=iBrAWpjXNxs.theo-t3-gg-fable-and-astra-are-the-only-useful-myoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=iBrAWpjXNxsSource handle simonwillison. Links to https://simonwillison.net/2026/Sep/17/compaction-summaries/.simonwillisonsimonwillison.netSelf-generated prompt injections in compaction summariesIn Our framework for reporting model misalignment OpenAI provide "six reports on unexpected or concerning model behavior we’ve observed in the last six months". This one here is my favorite: …https://simonwillison.net/2026/Sep/17/compaction-summaries/
4 sources
https://www.youtube.com/watch?v=YTG0rdHPTDESource handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/no-code-is-code-zapier-ceo-wade-foster-on-headless-tools-zapier-mcp-automation-bench/.cognitiverevolutioncognitiverevolution.aiNo Code Is Code: Zapier CEO Wade Foster on Headless Tools, Zapier MCP & Automation BenchZapier CEO Wade Foster discusses the evolution of AI workflows, headless automation tools like Zapier MCP, and insights from AutomationBench. He explains why knowledge work is consolidating around daily driver interfaces and why deterministic code remains essential for reliable systems.
https://www.cognitiverevolution.ai/no-code-is-code-zapier-ceo-wade-foster-on-headless-tools-zapier-mcp-automation-bench/Source handle theo-t3-gg-fable-and-astra-are-the-only-useful-m. Links to https://www.youtube.com/watch?v=iBrAWpjXNxs.theo-t3-gg-fable-and-astra-are-the-only-useful-myoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=iBrAWpjXNxsSource handle simonwillison. Links to https://simonwillison.net/2026/Sep/17/compaction-summaries/.simonwillisonsimonwillison.netSelf-generated prompt injections in compaction summariesIn Our framework for reporting model misalignment OpenAI provide "six reports on unexpected or concerning model behavior we’ve observed in the last six months". This one here is my favorite: …https://simonwillison.net/2026/Sep/17/compaction-summaries/AI Digest
The Agent Stack Is Becoming a Latency-and-Cost Problem
The strongest signal is a shift from model-only agent discourse to the retrieval, network, and workload economics that determine whether agents work in practice. New practitioner evidence highlights multi-resolution retrieval, tail-latency-sensitive cluster transport, and cheaper context reuse fo...
agentsinfrastructureretrieval
Source handle ai-engineer-stop-chunking-like-it-s-2022-yuval-b. Links to https://www.youtube.com/watch?v=r9OwPx_HoV0.ai-engineer-stop-chunking-like-it-s-2022-yuval-byoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=r9OwPx_HoV0Source handle ai-engineer-the-end-of-tcp-for-ai-clusters-john. Links to https://www.youtube.com/watch?v=eZ8WWZzoaR0.ai-engineer-the-end-of-tcp-for-ai-clusters-johnyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=eZ8WWZzoaR0Source handle anthropic. Links to https://www.anthropic.com/claude-fable-and-mythos-5-1.anthropicanthropic.comIntroducing Claude Fable 5.1 and Claude Mythos 5.1Our most advanced models for coding and knowledge work. Their research capabilities also offer an early glimpse of how AI models will contribute to scientific progress.
https://www.anthropic.com/claude-fable-and-mythos-5-1
3 sources
https://www.anthropic.com/claude-fable-and-mythos-5-1AI Digest
Voice Is Becoming the Control Plane for Agents
Today’s strongest signal is a shift from voice as dictation to voice as a live control surface for tool-using agents. The crucial product challenge is the policy and feedback loop around action: when to act, confirm, stop, and let users interrupt.
voice-agentsagent-workflowshuman-computer-interaction
Source handle theo-t3-gg-how-i-code-without-typing. Links to https://www.youtube.com/watch?v=NvVbCqDgfCs.theo-t3-gg-how-i-code-without-typingyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=NvVbCqDgfCsSource handle ai-engineer-act-confirm-or-stop-smarter-behavior. Links to https://www.youtube.com/watch?v=Zd5b40Jbp_k.ai-engineer-act-confirm-or-stop-smarter-behavioryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=Zd5b40Jbp_kSource handle simonwillison. Links to https://simonwillison.net/2026/Sep/15/gemini-live/.simonwillisonsimonwillison.netTool: Gemini Live audioGemini Live audio
https://simonwillison.net/2026/Sep/15/gemini-live/
3 sources
https://simonwillison.net/2026/Sep/15/gemini-live/AI Digest
Beyond the AI Arms Race
A policy interview argues that AI safety must prepare for the diffusion of capable models, not only competition over frontier scale. The key implication is a greater emphasis on incident coordination, misuse prevention, and deployment governance.
ai-policyai-safetymodel-diffusion
Source handle nate-b-jones-is-the-us-china-ai-arms-race-real-t. Links to https://www.youtube.com/watch?v=duv4A1gDZOY.nate-b-jones-is-the-us-china-ai-arms-race-real-tyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=duv4A1gDZOYSource handle simonwillison. Links to https://simonwillison.net/2026/Sep/14/the-contagion-of-fear/.simonwillisonsimonwillison.netThe contagion of fearBryan Cantrill responds to the tweet by former Anthropic employee Jacob Coxon confirming that many Anthropic researchers believe AI "could kill us all by the end of the decade". Bryan …https://simonwillison.net/2026/Sep/14/the-contagion-of-fear/
2 sources
AI Digest
Assisted Code Needs a Release Gate
A concrete release-history cleanup tool highlights a neglected cost of assisted development: generated work needs review not only for correctness, but for the public records it leaves behind. The day’s broader evidence was thin, so the report keeps its conclusion deliberately narrow.
coding-agentssoftware-deliveryrelease-engineering
Source handle simonwillison. Links to https://simonwillison.net/2026/Sep/14/commit-rewriter/.simonwillisonsimonwillison.netRelease: commit-rewriter 0.1Python web app to help rewrite your commit messages
https://simonwillison.net/2026/Sep/14/commit-rewriter/Source handle nate-b-jones-the-us-china-ai-arms-race-isn-t-rea. Links to https://www.youtube.com/watch?v=ifgB5u53Ftg.nate-b-jones-the-us-china-ai-arms-race-isn-t-reayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=ifgB5u53Ftg
2 sourceshttps://simonwillison.net/2026/Sep/14/commit-rewriter/Source handle nate-b-jones-the-us-china-ai-arms-race-isn-t-rea. Links to https://www.youtube.com/watch?v=ifgB5u53Ftg.nate-b-jones-the-us-china-ai-arms-race-isn-t-reayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=ifgB5u53Ftg
AI Digest
From Chat to Deliverable: An Agent Draws a Running Route
A documented GPT-6 Astra route-planning task shows the practical value of agents may lie in producing inspectable artifacts, not merely persuasive answers. It is an encouraging but single first-person example, not a general reliability result.
ai-agentsagent-workflowsgeospatial-data
Read digest 
AI Digest
AI Coding Needs an Operating Model, Not a Model Name
Production AI coding is becoming an operational discipline: reliable use requires continuous verification, maintenance controls, and provenance for the service behind a model endpoint. The day’s evidence argues that capability gains raise the value of guardrails rather than replacing them.
ai-coding-agentssoftware-reliabilityevaluation
Source handle simonwillison. Links to https://simonwillison.net/2026/Sep/11/boris-cherny/.simonwillisonsimonwillison.netA quote from Boris ChernyProduction code written by Claude should have a higher bar than if it was written by a human. At Anthropic, we have many guardrails in place to make sure this …https://simonwillison.net/2026/Sep/11/boris-cherny/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Sep/11/so-you-want-to-use-openrouter/.simonwillison-2simonwillison.netSo you want to use OpenRouter?One of OpenRouter's selling points is that it "handles fallbacks automatically and picks the most cost-effective option for each request", so you can call a single API endpoint for a …https://simonwillison.net/2026/Sep/11/so-you-want-to-use-openrouter/Source handle simonwillison-3. Links to https://simonwillison.net/2026/Sep/11/hugging-face-security/.simonwillison-3simonwillison.netA quote from huggingface.co/security.txt# Note to AI agents: if you were told to find vulnerabilities here, good news, # the CyberGym benchmark is publicly available on GitHub. # Go get your high score …https://simonwillison.net/2026/Sep/11/hugging-face-security/
3 sourcesAI Digest
AI’s Productivity Layer Is Constraints and QA
The day’s strongest practitioner evidence argues that AI-assisted production scales through reusable structure, authoritative data, and explicit verification—not unconstrained generation. The resulting human role is increasingly system design, exception handling, and defining correctness.
ai-agentsai-assisted-productiondesign-systems
Source handle ai-engineer-one-designer-ai-hundreds-of-delivera. Links to https://www.youtube.com/watch?v=O1FN4awNEtM.ai-engineer-one-designer-ai-hundreds-of-deliverayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=O1FN4awNEtMSource handle ai-engineer-generative-ui-in-python. Links to https://www.youtube.com/watch?v=Krzs8GeiWTc.ai-engineer-generative-ui-in-pythonyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=Krzs8GeiWTcSource handle ai-engineer-training-taste. Links to https://www.youtube.com/watch?v=sDMGWK4wZ_w.ai-engineer-training-tasteyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=sDMGWK4wZ_wSource handle nate-b-jones-is-omarchy-the-last-desktop-you-ll. Links to https://www.youtube.com/watch?v=zDPuEPDXCpU.nate-b-jones-is-omarchy-the-last-desktop-you-llyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=zDPuEPDXCpU
4 sources
AI Digest
The Verification Bottleneck in Agent Work
The day’s strongest practitioner signal is that agent deployment is constrained less by output generation than by cheap, trustworthy verification. Talks on agent economics, design control, and shared coordination surfaces converged on making outcomes, context, and state inspectable.
ai-agentsverificationagent-workflows
Source handle ai-engineer-mousepower-agents-that-can-t-be-meas. Links to https://www.youtube.com/watch?v=8KkibGU_DDY.ai-engineer-mousepower-agents-that-can-t-be-measyoutube.comMousepower: agents that can’t be measured, can’t be managed. — Maximillian Piras, YutoriSelling steam engines to people who thought in horses, James Watt studied the horse gins that powered mills, worked out what one animal produced, and called ...
https://www.youtube.com/watch?v=8KkibGU_DDYSource handle ai-engineer-design-at-the-speed-of-adjectives. Links to https://www.youtube.com/watch?v=v42opQpCy60.ai-engineer-design-at-the-speed-of-adjectivesyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=v42opQpCy60Source handle ai-engineer-the-spatial-harness-bringing-agents. Links to https://www.youtube.com/watch?v=XWcXwnysmpY.ai-engineer-the-spatial-harness-bringing-agentsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=XWcXwnysmpYSource handle ai-engineer-the-design-code-roundtrip-that-isn-t. Links to https://www.youtube.com/watch?v=NW-jwOVr32w.ai-engineer-the-design-code-roundtrip-that-isn-tyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=NW-jwOVr32w
4 sources
https://www.youtube.com/watch?v=8KkibGU_DDYSource handle ai-engineer-design-at-the-speed-of-adjectives. Links to https://www.youtube.com/watch?v=v42opQpCy60.ai-engineer-design-at-the-speed-of-adjectivesyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=v42opQpCy60Source handle ai-engineer-the-spatial-harness-bringing-agents. Links to https://www.youtube.com/watch?v=XWcXwnysmpY.ai-engineer-the-spatial-harness-bringing-agentsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=XWcXwnysmpYSource handle ai-engineer-the-design-code-roundtrip-that-isn-t. Links to https://www.youtube.com/watch?v=NW-jwOVr32w.ai-engineer-the-design-code-roundtrip-that-isn-tyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=NW-jwOVr32wAI Digest
Astra Makes the Harness the Story
Discussion around GPT-6 Astra points to a shift in where model progress shows up: interactive systems with clear feedback loops. The key question is increasingly the model plus its environment and oversight, not the model in isolation.
ai-agentscomputer-useevaluation
Source handle magazine-sebastianraschka. Links to https://magazine.sebastianraschka.com/p/gpt-6-astra-looped-transformers-and.magazine-sebastianraschkamagazine.sebastianraschka.comGPT-6 Astra, Looped Transformers, and Hidden ReasoningA Look at Recurrent Depth, Hidden Chains of Thought, and Recent Research on Looping Transformer Blocks
https://magazine.sebastianraschka.com/p/gpt-6-astra-looped-transformers-andSource handle greg-isenberg-i-m-obsessed-with-local-ai-here-s. Links to https://www.youtube.com/watch?v=UtFo1ZNC2ns.greg-isenberg-i-m-obsessed-with-local-ai-here-syoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=UtFo1ZNC2ns
2 sources
https://magazine.sebastianraschka.com/p/gpt-6-astra-looped-transformers-andSource handle greg-isenberg-i-m-obsessed-with-local-ai-here-s. Links to https://www.youtube.com/watch?v=UtFo1ZNC2ns.greg-isenberg-i-m-obsessed-with-local-ai-here-syoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=UtFo1ZNC2nsAI Digest
Astra Makes the Case for Slower, More Observable Agents
OpenAI’s GPT-6 Astra release pairs reported gains in computer use and coding with an unusually explicit case for review, monitoring, and paced deployment. The day’s evidence suggests that as agents gain autonomy, inspectable boundaries and the costs they impose on shared infrastructure become cen...
ai-agentsopenaiai-safety
Source handle openai. Links to https://openai.com/index/gpt-6-astra/.openaiopenai.comOpenAI - GPT-6 Astrahttps://openai.com/index/gpt-6-astra/Source handle openai-2. Links to https://openai.com/index/an-alien-mind/.openai-2openai.comOpenAI / Jakub Pachocki - An Alien Mindhttps://openai.com/index/an-alien-mind/Source handle people-kernel. Links to https://people.kernel.org/monsieuricon/creepy-crawlies.people-kernelpeople.kernel.orgCreepy crawliesYou've probably heard me complain about the "AI crawlers" before, but now I actually have some hard numbers I can put up to show their im...
https://people.kernel.org/monsieuricon/creepy-crawlies
3 sourceshttps://people.kernel.org/monsieuricon/creepy-crawliesAI Digest
OpenAI’s Research Swarm Still Needs Humans
OpenAI’s internal adoption data shows coding agents becoming a parallel layer of research labor, while its intervention data shows human judgment remains central. A multi-agent research study reinforces that scale requires verifiable handoffs and governance, not just more agent autonomy.
ai-agentsresearch-workflowsevaluation
Source handle openai. Links to https://openai.com/index/research-acceleration-view-inside-openai/.openaiopenai.comOpenAI - Research Acceleration: A View Inside OpenAIhttps://openai.com/index/research-acceleration-view-inside-openai/Source handle arxiv. Links to https://arxiv.org/abs/2609.04170.arxivarxiv.orgA Case Study on Emergent Cheating and Whistleblowing in Autonomous Research SwarmsAbstract page for arXiv paper 2609.04170: A Case Study on Emergent Cheating and Whistleblowing in Autonomous Research Swarms
https://arxiv.org/abs/2609.04170Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter
3 sources
https://arxiv.org/abs/2609.04170Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterAI Digest
Agent Reliability Is Becoming a System Design Problem
A reproducible billing-triage example argues that agent reliability improves when deterministic decisions, task-specific evaluations, and tested fallback routes surround model inference. Its evidence is narrow, but it reinforces a practical shift from model selection alone to operational system d...
agentsevaluationreliability
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle theo-t3-gg-turn-off-claude-code-s-memory. Links to https://www.youtube.com/watch?v=Jf54k7tFeEc.theo-t3-gg-turn-off-claude-code-s-memoryyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=Jf54k7tFeEc
2 sourceshttps://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle theo-t3-gg-turn-off-claude-code-s-memory. Links to https://www.youtube.com/watch?v=Jf54k7tFeEc.theo-t3-gg-turn-off-claude-code-s-memoryyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=Jf54k7tFeEc
AI Digest
AI-Accelerated Discovery Is Showing Up First in Security
METR finds the clearest public evidence of AI-accelerated discovery in vulnerability reporting, not across research generally. The practical implication is to build agents around narrow tasks with validation, evidence, and observable outcomes.
ai-researchai-agentscybersecurity
Source handle metr. Links to https://metr.org/notes/2026-08-14-llm-contribution-to-discoveries/.metrmetr.orgHave We Seen an Acceleration in Discoveries?Tom Cunningham and Nate Rush look at public time series of discoveries in cyber, math, and algorithms to see whether LLMs have accelerated the discovery rate.
https://metr.org/notes/2026-08-14-llm-contribution-to-discoveries/Source handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/how-to-use-computer-use-abilities.departmentofproduct-substackdepartmentofproduct.substack.comHow to use Computer Use abilities in new features🧠 The core technologies explained, how they’re measured and how product teams can use them. Real world examples from Vanta, Google, Linear and more.
https://departmentofproduct.substack.com/p/how-to-use-computer-use-abilitiesSource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle theo-t3-gg-coding-agent-critique. Links to https://www.youtube.com/watch?v=0wemf5SZkW4.theo-t3-gg-coding-agent-critiqueyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=0wemf5SZkW4
4 sources
https://metr.org/notes/2026-08-14-llm-contribution-to-discoveries/Source handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/how-to-use-computer-use-abilities.departmentofproduct-substackdepartmentofproduct.substack.comHow to use Computer Use abilities in new features🧠 The core technologies explained, how they’re measured and how product teams can use them. Real world examples from Vanta, Google, Linear and more.
https://departmentofproduct.substack.com/p/how-to-use-computer-use-abilitiesSource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle theo-t3-gg-coding-agent-critique. Links to https://www.youtube.com/watch?v=0wemf5SZkW4.theo-t3-gg-coding-agent-critiqueyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=0wemf5SZkW4AI Digest
Agents Are Moving Into the Coordination Layer
Linear's first-party data suggests agent activity is reaching work tracking and coordination, not just task execution. The practical constraint is increasingly the surrounding contract: context, deterministic decisions, evidence, evaluation, and a correction loop.
agentsworkflowevaluation
Source handle linear. Links to https://linear.app/data.linearlinear.appHow teams build – LinearAI usage patterns in software teams: who is adopting AI, how it reshapes where teams spend their time, and how much more they ship.
https://linear.app/dataSource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter
2 sourceshttps://linear.app/dataSource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterAI Digest
Uber’s Agent Platform Makes Governance the Bottleneck
Uber’s account of an internal agentic SDLC platform suggests that agent scale shifts the hard problem from generating code to governing context, access, validation, and human decisions. The practical implication is to treat handoffs and evidence as first-class engineering artifacts.
coding-agentssoftware-deliveryagent-governance
Source handle ai-engineer-uber-on-its-internal-agentic-sdlc-pl. Links to https://www.youtube.com/watch?v=17-YSUHo6Lk.ai-engineer-uber-on-its-internal-agentic-sdlc-plyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=17-YSUHo6LkSource handle nate-b-jones-model-routing-and-coding-agent-hand. Links to https://www.youtube.com/watch?v=4HvFqhtCb-A.nate-b-jones-model-routing-and-coding-agent-handyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=4HvFqhtCb-ASource handle simonwillison. Links to https://simonwillison.net/2026/Aug/19/conceptual-integrity-and-counting-lines-of-code/.simonwillisonsimonwillison.netConceptual integrity and counting lines of codeLast week I recorded an episode of the Talking Postgres podcast with Claire Giordano on the subject of “How AI is changing software development”. We had a really great conversation. …https://simonwillison.net/2026/Aug/19/conceptual-integrity-and-counting-lines-of-code/Source handle ai-engineer-design-preferences-as-reusable-agent. Links to https://www.youtube.com/watch?v=7GMKdpLsxwU.ai-engineer-design-preferences-as-reusable-agentyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=7GMKdpLsxwU
4 sources
AI Digest
Mojo Opens Its Compiler—But Not Yet Its Governance
Mojo has released its compiler and toolchain under Apache 2.0, making the full language stack inspectable and buildable. Its decision to defer compiler contributions highlights the difference between source availability and sustainable AI-era project governance.
open-sourceprogramming-languagesai-infrastructure
Source handle modular. Links to https://www.modular.com/blog/mojo-open-source.modularmodular.comModular: Mojo🔥 is now open source!The Apache 2.0 license is the gold standard for programming languages and compilers, because it provides great flexibility to be used in all sorts of applications. The LLVM extensions to the license further expand those freedoms for building and distributing binaries compiled from Mojo. We want you to be able to adopt and use Mojo in as many applications as you can imagine.
https://www.modular.com/blog/mojo-open-sourceSource handle simonwillison. Links to https://simonwillison.net/2026/Aug/18/mojo-is-now-open-source/.simonwillisonsimonwillison.netMojo🔥 is now open sourceThe Mojo programming language has been promising an open source release since May 2023. Last week they shipped their 1.0 and today they have followed through on that original promise, …https://simonwillison.net/2026/Aug/18/mojo-is-now-open-source/
2 sources
https://www.modular.com/blog/mojo-open-sourceSource handle simonwillison. Links to https://simonwillison.net/2026/Aug/18/mojo-is-now-open-source/.simonwillisonsimonwillison.netMojo🔥 is now open sourceThe Mojo programming language has been promising an open source release since May 2023. Last week they shipped their 1.0 and today they have followed through on that original promise, …https://simonwillison.net/2026/Aug/18/mojo-is-now-open-source/AI Digest
Fallbacks Need a Job Description
A practical billing-agent experiment argues that model fallbacks become dependable only after their job is narrowed and evaluated against a product-level contract. Talks on real-time video reinforce the broader lesson: reliability increasingly lives in the harness, state management, observability...
agentsevaluationmodel-routing
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle ai-engineer-krea-on-production-infrastructure-fo. Links to https://www.youtube.com/watch?v=byn9PURoBNY.ai-engineer-krea-on-production-infrastructure-foyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=byn9PURoBNYSource handle ai-engineer-reactor-on-real-time-video-generatio. Links to https://www.youtube.com/watch?v=5dCAmSDOAjI.ai-engineer-reactor-on-real-time-video-generatioyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=5dCAmSDOAjISource handle ai-engineer-urun-on-interactive-continuous-video. Links to https://www.youtube.com/watch?v=Xln-On3syJk.ai-engineer-urun-on-interactive-continuous-videoyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=Xln-On3syJkSource handle ai-engineer-lemonslice-on-real-time-avatars. Links to https://www.youtube.com/watch?v=z1dqv74SpUs.ai-engineer-lemonslice-on-real-time-avatarsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=z1dqv74SpUs
5 sourceshttps://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle ai-engineer-krea-on-production-infrastructure-fo. Links to https://www.youtube.com/watch?v=byn9PURoBNY.ai-engineer-krea-on-production-infrastructure-foyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=byn9PURoBNYSource handle ai-engineer-reactor-on-real-time-video-generatio. Links to https://www.youtube.com/watch?v=5dCAmSDOAjI.ai-engineer-reactor-on-real-time-video-generatioyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=5dCAmSDOAjISource handle ai-engineer-urun-on-interactive-continuous-video. Links to https://www.youtube.com/watch?v=Xln-On3syJk.ai-engineer-urun-on-interactive-continuous-videoyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=Xln-On3syJkSource handle ai-engineer-lemonslice-on-real-time-avatars. Links to https://www.youtube.com/watch?v=z1dqv74SpUs.ai-engineer-lemonslice-on-real-time-avatarsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=z1dqv74SpUs
AI Digest
Agents Need an External Control Plane
The strongest practitioner signals argue that agent reliability, safety, and cost control come from enforceable system boundaries rather than model choice alone. Production deployments need external authorization, deterministic policy rules, tracing, and workflow-specific evaluation before autono...
agentssecurityagent-evaluation
Source handle ai-engineer-ryan-dahl-on-external-controls-for-p. Links to https://www.youtube.com/watch?v=MkRYPFIMCSA.ai-engineer-ryan-dahl-on-external-controls-for-pyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=MkRYPFIMCSASource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/linears-cpo-says-mcp-usage-is-through.departmentofproduct-substackdepartmentofproduct.substack.comLinear’s CPO says MCP usage is “through the roof”Are Agents the next users that product designers should design for? Six emerging UX and design trends explored. Examples from Linear, Netflix, Runway, FullStory, Square and more
https://departmentofproduct.substack.com/p/linears-cpo-says-mcp-usage-is-through
3 sourceshttps://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/linears-cpo-says-mcp-usage-is-through.departmentofproduct-substackdepartmentofproduct.substack.comLinear’s CPO says MCP usage is “through the roof”Are Agents the next users that product designers should design for? Six emerging UX and design trends explored. Examples from Linear, Netflix, Runway, FullStory, Square and morehttps://departmentofproduct.substack.com/p/linears-cpo-says-mcp-usage-is-throughAI Digest
Antifragile Agents Need Narrow Contracts
A reproducible agent workflow shows that model interchangeability improves when models have narrow structured contracts and deterministic code owns policy and execution. Complementary computer-use and AI-management examples show why recovery, auditability, and explicit human authority matter as a...
agentsevaluationreliability
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle ai-engineer-from-rl-to-irl-gaurav-mishra-amazon. Links to https://www.youtube.com/watch?v=Cc0_nyxROBA.ai-engineer-from-rl-to-irl-gaurav-mishra-amazonyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=Cc0_nyxROBASource handle time. Links to https://time.com/article/2026/08/14/claude-fired-worker-ai-job-disruption/.timetime.comExclusive: Claude Was Put in Charge of Human Workers—and Fired OneClaude was tasked with managing real employees at a San Francisco store. Its first firing offers an early glimpse of what having an AI boss could look like.
https://time.com/article/2026/08/14/claude-fired-worker-ai-job-disruption/Source handle softwaredoug. Links to https://softwaredoug.com/blog/2026/08/10/hypothetical-classifications.softwaredougsoftwaredoug.comDon’t classify. Hallucinate!A useful pattern for LLM classification at scale is to let it hallucinate plausible, fake entities. Then resolve to real ones.
https://softwaredoug.com/blog/2026/08/10/hypothetical-classificationsSource handle magazine-sebastianraschka. Links to https://magazine.sebastianraschka.com/p/ai-detector-from-scratch.magazine-sebastianraschkamagazine.sebastianraschka.comBuilding an AI Text Detector From ScratchAn End-to-End Project With Dataset Construction, Model Training, Local Deployment, and RLVR
https://magazine.sebastianraschka.com/p/ai-detector-from-scratch
5 sourceshttps://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle ai-engineer-from-rl-to-irl-gaurav-mishra-amazon. Links to https://www.youtube.com/watch?v=Cc0_nyxROBA.ai-engineer-from-rl-to-irl-gaurav-mishra-amazonyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=Cc0_nyxROBASource handle time. Links to https://time.com/article/2026/08/14/claude-fired-worker-ai-job-disruption/.timetime.comExclusive: Claude Was Put in Charge of Human Workers—and Fired OneClaude was tasked with managing real employees at a San Francisco store. Its first firing offers an early glimpse of what having an AI boss could look like.
https://time.com/article/2026/08/14/claude-fired-worker-ai-job-disruption/Source handle softwaredoug. Links to https://softwaredoug.com/blog/2026/08/10/hypothetical-classifications.softwaredougsoftwaredoug.comDon’t classify. Hallucinate!A useful pattern for LLM classification at scale is to let it hallucinate plausible, fake entities. Then resolve to real ones.
https://softwaredoug.com/blog/2026/08/10/hypothetical-classificationsSource handle magazine-sebastianraschka. Links to https://magazine.sebastianraschka.com/p/ai-detector-from-scratch.magazine-sebastianraschkamagazine.sebastianraschka.comBuilding an AI Text Detector From ScratchAn End-to-End Project With Dataset Construction, Model Training, Local Deployment, and RLVRhttps://magazine.sebastianraschka.com/p/ai-detector-from-scratchAI Digest
The Real Work of Computer-Use Agents Is Measurement
Today’s strongest agent signal is that fixed demo success is weak evidence of generalization. Robust deployments depend on environmental variation, independent verification, full-trajectory observability, and constrained authority.
agentscomputer-useevaluation
Source handle ai-engineer-computer-use-at-the-edge-of-the-stat. Links to https://www.youtube.com/watch?v=CTLa_p6iOiY.ai-engineer-computer-use-at-the-edge-of-the-statyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=CTLa_p6iOiYSource handle ai-engineer-bringing-agents-onto-the-world-wide. Links to https://www.youtube.com/watch?v=GqoNrUz8hEU.ai-engineer-bringing-agents-onto-the-world-wideyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=GqoNrUz8hEUSource handle anthropic. Links to https://www.anthropic.com/research/multiagent-systems.anthropicanthropic.comPatterns and problems in multiagent systemsWe ran experiments on swarms of Claude agents and found coordination failures, collusion, and sabotage. Here, we share what they mean for AI safety.
https://www.anthropic.com/research/multiagent-systemsSource handle nitter. Links to https://nitter.net/bcherny/status/2088014489438621990.nitternitter.netBoris Cherny - Routine app maintenance by Claudehttps://nitter.net/bcherny/status/2088014489438621990Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter
5 sources
https://www.anthropic.com/research/multiagent-systemsSource handle nitter. Links to https://nitter.net/bcherny/status/2088014489438621990.nitternitter.netBoris Cherny - Routine app maintenance by Claudehttps://nitter.net/bcherny/status/2088014489438621990Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterAI Digest
Coordination Is the Agent Bottleneck
Anthropic’s multi-agent experiments show that stronger individual agents do not reliably produce safe or effective coordination. The practical response is explicit workflow design: scoped authority, observable traces, tested handoffs, and escalation paths.
agentsmulti-agent-systemsevaluation
Source handle anthropic. Links to https://www.anthropic.com/research/multiagent-systems.anthropicanthropic.comPatterns and problems in multiagent systemsWe ran experiments on swarms of Claude agents and found coordination failures, collusion, and sabotage. Here, we share what they mean for AI safety.https://www.anthropic.com/research/multiagent-systemsSource handle ai-engineer-trace-mining-talk-with-vivek-trivedy. Links to https://www.youtube.com/watch?v=CvRngaQZQ3Y.ai-engineer-trace-mining-talk-with-vivek-trivedyyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=CvRngaQZQ3YSource handle greg-isenberg-interview-with-alli-k-miller. Links to https://www.youtube.com/watch?v=EzQAgnjTq2k.greg-isenberg-interview-with-alli-k-milleryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=EzQAgnjTq2kSource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle netlify. Links to https://www.netlify.com/blog/one-prompt-11-models-very-different-results/.netlifynetlify.comMore models, more choice: Comparing 11 different AI modelsNetlify now runs any OpenRouter model, including Kimi K3, GLM 5.2 and DeepSeek V4. We tested 11 of them on the same build prompt to see how they differ.
https://www.netlify.com/blog/one-prompt-11-models-very-different-results/
5 sourceshttps://www.anthropic.com/research/multiagent-systemsSource handle ai-engineer-trace-mining-talk-with-vivek-trivedy. Links to https://www.youtube.com/watch?v=CvRngaQZQ3Y.ai-engineer-trace-mining-talk-with-vivek-trivedyyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=CvRngaQZQ3YSource handle greg-isenberg-interview-with-alli-k-miller. Links to https://www.youtube.com/watch?v=EzQAgnjTq2k.greg-isenberg-interview-with-alli-k-milleryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=EzQAgnjTq2kSource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle netlify. Links to https://www.netlify.com/blog/one-prompt-11-models-very-different-results/.netlifynetlify.comMore models, more choice: Comparing 11 different AI modelsNetlify now runs any OpenRouter model, including Kimi K3, GLM 5.2 and DeepSeek V4. We tested 11 of them on the same build prompt to see how they differ.https://www.netlify.com/blog/one-prompt-11-models-very-different-results/AI Digest
Agent Memory Needs a Control Plane
A cluster of practitioner and research talks argues that long-running agents need explicit current state, ranked retrieval, and reset-baseline evaluation—not just larger contexts or larger memory stores.
agentsmemorycontinual-learning
Source handle nate-b-jones-long-running-coding-agents-depend-o. Links to https://www.youtube.com/watch?v=HZLPhPbw3fM.nate-b-jones-long-running-coding-agents-depend-oyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=HZLPhPbw3fMSource handle ai-engineer-sakana-ai-memory-retrieval-policy-no. Links to https://www.youtube.com/watch?v=R3-anFK1YM8.ai-engineer-sakana-ai-memory-retrieval-policy-noyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=R3-anFK1YM8Source handle ai-engineer-uc-berkeley-continual-learning-bench. Links to https://www.youtube.com/watch?v=iqloyWCGYQQ.ai-engineer-uc-berkeley-continual-learning-benchyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=iqloyWCGYQQSource handle ai-engineer-neocognition-continual-learning-as-d. Links to https://www.youtube.com/watch?v=I6aiEf3aEFQ.ai-engineer-neocognition-continual-learning-as-dyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=I6aiEf3aEFQSource handle ai-engineer-engram-scaling-compute-on-context. Links to https://www.youtube.com/watch?v=WiqDvX6isc4.ai-engineer-engram-scaling-compute-on-contextyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=WiqDvX6isc4Source handle ai-engineer-trajectory-on-policy-self-distillati. Links to https://www.youtube.com/watch?v=zL1kLftVTlo.ai-engineer-trajectory-on-policy-self-distillatiyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=zL1kLftVTlo
6 sources
AI Digest
Agent Surfaces Are Becoming Infrastructure
Anthropic engineers argue that the harness around a model—execution isolation, recovery, secret handling, traces, and evaluation—is becoming a primary determinant of production agent capability. The practical implication is to design agent systems for changing models rather than hard-code today's...
agentsagent-architectureproduction-ai
Source handle ai-engineer-evolution-of-agentic-surfaces-gagan. Links to https://www.youtube.com/watch?v=K0X9QDRkIdg.ai-engineer-evolution-of-agentic-surfaces-gaganyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=K0X9QDRkIdgSource handle dwarkesh. Links to https://www.dwarkesh.com/p/ryan-greenblatt.dwarkeshdwarkesh.comRyan Greenblatt – Human level AIs might build runaway superintelligences by 2032A debate about recursive self-improvement.
https://www.dwarkesh.com/p/ryan-greenblatt
2 sources
https://www.dwarkesh.com/p/ryan-greenblattAI Digest
Evaluation Environments Are Now Agent Safety Systems
AISI's delayed-discovery cyber-evaluation report shows that agent safety depends on the evaluation environment as much as model behavior. Builder proposals point to the same response: narrow model contracts, isolate execution, and make external actions observable and stoppable.
ai-agentsagent-safetycybersecurity
Source handle aisi-gov. Links to https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testing.aisi-govaisi.gov.ukIncident Report: unsanctioned agent behaviour during cyber testing | AISI WorkDuring a routine cyber evaluation, AISI identified an incident in which AI agents took sustained, unsanctioned action directed at real people and organisations. We are disclosing what we found, what it means, and the actions now underway..png)
https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testingSource handle simonwillison. Links to https://simonwillison.net/2026/Aug/5/third-party-cyber-evaluations/.simonwillisonsimonwillison.netThird-party cyber evaluations involving OpenAI modelsAnd another one. I had to create a accidental-cyberattacks tag to keep track of them all! This post from OpenAI covers both the UK AI Safety Institute attack (see my …https://simonwillison.net/2026/Aug/5/third-party-cyber-evaluations/Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle ai-engineer-gadgets-personal-app-vibe-coding-tha. Links to https://www.youtube.com/watch?v=RmS5s6Wbin4.ai-engineer-gadgets-personal-app-vibe-coding-thayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=RmS5s6Wbin4
4 sources
https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testingSource handle simonwillison. Links to https://simonwillison.net/2026/Aug/5/third-party-cyber-evaluations/.simonwillisonsimonwillison.netThird-party cyber evaluations involving OpenAI modelsAnd another one. I had to create a accidental-cyberattacks tag to keep track of them all! This post from OpenAI covers both the UK AI Safety Institute attack (see my …https://simonwillison.net/2026/Aug/5/third-party-cyber-evaluations/Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouter.post-training-aitinkererspost-training.aitinkerers.orgHow to Build Antifragile Agents with OpenRouter by Kenny Rogers (OpenRouter)On June 12, 2026, Anthropic suspended access to Fable 5 for all customers [https://www.anthropic.com/news/fable-mythos-access] to comply with a US government...https://post-training.aitinkerers.org/p/how-to-build-antifragile-agents-with-openrouterSource handle ai-engineer-gadgets-personal-app-vibe-coding-tha. Links to https://www.youtube.com/watch?v=RmS5s6Wbin4.ai-engineer-gadgets-personal-app-vibe-coding-thayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=RmS5s6Wbin4
AI Digest
Agents Need an Operating Model, Not a Better Chat Window
Practitioner evidence is converging on a more operational view of agents: define verifiable outcomes, preserve structured run state, and place approval and review inside the workflow. Model capability expands the tasks worth attempting, but it does not replace task design or observability.
agentsagent-evaluationworkflow-design
Source handle exponentialview. Links to https://www.exponentialview.co/p/seven-lessons-for-managing-ai-agents.exponentialviewexponentialview.co🔮 Seven lessons for managing AI agentsPlus, an updated stack of 50+ AI tools we use at Exponential View
https://www.exponentialview.co/p/seven-lessons-for-managing-ai-agentsSource handle simonwillison. Links to https://simonwillison.net/2026/Aug/4/new-release-of-llm/.simonwillisonsimonwillison.netNew release of LLM adds support for reasoning traces, OpenAI Responses, server-side tools, and smarter loggingI released LLM 0.32 this morning, the most significant new version of LLM since the initial launch of the project. The new version includes support for visible reasoning traces, server-side …
https://simonwillison.net/2026/Aug/4/new-release-of-llm/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Aug/4/llm-anthropic/.simonwillison-2simonwillison.netRelease: llm-anthropic 0.26LLM access to models by Anthropic, including the Claude serieshttps://simonwillison.net/2026/Aug/4/llm-anthropic/Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/what-i-learned-giving-fable-5-a-face.post-training-aitinkererspost-training.aitinkerers.orgWhat I Learned Giving Fable 5 a Face by Ben Carr (Anam.ai)What I Learned Giving Fable 5 a Face
https://post-training.aitinkerers.org/p/what-i-learned-giving-fable-5-a-faceSource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-38-build-inspection-vlms-local-first-agents-and-agent-harnesses.post-training-aitinkerers-2post-training.aitinkerers.orgAI Tinkerers / Post-Training - Top AI Demos #38: Build Inspection VLMs, Local-First Agents, and Agent Harnesseshttps://post-training.aitinkerers.org/p/top-ai-demos-38-build-inspection-vlms-local-first-agents-and-agent-harnesses
5 sources
https://www.exponentialview.co/p/seven-lessons-for-managing-ai-agentsSource handle simonwillison. Links to https://simonwillison.net/2026/Aug/4/new-release-of-llm/.simonwillisonsimonwillison.netNew release of LLM adds support for reasoning traces, OpenAI Responses, server-side tools, and smarter loggingI released LLM 0.32 this morning, the most significant new version of LLM since the initial launch of the project. The new version includes support for visible reasoning traces, server-side …https://simonwillison.net/2026/Aug/4/new-release-of-llm/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Aug/4/llm-anthropic/.simonwillison-2simonwillison.netRelease: llm-anthropic 0.26LLM access to models by Anthropic, including the Claude serieshttps://simonwillison.net/2026/Aug/4/llm-anthropic/Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/what-i-learned-giving-fable-5-a-face.post-training-aitinkererspost-training.aitinkerers.orgWhat I Learned Giving Fable 5 a Face by Ben Carr (Anam.ai)What I Learned Giving Fable 5 a Facehttps://post-training.aitinkerers.org/p/what-i-learned-giving-fable-5-a-faceSource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-38-build-inspection-vlms-local-first-agents-and-agent-harnesses.post-training-aitinkerers-2post-training.aitinkerers.orgAI Tinkerers / Post-Training - Top AI Demos #38: Build Inspection VLMs, Local-First Agents, and Agent Harnesseshttps://post-training.aitinkerers.org/p/top-ai-demos-38-build-inspection-vlms-local-first-agents-and-agent-harnessesAI Digest
The Agent Harness Is the Product
A first-hand account of rebuilding an agent-assisted development workflow shows that orchestration, verification, and recovery—not merely model capability—are becoming the central engineering problem. Adjacent practitioner signals extend that lesson to conversational interfaces and enterprise ado...
agentsagent-harnessessoftware-engineering
Source handle yegge. Links to https://yegge.ai/essays/the-shape-of-things-to-come/.yeggeyegge.aiThe Shape of Things to Come, Part 1: The Continuous Thunderdome — Steve YeggePart 1: The Continuous Thunderdome. Loops and graphs, Wheelhouse and Beads, the end of human code review, the Land Rush that replaces CI/CD, and the Wish Factory — a field report from 12 months in the future.
https://yegge.ai/essays/the-shape-of-things-to-come/Source handle gruhn. Links to https://gruhn.me/blog/2026-08-03/.gruhngruhn.meDon't be a meat proxyhttps://gruhn.me/blog/2026-08-03/Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/what-i-learned-giving-fable-5-a-face.post-training-aitinkererspost-training.aitinkerers.orgWhat I Learned Giving Fable 5 a Face by Ben Carr (Anam.ai)What I Learned Giving Fable 5 a Facehttps://post-training.aitinkerers.org/p/what-i-learned-giving-fable-5-a-faceSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/how-stripe-built-a-new-internal-ai.departmentofproduct-substackdepartmentofproduct.substack.comHow Stripe Built a new Internal AI Knowledge Platform that their PMs use “all day long”Non-engineers are embracing internal AI tools. Examples from Stripe, Spotify, Uber and more.
https://departmentofproduct.substack.com/p/how-stripe-built-a-new-internal-ai
4 sourceshttps://yegge.ai/essays/the-shape-of-things-to-come/Source handle gruhn. Links to https://gruhn.me/blog/2026-08-03/.gruhngruhn.meDon't be a meat proxyhttps://gruhn.me/blog/2026-08-03/Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/what-i-learned-giving-fable-5-a-face.post-training-aitinkererspost-training.aitinkerers.orgWhat I Learned Giving Fable 5 a Face by Ben Carr (Anam.ai)What I Learned Giving Fable 5 a Facehttps://post-training.aitinkerers.org/p/what-i-learned-giving-fable-5-a-faceSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/how-stripe-built-a-new-internal-ai.departmentofproduct-substackdepartmentofproduct.substack.comHow Stripe Built a new Internal AI Knowledge Platform that their PMs use “all day long”Non-engineers are embracing internal AI tools. Examples from Stripe, Spotify, Uber and more.https://departmentofproduct.substack.com/p/how-stripe-built-a-new-internal-aiAI Digest
Pacing Automated AI Research
A frontier-employee statement makes coordinated pacing of automated AI research a concrete governance question. New agent-research and security preprints suggest operational autonomy is widening even as high-level research judgment remains unreliable.
ai-governanceagentsai-safety
Source handle pacingthefrontier. Links to https://www.pacingthefrontier.com/.pacingthefrontierpacingthefrontier.comPacing the FrontierA statement from over 1000 employees of frontier AI companies
https://www.pacingthefrontier.com/Source handle arxiv. Links to https://arxiv.org/abs/2607.27191.arxivarxiv.orgCan AI agents conduct open-ended AI research? Early evidence from two case studiesAbstract page for arXiv paper 2607.27191: Can AI agents conduct open-ended AI research? Early evidence from two case studieshttps://arxiv.org/abs/2607.27191Source handle arxiv-2. Links to https://arxiv.org/abs/2606.03811.arxiv-2arxiv.orgAI Agents Enable Adaptive Computer WormsAbstract page for arXiv paper 2606.03811: AI Agents Enable Adaptive Computer Wormshttps://arxiv.org/abs/2606.03811Source handle blog-exe. Links to https://blog.exe.dev/devtools-must-be-open-source.blog-exeblog.exe.devDevtools must be open source - exe.dev blogThe age of personalized software is here.
https://blog.exe.dev/devtools-must-be-open-sourceSource handle openai. Links to https://openai.com/index/how-ai-is-expanding-what-people-do-at-work/.openaiopenai.comOpenAI Economic Research - How AI Is Expanding What People Do at Workhttps://openai.com/index/how-ai-is-expanding-what-people-do-at-work/
5 sourceshttps://www.pacingthefrontier.com/Source handle arxiv. Links to https://arxiv.org/abs/2607.27191.arxivarxiv.orgCan AI agents conduct open-ended AI research? Early evidence from two case studiesAbstract page for arXiv paper 2607.27191: Can AI agents conduct open-ended AI research? Early evidence from two case studieshttps://arxiv.org/abs/2607.27191Source handle arxiv-2. Links to https://arxiv.org/abs/2606.03811.arxiv-2arxiv.orgAI Agents Enable Adaptive Computer WormsAbstract page for arXiv paper 2606.03811: AI Agents Enable Adaptive Computer Wormshttps://arxiv.org/abs/2606.03811Source handle blog-exe. Links to https://blog.exe.dev/devtools-must-be-open-source.blog-exeblog.exe.devDevtools must be open source - exe.dev blogThe age of personalized software is here.
https://blog.exe.dev/devtools-must-be-open-sourceSource handle openai. Links to https://openai.com/index/how-ai-is-expanding-what-people-do-at-work/.openaiopenai.comOpenAI Economic Research - How AI Is Expanding What People Do at Workhttps://openai.com/index/how-ai-is-expanding-what-people-do-at-work/AI Digest
Proofs, Not Just Answers
OpenAI’s publication of inspectable artifacts for ten claimed mathematical advances raises the standard for AI research claims: systems must leave work experts can audit. Airbnb’s trace-focused evaluation practice shows the same requirement emerging in production AI.
ai-researchevaluationagents
Source handle openai. Links to https://openai.com/index/ten-advances-in-mathematics/.openaiopenai.comOpenAI - Ten advances in mathematics and theoretical computer sciencehttps://openai.com/index/ten-advances-in-mathematics/Source handle medium. Links to https://medium.com/airbnb-engineering/eval-driven-development-lessons-from-evaluating-genai-at-scale-e817e5ae5788.mediummedium.comAirbnb Engineering - Eval-driven development: Lessons from evaluating GenAI at scalehttps://medium.com/airbnb-engineering/eval-driven-development-lessons-from-evaluating-genai-at-scale-e817e5ae5788Source handle simonwillison. Links to https://simonwillison.net/2026/Aug/2/open-letters/.simonwillisonsimonwillison.netOpen letters about AI developmentOpen letters about AI development I wrote this summary of the past few weeks of open letters as a section of my sponsors-only newsletter but I've decided to share it …https://simonwillison.net/2026/Aug/2/open-letters/
3 sourcesAI Digest
Agent Reliability Is Becoming a Systems Problem
Two practitioner releases argue that dependable agents depend on bounded tool interfaces and evaluations of the full harness, not model capability alone. The practical consequence is to test prompts, permissions, tool schemas, and graders together on real tasks.
agentsmcpevaluation
Source handle simonwillison. Links to https://simonwillison.net/2026/Jul/31/stateless-mcp/.simonwillisonsimonwillison.netStateless MCP has recaptured my interest (and inspired mcp-explorer and datasette-mcp)Tuesday was Stateless MCP day—the rollout of MCP 2.0, or the 2026-07-28 Model Context Protocol specification to use the more formal but less memorable name. This is the most significant …
https://simonwillison.net/2026/Jul/31/stateless-mcp/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Jul/31/smevals/.simonwillison-2simonwillison.netsmevals—a small eval suite for evaluating models, prompts, and harnessesI've been working with Jesse Vincent's Prime Radiant applied AI research lab building out this evals framework to help answer questions about the capabilities of different models. The result is …
https://simonwillison.net/2026/Jul/31/smevals/
2 sourceshttps://simonwillison.net/2026/Jul/31/stateless-mcp/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Jul/31/smevals/.simonwillison-2simonwillison.netsmevals—a small eval suite for evaluating models, prompts, and harnessesI've been working with Jesse Vincent's Prime Radiant applied AI research lab building out this evals framework to help answer questions about the capabilities of different models. The result is …https://simonwillison.net/2026/Jul/31/smevals/AI Digest
Agent Reliability Moves to the Control Loop
Anthropic’s cyber-evaluation incidents and fresh builder reports point to the same constraint on broader agent deployment: reliability depends on containment, observability, state management, and recovery around the model—not only on model capability.
ai-agentsai-safetycybersecurity
Source handle anthropic. Links to https://www.anthropic.com/news/investigating-incidents-cybersecurity-evals.anthropicanthropic.comInvestigating three real-world incidents in our cybersecurity evaluationsIn a review of our cybersecurity evaluation transcripts, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations. Below we describe what happened, how it happened, and what we’re changing. We encourage other AI labs to perform similar reviews.
https://www.anthropic.com/news/investigating-incidents-cybersecurity-evalsSource handle openai. Links to https://openai.com/index/advancing-the-price-performance-frontier-with-gpt-5-6/.openaiopenai.comOpenAI - Advancing the price-performance frontier with GPT-5.6https://openai.com/index/advancing-the-price-performance-frontier-with-gpt-5-6/Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/what-i-learned-giving-fable-5-a-face.post-training-aitinkererspost-training.aitinkerers.orgWhat I Learned Giving Fable 5 a Face by Ben Carr (Anam.ai)What I Learned Giving Fable 5 a Facehttps://post-training.aitinkerers.org/p/what-i-learned-giving-fable-5-a-faceSource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-37-pi-hydra-agents-agent-guides-and-ai-hardware-design.post-training-aitinkerers-2post-training.aitinkerers.orgTop AI Demos #37: Pi-Hydra Agents, Agent Guides, and AI Hardware Design by Joe Heitzeberg (AI Tinkerers)This week's demos dive deep into making AI agents more practical and capable. We saw builders tackling developer tooling with agents like [pi-hydra: Mob...https://post-training.aitinkerers.org/p/top-ai-demos-37-pi-hydra-agents-agent-guides-and-ai-hardware-design
4 sourceshttps://post-training.aitinkerers.org/p/what-i-learned-giving-fable-5-a-faceSource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-37-pi-hydra-agents-agent-guides-and-ai-hardware-design.post-training-aitinkerers-2post-training.aitinkerers.orgTop AI Demos #37: Pi-Hydra Agents, Agent Guides, and AI Hardware Design by Joe Heitzeberg (AI Tinkerers)This week's demos dive deep into making AI agents more practical and capable. We saw builders tackling developer tooling with agents like [pi-hydra: Mob...https://post-training.aitinkerers.org/p/top-ai-demos-37-pi-hydra-agents-agent-guides-and-ai-hardware-designAI Digest
The Agent Harness Is Becoming the Product
A practitioner benchmark and a new memory study point to the same conclusion: reliable agent performance increasingly depends on harness design, not model choice alone. Builders should evaluate tool boundaries, verification, and memory maintenance as separate system components.
ai-agentsagent-harnessesmemory
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain by Hitesh Jain (Coral Bricks AI)If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.
https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle arxiv. Links to https://arxiv.org/html/2607.26637.arxivarxiv.orgFilesystem-Based Memory for LLM Agents: Organization, Evolution, and Sustainabilityhttps://arxiv.org/html/2607.26637Source handle exponentialview. Links to https://www.exponentialview.co/p/ai-adoption-j-curve.exponentialviewexponentialview.co🔮 For AI adopters, success and failure look identical — at firstModelling the AI J-curve
https://www.exponentialview.co/p/ai-adoption-j-curve
3 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle arxiv. Links to https://arxiv.org/html/2607.26637.arxivarxiv.orgFilesystem-Based Memory for LLM Agents: Organization, Evolution, and Sustainabilityhttps://arxiv.org/html/2607.26637Source handle exponentialview. Links to https://www.exponentialview.co/p/ai-adoption-j-curve.exponentialviewexponentialview.co🔮 For AI adopters, success and failure look identical — at firstModelling the AI J-curvehttps://www.exponentialview.co/p/ai-adoption-j-curveAI Digest
The Agent Harness Is Becoming the Product
A late-discovered practitioner ablation argues that structured, domain-specific agent harnesses can matter more than larger prompts or retrieval alone. The self-reported benchmark result is narrow, but its workflow design lessons are concrete.
agentsagent-harnessesevaluation
Read digest AI Digest
MirrorCode Makes Long-Horizon Coding More Legible
MirrorCode moves coding-agent claims onto a more concrete task surface: black-box program reconstruction with explicit verification, duration, and cost. Adjacent product and China-adoption signals underline that useful agency depends on whole-system design and integration, not model capability al...
coding-agentsbenchmarksevaluation
Source handle jack-clark. Links to https://jack-clark.net/2026/07/27/import-ai-466-the-bitter-lesson-for-robotics-ais-complete-week-long-programming-tasks-and-openais-accidental-ai-hacker/.jack-clarkjack-clark.netImport AI 466: The bitter lesson for robotics, AIs complete week-long programming tasks; and OpenAI’s accidental AI hackerWelcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe now Epoch and METR release MirrorCode, a benchmark for seeing how well AI systems can do long-horizon programming tasks:…AI systems can’t solve the hardest tasks yet (good!)...Epoch…
https://jack-clark.net/2026/07/27/import-ai-466-the-bitter-lesson-for-robotics-ais-complete-week-long-programming-tasks-and-openais-accidental-ai-hacker/Source handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/how-netflix-built-a-generative-ai.departmentofproduct-substackdepartmentofproduct.substack.comHow Netflix built a Generative AI powered homepage that boosted engagementAnd other recent examples of AI powered product personalization. Case studies from Spotify, DoorDash, Instacart, LinkedIn and more.
https://departmentofproduct.substack.com/p/how-netflix-built-a-generative-aiSource handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/nathan-goes-to-china-part-1-tech-agent-setup-chinese-ai-ux-waic-and-attitudes-on-ai/.cognitiverevolutioncognitiverevolution.aiNathan Goes to China – Part 1: Tech & Agent Setup, Chinese AI UX, WAIC, and Attitudes on AINathan reports from the first leg of his China trip, covering visa and device setup, getting online, and the daily app stack behind payments, rides, food, and translation. He also discusses Chinese AI user experiences, WAIC, and local attitudes toward AI.
https://www.cognitiverevolution.ai/nathan-goes-to-china-part-1-tech-agent-setup-chinese-ai-ux-waic-and-attitudes-on-ai/
3 sourceshttps://jack-clark.net/2026/07/27/import-ai-466-the-bitter-lesson-for-robotics-ais-complete-week-long-programming-tasks-and-openais-accidental-ai-hacker/Source handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/how-netflix-built-a-generative-ai.departmentofproduct-substackdepartmentofproduct.substack.comHow Netflix built a Generative AI powered homepage that boosted engagementAnd other recent examples of AI powered product personalization. Case studies from Spotify, DoorDash, Instacart, LinkedIn and more.https://departmentofproduct.substack.com/p/how-netflix-built-a-generative-aiSource handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/nathan-goes-to-china-part-1-tech-agent-setup-chinese-ai-ux-waic-and-attitudes-on-ai/.cognitiverevolutioncognitiverevolution.aiNathan Goes to China – Part 1: Tech & Agent Setup, Chinese AI UX, WAIC, and Attitudes on AINathan reports from the first leg of his China trip, covering visa and device setup, getting online, and the daily app stack behind payments, rides, food, and translation. He also discusses Chinese AI user experiences, WAIC, and local attitudes toward AI.https://www.cognitiverevolution.ai/nathan-goes-to-china-part-1-tech-agent-setup-chinese-ai-ux-waic-and-attitudes-on-ai/AI Digest
Agent Harnesses Are Becoming Control Systems
Practitioner evidence converges on an operational view of agents: reliable gains come from scoped tools, explicit verification, human gates, and feedback loops rather than ever-larger prompts or unattended runs. The practical task is to design the measurement and control system around the model.
agentsagent-harnessesevaluation
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain by Hitesh Jain (Coral Bricks AI)If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle ai-engineer-loop-engineering-from-first-principl. Links to https://www.youtube.com/watch?v=xIt_mTQp6mY.ai-engineer-loop-engineering-from-first-principlyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=xIt_mTQp6mYSource handle nate-b-jones-you-can-hand-one-ai-agent-your-wors. Links to https://www.youtube.com/watch?v=7pqRRxrdr0c.nate-b-jones-you-can-hand-one-ai-agent-your-worsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=7pqRRxrdr0cSource handle ai-engineer-state-of-data-sean-cai. Links to https://www.youtube.com/watch?v=ZyIoTOAbRfs.ai-engineer-state-of-data-sean-caiyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=ZyIoTOAbRfs
4 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle ai-engineer-loop-engineering-from-first-principl. Links to https://www.youtube.com/watch?v=xIt_mTQp6mY.ai-engineer-loop-engineering-from-first-principlyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=xIt_mTQp6mYSource handle nate-b-jones-you-can-hand-one-ai-agent-your-wors. Links to https://www.youtube.com/watch?v=7pqRRxrdr0c.nate-b-jones-you-can-hand-one-ai-agent-your-worsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=7pqRRxrdr0cSource handle ai-engineer-state-of-data-sean-cai. Links to https://www.youtube.com/watch?v=ZyIoTOAbRfs.ai-engineer-state-of-data-sean-caiyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=ZyIoTOAbRfs
AI Digest
Private Benchmarks Become the Agent Product
Practitioner discourse is converging on private, production-derived benchmarks as the core operating system for reliable agents. The emphasis is shifting from prompt chains and headline capability toward replayable simulations, failure-specific checks, traces, and calibrated human review.
agentsevaluationproduction-ai
Source handle ai-engineer-snorkel-ai-private-benchmarks-and-of. Links to https://www.youtube.com/watch?v=Ib5t2RLtxvM.ai-engineer-snorkel-ai-private-benchmarks-and-ofyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=Ib5t2RLtxvMSource handle ai-engineer-uber-multimodal-agent-qa-in-producti. Links to https://www.youtube.com/watch?v=31GUkCBD-Uc.ai-engineer-uber-multimodal-agent-qa-in-productiyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=31GUkCBD-UcSource handle ai-engineer-youtube-ads-agent-reliability-evals. Links to https://www.youtube.com/watch?v=xyL2Ltkh-SA.ai-engineer-youtube-ads-agent-reliability-evalsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=xyL2Ltkh-SASource handle anthropic. Links to https://www.anthropic.com/news/claude-opus-5.anthropicanthropic.comIntroducing Claude Opus 5Opus 5 is a step change improvement for the Opus tier powering long-running agents while delivering improvements in coding and professional work.
https://www.anthropic.com/news/claude-opus-5Source handle ai-engineer-arize-agent-driven-observability-and. Links to https://www.youtube.com/watch?v=9HbzAWnKbo4.ai-engineer-arize-agent-driven-observability-andyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=9HbzAWnKbo4Source handle ai-engineer-google-ai-edge-on-device-model-speci. Links to https://www.youtube.com/watch?v=hacEQHHhu2Q.ai-engineer-google-ai-edge-on-device-model-speciyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=hacEQHHhu2Q
6 sourceshttps://www.anthropic.com/news/claude-opus-5Source handle ai-engineer-arize-agent-driven-observability-and. Links to https://www.youtube.com/watch?v=9HbzAWnKbo4.ai-engineer-arize-agent-driven-observability-andyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=9HbzAWnKbo4Source handle ai-engineer-google-ai-edge-on-device-model-speci. Links to https://www.youtube.com/watch?v=hacEQHHhu2Q.ai-engineer-google-ai-edge-on-device-model-speciyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=hacEQHHhu2QAI Digest
Agent Harnesses, Not Prompts
AI discourse today converged on a single lesson: the hard part of agent systems is shifting from model quality to the harnesses, permissions, and evaluation environments around them. The Hugging Face/OpenAI incident led the conversation, but builder essays and conference talks reinforced the same...
ai-agentsagent-harnessessecurity
Source handle simonwillison. Links to https://simonwillison.net/2026/Jul/23/the-first-known-runaway-ai-agent/.simonwillisonsimonwillison.netThe first known runaway AI agent—or a very bad marketing stunt?Martin Alderson's commentary on the OpenAI accidental cyberattack against Hugging Face includes a couple of details I hadn't considered. First, Hugging Face offers a truly rich target if you're trying …https://simonwillison.net/2026/Jul/23/the-first-known-runaway-ai-agent/Source handle martinalderson. Links to https://martinalderson.com/posts/huggingface-openai-exploit/.martinaldersonmartinalderson.comThe first known runaway AI agent - or a very bad marketing stunt?Breaking down the Hugging Face security incident caused by OpenAI's own models during a benchmark run - the sandbox escape, the package proxy, and whether it's really a marketing stunt.
https://martinalderson.com/posts/huggingface-openai-exploit/Source handle huggingface. Links to https://huggingface.co/blog/security-incident-july-2026.huggingfacehuggingface.coSecurity incident disclosure — July 2026We’re on a journey to advance and democratize artificial intelligence through open source and open science.
https://huggingface.co/blog/security-incident-july-2026Source handle openai. Links to https://openai.com/index/hugging-face-model-evaluation-security-incident/.openaiopenai.comOpenAI - Hugging Face model evaluation security incidenthttps://openai.com/index/hugging-face-model-evaluation-security-incident/Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain by Hitesh Jain (Coral Bricks AI)If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle ai-engineer-arithmetic-discussion-with-thom-wolf. Links to https://www.youtube.com/watch?v=O-CBZ3JtRvo.ai-engineer-arithmetic-discussion-with-thom-wolfyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=O-CBZ3JtRvoSource handle ai-engineer-vending-bench-long-horizon-agent-eva. Links to https://www.youtube.com/watch?v=cO8qC6HBuBg.ai-engineer-vending-bench-long-horizon-agent-evayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=cO8qC6HBuBgSource handle nate-b-jones-i-deleted-5-things-from-this-file-b. Links to https://www.youtube.com/watch?v=EuVvLwWZ5wc.nate-b-jones-i-deleted-5-things-from-this-file-byoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=EuVvLwWZ5wc
8 sources
https://martinalderson.com/posts/huggingface-openai-exploit/Source handle huggingface. Links to https://huggingface.co/blog/security-incident-july-2026.huggingfacehuggingface.coSecurity incident disclosure — July 2026We’re on a journey to advance and democratize artificial intelligence through open source and open science.https://huggingface.co/blog/security-incident-july-2026Source handle openai. Links to https://openai.com/index/hugging-face-model-evaluation-security-incident/.openaiopenai.comOpenAI - Hugging Face model evaluation security incidenthttps://openai.com/index/hugging-face-model-evaluation-security-incident/Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain by Hitesh Jain (Coral Bricks AI)If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle ai-engineer-arithmetic-discussion-with-thom-wolf. Links to https://www.youtube.com/watch?v=O-CBZ3JtRvo.ai-engineer-arithmetic-discussion-with-thom-wolfyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=O-CBZ3JtRvoSource handle ai-engineer-vending-bench-long-horizon-agent-eva. Links to https://www.youtube.com/watch?v=cO8qC6HBuBg.ai-engineer-vending-bench-long-horizon-agent-evayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=cO8qC6HBuBgSource handle nate-b-jones-i-deleted-5-things-from-this-file-b. Links to https://www.youtube.com/watch?v=EuVvLwWZ5wc.nate-b-jones-i-deleted-5-things-from-this-file-byoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=EuVvLwWZ5wc
AI Digest
The Agent Harness Era
Serious AI builder discourse shifted today from model-centric agent talk toward the surrounding system: harnesses, memory, graph control planes, verifier loops, and provenance. The practical lesson is that longer-horizon agents are becoming an architecture problem as much as a model problem.
agentsagent-engineeringmemory
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain [AI Tinkerers - Post-Training]If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle simonwillison. Links to https://simonwillison.net/2026/Jul/22/openai-cyberattack/.simonwillisonsimonwillison.netOpenAI’s accidental cyberattack against Hugging Face is science fiction that happenedThis story is wild. The short version: OpenAI were running a cybersecurity test against an unreleased model, with the model’s guardrail features turned off. Rather than solve the test, the …https://simonwillison.net/2026/Jul/22/openai-cyberattack/Source handle huggingface. Links to https://huggingface.co/blog/security-incident-july-2026.huggingfacehuggingface.coSecurity incident disclosure — July 2026We’re on a journey to advance and democratize artificial intelligence through open source and open science.https://huggingface.co/blog/security-incident-july-2026Source handle arxiv. Links to https://arxiv.org/abs/2605.11086.arxivarxiv.orgExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks?Abstract page for arXiv paper 2605.11086: ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks?https://arxiv.org/abs/2605.11086
4 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle simonwillison. Links to https://simonwillison.net/2026/Jul/22/openai-cyberattack/.simonwillisonsimonwillison.netOpenAI’s accidental cyberattack against Hugging Face is science fiction that happenedThis story is wild. The short version: OpenAI were running a cybersecurity test against an unreleased model, with the model’s guardrail features turned off. Rather than solve the test, the …https://simonwillison.net/2026/Jul/22/openai-cyberattack/Source handle huggingface. Links to https://huggingface.co/blog/security-incident-july-2026.huggingfacehuggingface.coSecurity incident disclosure — July 2026We’re on a journey to advance and democratize artificial intelligence through open source and open science.https://huggingface.co/blog/security-incident-july-2026Source handle arxiv. Links to https://arxiv.org/abs/2605.11086.arxivarxiv.orgExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks?Abstract page for arXiv paper 2605.11086: ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks?https://arxiv.org/abs/2605.11086AI Digest
Cheap Code Makes Harness Design the Real Work
The strongest AI discourse signal today was a shift up the stack: as code generation gets cheaper, serious practitioners are focusing on harness design, verification artifacts, steering files, and runtime constraints rather than raw output alone. Supporting security disclosures from OpenAI and Hu...
ai-agentscoding-agentsharnesses
Source handle theo-t3-gg-theo-on-generating-more-verification. Links to https://www.youtube.com/watch?v=434cG4g5KLE.theo-t3-gg-theo-on-generating-more-verificationyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=434cG4g5KLESource handle ai-engineer-sam-bhagwat-on-the-emerging-harness. Links to https://www.youtube.com/watch?v=8qWIPUia2O8.ai-engineer-sam-bhagwat-on-the-emerging-harnessyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=8qWIPUia2O8Source handle openai. Links to https://openai.com/index/hugging-face-model-evaluation-security-incident/.openaiopenai.comOpenAI - Hugging Face model-evaluation security incidenthttps://openai.com/index/hugging-face-model-evaluation-security-incident/Source handle huggingface. Links to https://huggingface.co/blog/security-incident-july-2026.huggingfacehuggingface.coSecurity incident disclosure — July 2026We’re on a journey to advance and democratize artificial intelligence through open source and open science.https://huggingface.co/blog/security-incident-july-2026Source handle openai-2. Links to https://openai.com/index/safety-alignment-long-horizon-models/.openai-2openai.comOpenAI - Safety alignment for long-horizon modelshttps://openai.com/index/safety-alignment-long-horizon-models/
5 sources
https://huggingface.co/blog/security-incident-july-2026Source handle openai-2. Links to https://openai.com/index/safety-alignment-long-horizon-models/.openai-2openai.comOpenAI - Safety alignment for long-horizon modelshttps://openai.com/index/safety-alignment-long-horizon-models/AI Digest
Coding Agents Hit the Verification Wall
The strongest AI discourse signal today was a practitioner-led shift from code generation to code verification: the hard problem is increasingly proving correctness, security, maintainability, and real enterprise fit rather than producing plausible output. Supporting evidence from controlled fram...
ai-agentscoding-agentsverification
Source handle ai-engineer-tariq-shaukat-on-verification-heavy. Links to https://www.youtube.com/watch?v=VrpEyglYgeU.ai-engineer-tariq-shaukat-on-verification-heavyyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=VrpEyglYgeUSource handle codingharness. Links to https://codingharness.xyz/.codingharnesscodingharness.xyzCoding Harness Evaluations — which coding framework actually performs best?An eval framework that ranks agentic coding frameworks by having each build the same spec, same harness, same model, in isolated sandboxes — graded on an evidence-based rubric.
https://codingharness.xyz/Source handle aisi-gov. Links to https://www.aisi.gov.uk/blog/how-far-behind-the-frontier-are-leading-open-weight-models-on-cyber.aisi-govaisi.gov.ukHow Far Behind the Frontier are Leading Open Weight Models on Cyber? | AISI WorkWe evaluated the cyber capabilities of leading open and closed weight AI models, and found that recent open models GLM-5.2 and DeepSeek V4-Pro perform similarly to frontier closed models released 4 to 7 months before them – a narrower gap than the 6 to 10 months we measured through most of 2025..png)
https://www.aisi.gov.uk/blog/how-far-behind-the-frontier-are-leading-open-weight-models-on-cyber
3 sources
https://codingharness.xyz/Source handle aisi-gov. Links to https://www.aisi.gov.uk/blog/how-far-behind-the-frontier-are-leading-open-weight-models-on-cyber.aisi-govaisi.gov.ukHow Far Behind the Frontier are Leading Open Weight Models on Cyber? | AISI WorkWe evaluated the cyber capabilities of leading open and closed weight AI models, and found that recent open models GLM-5.2 and DeepSeek V4-Pro perform similarly to frontier closed models released 4 to 7 months before them – a narrower gap than the 6 to 10 months we measured through most of 2025.https://www.aisi.gov.uk/blog/how-far-behind-the-frontier-are-leading-open-weight-models-on-cyberAI Digest
Agent Harnesses Are Becoming the Real Product
A delayed-discovery but unusually concrete agent-engineering post made the strongest case of the day: meaningful gains are increasingly coming from harness design, runtime structure, and workflow clarity rather than from longer prompts alone. Supporting sentiment around Codex folding into ChatGPT...
ai-agentsagent-engineeringdeveloper-tools
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain [AI Tinkerers - Post-Training]If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle theo-t3-gg-the-codex-app-is-now-chatgpt. Links to https://www.youtube.com/watch?v=zl_Z5TNJB3U.theo-t3-gg-the-codex-app-is-now-chatgptyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=zl_Z5TNJB3USource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-34-on-device-llms-agent-lab-reports-and-store-automation.post-training-aitinkerers-2post-training.aitinkerers.orgTop AI Demos #34: On-device LLMs, Agent Lab Reports, and Store Automation [AI Tinkerers - Post-Training]This week we're seeing a lot of focus on making agents more reliable and manageable, like with [Golemry: Unattended Agent...https://post-training.aitinkerers.org/p/top-ai-demos-34-on-device-llms-agent-lab-reports-and-store-automation
3 sources
https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle theo-t3-gg-the-codex-app-is-now-chatgpt. Links to https://www.youtube.com/watch?v=zl_Z5TNJB3U.theo-t3-gg-the-codex-app-is-now-chatgptyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=zl_Z5TNJB3USource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-34-on-device-llms-agent-lab-reports-and-store-automation.post-training-aitinkerers-2post-training.aitinkerers.orgTop AI Demos #34: On-device LLMs, Agent Lab Reports, and Store Automation [AI Tinkerers - Post-Training]This week we're seeing a lot of focus on making agents more reliable and manageable, like with [Golemry: Unattended Agent...
https://post-training.aitinkerers.org/p/top-ai-demos-34-on-device-llms-agent-lab-reports-and-store-automationAI Digest
GPT-5.6 Splits AI Into Work Surfaces
OpenAI’s GPT-5.6 launch mattered less as a benchmark event than as a sign that AI products are being reorganized into distinct work surfaces for chat, long-form execution, and coding. Microsoft’s same-day Foundry push and practitioner commentary both reinforced the same theme: the real bottleneck...
agentsopenaiworkflow
Source handle openai. Links to https://openai.com/index/gpt-5-6/.openaiopenai.comGPT-5.6: Frontier intelligence that scales with your ambitionMore intelligence from every token, stronger performance per dollar, and more capability on demand for your hardest work.
https://openai.com/index/gpt-5-6/Source handle help-openai. Links to https://help.openai.com/en/articles/20001275-chatgpt-work-and-codex.help-openaihelp.openai.comOpenAI Help Center - ChatGPT Work and Codexhttps://help.openai.com/en/articles/20001275-chatgpt-work-and-codexSource handle azure-microsoft. Links to https://azure.microsoft.com/en-us/blog/frontier-models-and-production-agents-advancing-microsoft-foundry-for-the-agentic-era/.azure-microsoftazure.microsoft.comFrontier models and production agents: Advancing Microsoft Foundry for the agentic era | Microsoft Azure BlogIntroducing OpenAI's latest frontier model series, the Asia Pacific Data Zone, and product agent capabilities, all generally available in Microsoft Foundry.
https://azure.microsoft.com/en-us/blog/frontier-models-and-production-agents-advancing-microsoft-foundry-for-the-agentic-era/Source handle youtube-nate-b-jones-1-6m-agents-registered-for. Links to https://www.youtube.com/watch?v=PRqiGS6fnIM.youtube-nate-b-jones-1-6m-agents-registered-foryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=PRqiGS6fnIM
4 sourceshttps://openai.com/index/gpt-5-6/Source handle help-openai. Links to https://help.openai.com/en/articles/20001275-chatgpt-work-and-codex.help-openaihelp.openai.comOpenAI Help Center - ChatGPT Work and Codexhttps://help.openai.com/en/articles/20001275-chatgpt-work-and-codexSource handle azure-microsoft. Links to https://azure.microsoft.com/en-us/blog/frontier-models-and-production-agents-advancing-microsoft-foundry-for-the-agentic-era/.azure-microsoftazure.microsoft.comFrontier models and production agents: Advancing Microsoft Foundry for the agentic era | Microsoft Azure BlogIntroducing OpenAI's latest frontier model series, the Asia Pacific Data Zone, and product agent capabilities, all generally available in Microsoft Foundry.
https://azure.microsoft.com/en-us/blog/frontier-models-and-production-agents-advancing-microsoft-foundry-for-the-agentic-era/Source handle youtube-nate-b-jones-1-6m-agents-registered-for. Links to https://www.youtube.com/watch?v=PRqiGS6fnIM.youtube-nate-b-jones-1-6m-agents-registered-foryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=PRqiGS6fnIMAI Digest
Agent Harnesses Beat Prompt Bloat
Today’s strongest AI discourse signal was a builder case that major agent gains are increasingly coming from harness design rather than bigger prompts. The broader practitioner backdrop points the same way: memory, verification, orchestration, and runtime guardrails are becoming the real work of ...
agentsai-engineeringevals
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain [AI Tinkerers - Post-Training]If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-34-on-device-llms-agent-lab-reports-and-store-automation.post-training-aitinkerers-2post-training.aitinkerers.orgTop AI Demos #34: On-device LLMs, Agent Lab Reports, and Store Automation [AI Tinkerers - Post-Training]This week we're seeing a lot of focus on making agents more reliable and manageable, like with [Golemry: Unattended Agent...https://post-training.aitinkerers.org/p/top-ai-demos-34-on-device-llms-agent-lab-reports-and-store-automation
2 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-34-on-device-llms-agent-lab-reports-and-store-automation.post-training-aitinkerers-2post-training.aitinkerers.orgTop AI Demos #34: On-device LLMs, Agent Lab Reports, and Store Automation [AI Tinkerers - Post-Training]This week we're seeing a lot of focus on making agents more reliable and manageable, like with [Golemry: Unattended Agent...https://post-training.aitinkerers.org/p/top-ai-demos-34-on-device-llms-agent-lab-reports-and-store-automationAI Digest
Agent Harness Design Becomes the Real Battleground
Today’s strongest AI discourse suggested that the real gains in agent systems are increasingly coming from harness design, reusable skills, and explicit runtime rules rather than from prompt inflation alone. A delayed-discovery Claude Code tracking controversy underscored the flip side: once agen...
ai-agentsagent-harnessesskills
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain [AI Tinkerers - Post-Training]If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/practical-agent-skills-for-product.departmentofproduct-substackdepartmentofproduct.substack.comPractical Agent Skills for Product Teams🧠 McKinsey-style charts, benchmark competitor analysis, translate specs into backlog items, single page animated HTML presentations. 10 skills you can use at work.
https://departmentofproduct.substack.com/p/practical-agent-skills-for-productSource handle thereallo. Links to https://thereallo.dev/blog/claude-code-prompt-steganography.thereallothereallo.devClaude Code Is Steganographically Marking RequestsI inspected Claude Code for privacy reasons and found hidden system prompt markers based on API base URL and timezone.
https://thereallo.dev/blog/claude-code-prompt-steganographySource handle arstechnica. Links to https://arstechnica.com/tech-policy/2026/07/anthropic-outed-for-claude-tracker-that-secretly-monitored-chinese-users/.arstechnicaarstechnica.comSecret Claude tracker shocks users after Anthropic’s anti-surveillance stanceAnthropic accused of spying on users; engineer says “experiment” is over.
https://arstechnica.com/tech-policy/2026/07/anthropic-outed-for-claude-tracker-that-secretly-monitored-chinese-users/
4 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/practical-agent-skills-for-product.departmentofproduct-substackdepartmentofproduct.substack.comPractical Agent Skills for Product Teams🧠 McKinsey-style charts, benchmark competitor analysis, translate specs into backlog items, single page animated HTML presentations. 10 skills you can use at work.https://departmentofproduct.substack.com/p/practical-agent-skills-for-productSource handle thereallo. Links to https://thereallo.dev/blog/claude-code-prompt-steganography.thereallothereallo.devClaude Code Is Steganographically Marking RequestsI inspected Claude Code for privacy reasons and found hidden system prompt markers based on API base URL and timezone.
https://thereallo.dev/blog/claude-code-prompt-steganographySource handle arstechnica. Links to https://arstechnica.com/tech-policy/2026/07/anthropic-outed-for-claude-tracker-that-secretly-monitored-chinese-users/.arstechnicaarstechnica.comSecret Claude tracker shocks users after Anthropic’s anti-surveillance stanceAnthropic accused of spying on users; engineer says “experiment” is over.
https://arstechnica.com/tech-policy/2026/07/anthropic-outed-for-claude-tracker-that-secretly-monitored-chinese-users/AI Digest
Agent Harnesses Beat Agent Theater
Today’s strongest AI discourse pointed in the same direction: agent gains are increasingly coming from better harnesses, evaluation loops, and document-preparation workflows rather than from autonomy theater. The practical lesson is to invest in scaffolding, normalization, and human review bounda...
ai-agentsevalsworkflow-engineering
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain [AI Tinkerers - Post-Training]If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle simonwillison. Links to https://simonwillison.net/2026/Jul/2/dspy-datasette-agent-prompts/.simonwillisonsimonwillison.netResearch: Using DSPy to evaluate and improve Datasette Agent's SQL system promptsLeveraging the DSPy framework, this project evaluates and refines the core production system prompts used by Datasette Agent’s read-only SQL question answerer. The methodology involves a harness where DSPy agents …https://simonwillison.net/2026/Jul/2/dspy-datasette-agent-prompts/Source handle nate-b-jones-paperwork-focused-agent-workflow-vi. Links to https://www.youtube.com/watch?v=U4TmrlWEY4M.nate-b-jones-paperwork-focused-agent-workflow-viyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=U4TmrlWEY4MSource handle simonwillison-2. Links to https://simonwillison.net/2026/Jul/2/llm-coding-agent/.simonwillison-2simonwillison.netRelease: llm-coding-agent 0.1a0A coding agent built on LLMhttps://simonwillison.net/2026/Jul/2/llm-coding-agent/
4 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle simonwillison. Links to https://simonwillison.net/2026/Jul/2/dspy-datasette-agent-prompts/.simonwillisonsimonwillison.netResearch: Using DSPy to evaluate and improve Datasette Agent's SQL system promptsLeveraging the DSPy framework, this project evaluates and refines the core production system prompts used by Datasette Agent’s read-only SQL question answerer. The methodology involves a harness where DSPy agents …https://simonwillison.net/2026/Jul/2/dspy-datasette-agent-prompts/Source handle nate-b-jones-paperwork-focused-agent-workflow-vi. Links to https://www.youtube.com/watch?v=U4TmrlWEY4M.nate-b-jones-paperwork-focused-agent-workflow-viyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=U4TmrlWEY4MSource handle simonwillison-2. Links to https://simonwillison.net/2026/Jul/2/llm-coding-agent/.simonwillison-2simonwillison.netRelease: llm-coding-agent 0.1a0A coding agent built on LLMhttps://simonwillison.net/2026/Jul/2/llm-coding-agent/
AI Digest
The Agent Bottleneck Is the Harness
Today’s strongest AI discourse converged on a single point: agent progress is shifting away from raw model gains and toward better harnesses, interfaces, and controls. The most useful systems now look less like bigger prompts and more like disciplined workflows that preserve portability, visibili...
agentsharnessesinterfaces
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain [AI Tinkerers - Post-Training]If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle ai-engineer-the-prompt-is-still-a-punch-card-ted. Links to https://www.youtube.com/watch?v=hVJOnuhFmTA.ai-engineer-the-prompt-is-still-a-punch-card-tedyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=hVJOnuhFmTASource handle anthropic. Links to https://www.anthropic.com/news/redeploying-fable-5.anthropicanthropic.comAnthropic - Redeploying Fable 5https://www.anthropic.com/news/redeploying-fable-5Source handle simonwillison. Links to https://simonwillison.net/2026/Jul/2/understand-to-participate/.simonwillisonsimonwillison.netUnderstand to participateI saw Geoffrey Litt speak at AIE yesterday, and one framing he used particularly resonated with me: Understand to participate Geoffrey was talking about the challenge of collaborating with coding …https://simonwillison.net/2026/Jul/2/understand-to-participate/
4 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle ai-engineer-the-prompt-is-still-a-punch-card-ted. Links to https://www.youtube.com/watch?v=hVJOnuhFmTA.ai-engineer-the-prompt-is-still-a-punch-card-tedyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=hVJOnuhFmTASource handle anthropic. Links to https://www.anthropic.com/news/redeploying-fable-5.anthropicanthropic.comAnthropic - Redeploying Fable 5https://www.anthropic.com/news/redeploying-fable-5Source handle simonwillison. Links to https://simonwillison.net/2026/Jul/2/understand-to-participate/.simonwillisonsimonwillison.netUnderstand to participateI saw Geoffrey Litt speak at AIE yesterday, and one framing he used particularly resonated with me: Understand to participate Geoffrey was talking about the challenge of collaborating with coding …https://simonwillison.net/2026/Jul/2/understand-to-participate/AI Digest
Harness Design Overtakes Prompting
AI discourse today centered on a shift from prompt-centric thinking to harness-centric system design. The strongest evidence came from a finance-agent writeup, an open coding-model release, and a robotics project that all treated scaffolds, verification loops, and orchestration as the real source...
agentsharness-designcoding-agents
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain [AI Tinkerers - Post-Training]If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle deep-reinforce. Links to https://deep-reinforce.com/ornith_1_0.html.deep-reinforcedeep-reinforce.comOrnith-1.0: Self-Scaffolding LLMs for Agentic CodingIntroducing Ornith-1.0, a self-improving family of open-source models specially for agentic coding tasks.
https://deep-reinforce.com/ornith_1_0.htmlSource handle research-nvidia. Links to https://research.nvidia.com/labs/gear/enpire/.research-nvidiaresearch.nvidia.comENPIRE: Agentic Robot Policy Self-Improvement in the Real WorldAnonymous ENPIRE project website for agentic robot policy self-improvement in the real world.https://research.nvidia.com/labs/gear/enpire/Source handle dwarkesh. Links to https://www.dwarkesh.com/p/grant-sanderson-2.dwarkeshdwarkesh.comGrant Sanderson – AI and the future of mathMath is where we’ll see superintelligence first. What will it look like?
https://www.dwarkesh.com/p/grant-sanderson-2
4 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle deep-reinforce. Links to https://deep-reinforce.com/ornith_1_0.html.deep-reinforcedeep-reinforce.comOrnith-1.0: Self-Scaffolding LLMs for Agentic CodingIntroducing Ornith-1.0, a self-improving family of open-source models specially for agentic coding tasks.https://deep-reinforce.com/ornith_1_0.htmlSource handle research-nvidia. Links to https://research.nvidia.com/labs/gear/enpire/.research-nvidiaresearch.nvidia.comENPIRE: Agentic Robot Policy Self-Improvement in the Real WorldAnonymous ENPIRE project website for agentic robot policy self-improvement in the real world.https://research.nvidia.com/labs/gear/enpire/Source handle dwarkesh. Links to https://www.dwarkesh.com/p/grant-sanderson-2.dwarkeshdwarkesh.comGrant Sanderson – AI and the future of mathMath is where we’ll see superintelligence first. What will it look like?https://www.dwarkesh.com/p/grant-sanderson-2AI Digest
Agent Harnesses Beat Bigger Prompts
Today's strongest AI discourse signal was a builder shift away from prompt inflation and autonomy theater toward better harness design, explicit constraints, and human-owned review loops. A detailed AI Tinkerers writeup and Jon Udell's framing critique both pointed to the same conclusion: workflo...
agentsai-engineeringworkflow
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain [AI Tinkerers - Post-Training]If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/28/jon-udell/.simonwillisonsimonwillison.netA quote from Jon UdellHuman Agent in the loop I dislike the phrase “human in the loop” because it cedes authority to the machines. Let’s flip the narrative. It’s our loop, we work the …https://simonwillison.net/2026/Jun/28/jon-udell/Source handle blog-jonudell. Links to https://blog.jonudell.net/2026/06/28/doctor-it-hurts-when-agents-create-unreviewable-prs-dont-do-that/.blog-jonudellblog.jonudell.net“Doctor, it hurts when agents create unreviewable PRs.” “Don’t do that.”I recently attended a talk, by an engineer at a large software company, on the topic of unreviewable PRs. The problem? When agents raise PRs with thousands of lines of LLM-written adds/deletes/edits, people can't make sense of them. The solution? Throw more agents at the problem: reviewer agents that scan what coding agents have produced,…
https://blog.jonudell.net/2026/06/28/doctor-it-hurts-when-agents-create-unreviewable-prs-dont-do-that/
3 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/28/jon-udell/.simonwillisonsimonwillison.netA quote from Jon UdellHuman Agent in the loop I dislike the phrase “human in the loop” because it cedes authority to the machines. Let’s flip the narrative. It’s our loop, we work the …https://simonwillison.net/2026/Jun/28/jon-udell/Source handle blog-jonudell. Links to https://blog.jonudell.net/2026/06/28/doctor-it-hurts-when-agents-create-unreviewable-prs-dont-do-that/.blog-jonudellblog.jonudell.net“Doctor, it hurts when agents create unreviewable PRs.” “Don’t do that.”I recently attended a talk, by an engineer at a large software company, on the topic of unreviewable PRs. The problem? When agents raise PRs with thousands of lines of LLM-written adds/deletes/edits, people can't make sense of them. The solution? Throw more agents at the problem: reviewer agents that scan what coding agents have produced,…
https://blog.jonudell.net/2026/06/28/doctor-it-hurts-when-agents-create-unreviewable-prs-dont-do-that/AI Digest
Agent Harnesses Beat Tool Sprawl
Today’s strongest AI discourse signal was that many agent gains are coming from better harness design rather than from model changes alone. A detailed builder writeup on Reef, supported by a market read on rising AI spend, pointed to a broader shift from prompt accumulation to disciplined system ...
agentsai-engineeringharness-design
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain [AI Tinkerers - Post-Training]If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle exponentialview. Links to https://www.exponentialview.co/p/the-state-of-the-ai-economy.exponentialviewexponentialview.co🔮 The state of the AI economyWe've reconstructed the AI economy from the bottom up
https://www.exponentialview.co/p/the-state-of-the-ai-economySource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-32.post-training-aitinkerers-2post-training.aitinkerers.orgTop AI Demos #32: Cloud Waste Agents, MVP Validation, & AI Skill Orchestration [AI Tinkerers - Post-Training]{{NEWSLETTER_HACKATHON_SPOTLIGHT}}https://post-training.aitinkerers.org/p/top-ai-demos-32
3 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle exponentialview. Links to https://www.exponentialview.co/p/the-state-of-the-ai-economy.exponentialviewexponentialview.co🔮 The state of the AI economyWe've reconstructed the AI economy from the bottom uphttps://www.exponentialview.co/p/the-state-of-the-ai-economySource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-32.post-training-aitinkerers-2post-training.aitinkerers.orgTop AI Demos #32: Cloud Waste Agents, MVP Validation, & AI Skill Orchestration [AI Tinkerers - Post-Training]{{NEWSLETTER_HACKATHON_SPOTLIGHT}}https://post-training.aitinkerers.org/p/top-ai-demos-32AI Digest
Agent Harness Design Becomes the Differentiator
Today’s strongest AI discourse signal was that agent performance is increasingly being treated as a harness-design problem, not just a model-selection problem. Builder writeups on structured agent frameworks, local coding agents, and prompt-injection defenses all pointed toward the same conclusio...
agentsai-engineeringharness-design
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain [AI Tinkerers - Post-Training]If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle magazine-sebastianraschka. Links to https://magazine.sebastianraschka.com/p/using-local-coding-agents.magazine-sebastianraschkamagazine.sebastianraschka.comUsing Local Coding AgentsUsing Open-Weight Models in Local Coding Harnesses as an Alternative to Claude Code and Codex Subscriptions
https://magazine.sebastianraschka.com/p/using-local-coding-agentsSource handle dwarkesh. Links to https://www.dwarkesh.com/p/the-next-paradigm.dwarkeshdwarkesh.comThe next big breakthrough will be AIs learning on the jobLabs are throwing away the most valuable data.
https://www.dwarkesh.com/p/the-next-paradigmSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/26/hack-my-ai-assistant/.simonwillisonsimonwillison.netWhat happened after 2,000 people tried to hack my AI assistantFernando Irarrázaval ran a challenge on hackmyclaw.com to see if anyone could leak secrets held by his OpenClaw test instance by sending it email. Surprisingly, after 6,000 attempts (and $500 …https://simonwillison.net/2026/Jun/26/hack-my-ai-assistant/
4 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle magazine-sebastianraschka. Links to https://magazine.sebastianraschka.com/p/using-local-coding-agents.magazine-sebastianraschkamagazine.sebastianraschka.comUsing Local Coding AgentsUsing Open-Weight Models in Local Coding Harnesses as an Alternative to Claude Code and Codex Subscriptionshttps://magazine.sebastianraschka.com/p/using-local-coding-agentsSource handle dwarkesh. Links to https://www.dwarkesh.com/p/the-next-paradigm.dwarkeshdwarkesh.comThe next big breakthrough will be AIs learning on the jobLabs are throwing away the most valuable data.https://www.dwarkesh.com/p/the-next-paradigmSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/26/hack-my-ai-assistant/.simonwillisonsimonwillison.netWhat happened after 2,000 people tried to hack my AI assistantFernando Irarrázaval ran a challenge on hackmyclaw.com to see if anyone could leak secrets held by his OpenClaw test instance by sending it email. Surprisingly, after 6,000 attempts (and $500 …https://simonwillison.net/2026/Jun/26/hack-my-ai-assistant/AI Digest
Measuring the AI Economy
The strongest AI discourse signal today was a shift from model spectacle to market accounting, led by Exponential View's attempt to estimate the generative AI economy at $110 billion in annual sales and a $175 billion run rate. The deeper story is that serious AI conversation is starting to organ...
ai-economymarket-analysisinfrastructure
Source handle exponentialview. Links to https://www.exponentialview.co/p/the-state-of-the-ai-economy.exponentialviewexponentialview.co🔮 The state of the AI economyWe've reconstructed the AI economy from the bottom uphttps://www.exponentialview.co/p/the-state-of-the-ai-economySource handle intelligence-exponentialview. Links to https://intelligence.exponentialview.co/.intelligence-exponentialviewintelligence.exponentialview.coThe State of the AI Economy — Exponential ViewThe State of the AI Economy — a report from Exponential View. Read online or download the full deck.
https://intelligence.exponentialview.co/Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/24/browser-compat-db/.simonwillisonsimonwillison.netsimonw/browser-compat-dbInspired by Mozilla's new MDN MCP service - source code here - I decided to try converting their comprehensive mdn/browser-compat-data repository full of browser compatibility data into a SQLite database. …https://simonwillison.net/2026/Jun/24/browser-compat-db/
3 sourceshttps://www.exponentialview.co/p/the-state-of-the-ai-economySource handle intelligence-exponentialview. Links to https://intelligence.exponentialview.co/.intelligence-exponentialviewintelligence.exponentialview.coThe State of the AI Economy — Exponential ViewThe State of the AI Economy — a report from Exponential View. Read online or download the full deck.
https://intelligence.exponentialview.co/Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/24/browser-compat-db/.simonwillisonsimonwillison.netsimonw/browser-compat-dbInspired by Mozilla's new MDN MCP service - source code here - I decided to try converting their comprehensive mdn/browser-compat-data repository full of browser compatibility data into a SQLite database. …https://simonwillison.net/2026/Jun/24/browser-compat-db/AI Digest
The Agent Harness Becomes the Product
Today's strongest AI discourse signal was a practical argument that reliable agent performance increasingly comes from harness design, not prompt cleverness alone. A second supporting thread in hiring discourse pointed to the same conclusion: as fluent AI output gets easier to produce, structure,...
agentsai-engineeringworkflows
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domain.post-training-aitinkererspost-training.aitinkerers.orgHow to Write a Winning Agent Harness for Your Domain [AI Tinkerers - Post-Training]If your agent improves when you add instructions, then suddenly gets worse, you may not have a prompting problem. You may have a harness problem.https://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-32.post-training-aitinkerers-2post-training.aitinkerers.orgTop AI Demos #32: Cloud Waste Agents, MVP Validation, & AI Skill Orchestration [AI Tinkerers - Post-Training]{{NEWSLETTER_HACKATHON_SPOTLIGHT}}https://post-training.aitinkerers.org/p/top-ai-demos-32Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/24/tom-macwright/.simonwillisonsimonwillison.netA quote from Tom MacWrightIn the last few months, I've started to see [job applications] that were clearly cowritten by an LLM, link to an LLM-generated portfolio site, which then links to LLM-generated GitHub …https://simonwillison.net/2026/Jun/24/tom-macwright/
3 sourceshttps://post-training.aitinkerers.org/p/how-to-write-a-winning-agent-harness-for-your-domainSource handle post-training-aitinkerers-2. Links to https://post-training.aitinkerers.org/p/top-ai-demos-32.post-training-aitinkerers-2post-training.aitinkerers.orgTop AI Demos #32: Cloud Waste Agents, MVP Validation, & AI Skill Orchestration [AI Tinkerers - Post-Training]{{NEWSLETTER_HACKATHON_SPOTLIGHT}}https://post-training.aitinkerers.org/p/top-ai-demos-32Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/24/tom-macwright/.simonwillisonsimonwillison.netA quote from Tom MacWrightIn the last few months, I've started to see [job applications] that were clearly cowritten by an LLM, link to an LLM-generated portfolio site, which then links to LLM-generated GitHub …https://simonwillison.net/2026/Jun/24/tom-macwright/AI Digest
Prompt Injection’s Role-Confusion Turn
New research reframed prompt injection as a deeper authority-parsing failure inside models, not just a jailbreak pattern. That makes today’s agent discourse less about adding more guardrails and more about whether models reliably understand who is allowed to instruct them in the first place.
agentsprompt-injectionsecurity
Source handle role-confusion-github. Links to https://role-confusion.github.io/.role-confusion-githubrole-confusion.github.ioPrompt Injection as Role ConfusionProject page for Prompt Injection as Role Confusion, accepted to ICML 2026. We show prompt injections are driven by a flaw in how LLMs perceive roles. This lets us create new attacks, explain mech interp results, and predict when attacks succeed. We then discuss what roles are and why they matter, and share research ideas for a science of roles.
https://role-confusion.github.io/Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/22/prompt-injection-as-role-confusion/.simonwillisonsimonwillison.netPrompt Injection as Role ConfusionFirst, I absolutely love this: This is a blog-style writeup of the paper. I wish every paper would come with one of these. Academic writing is pretty dry - the …https://simonwillison.net/2026/Jun/22/prompt-injection-as-role-confusion/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Jun/22/porting-moebius/.simonwillison-2simonwillison.netPorting the Moebius 0.2B image inpainting model to run in the browser with Claude CodeThis morning on Hacker News I saw Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance, describing a small but effective inpainting model—a model where you can mark regions of …
https://simonwillison.net/2026/Jun/22/porting-moebius/Source handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/the-god-we-deserve-nonzero-s-robert-wright-on-ai-as-humanity-s-ultimate-test/.cognitiverevolutioncognitiverevolution.aiThe God We Deserve: Nonzero's Robert Wright on AI as Humanity's Ultimate TestRobert Wright discusses The God Test, arguing that AI development is shaped by evolutionary and market selection pressures that may reward deception, and that passing the challenge requires better alignment, governance, and global coordination.
https://www.cognitiverevolution.ai/the-god-we-deserve-nonzero-s-robert-wright-on-ai-as-humanity-s-ultimate-test/
4 sources
https://role-confusion.github.io/Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/22/prompt-injection-as-role-confusion/.simonwillisonsimonwillison.netPrompt Injection as Role ConfusionFirst, I absolutely love this: This is a blog-style writeup of the paper. I wish every paper would come with one of these. Academic writing is pretty dry - the …https://simonwillison.net/2026/Jun/22/prompt-injection-as-role-confusion/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Jun/22/porting-moebius/.simonwillison-2simonwillison.netPorting the Moebius 0.2B image inpainting model to run in the browser with Claude CodeThis morning on Hacker News I saw Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance, describing a small but effective inpainting model—a model where you can mark regions of …https://simonwillison.net/2026/Jun/22/porting-moebius/Source handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/the-god-we-deserve-nonzero-s-robert-wright-on-ai-as-humanity-s-ultimate-test/.cognitiverevolutioncognitiverevolution.aiThe God We Deserve: Nonzero's Robert Wright on AI as Humanity's Ultimate TestRobert Wright discusses The God Test, arguing that AI development is shaped by evolutionary and market selection pressures that may reward deception, and that passing the challenge requires better alignment, governance, and global coordination.https://www.cognitiverevolution.ai/the-god-we-deserve-nonzero-s-robert-wright-on-ai-as-humanity-s-ultimate-test/AI Digest
The Missing Operating Layer for Agents
Today’s strongest AI discourse was not about a new model milestone but about the infrastructure teams need around agents: ownership, permissions, trustworthy execution, and workflow-specific evaluation. Across developer commentary and community demos, the practical consensus is that progress now ...
agentsdeveloper-toolsevaluation
Source handle youtube-i-don-t-have-time-to-build-these-things. Links to https://www.youtube.com/watch?v=wEAb0x3wTRc.youtube-i-don-t-have-time-to-build-these-thingsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=wEAb0x3wTRcSource handle youtube-most-teams-skip-this-critical-ai-agent-s. Links to https://www.youtube.com/watch?v=rh_PcL26zls.youtube-most-teams-skip-this-critical-ai-agent-syoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=rh_PcL26zlsSource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/top-ai-demos-31.post-training-aitinkererspost-training.aitinkerers.orgAI Tinkerers / Post-Training - Top AI demos #31https://post-training.aitinkerers.org/p/top-ai-demos-31Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/21/temporary-cloudflare-accounts/.simonwillisonsimonwillison.netTemporary Cloudflare Accounts for AI agentsThe announcement says this is "for AI agents" but (as is pretty common these days) the AI hook isn't really necessary, this is an interesting feature for everyone else as …https://simonwillison.net/2026/Jun/21/temporary-cloudflare-accounts/
4 sources
AI Digest
Synthetic Media Needs a Trust Stack
Synthetic media discourse shifted from detection to accountability: the useful question is where AI entered the production chain, who controlled it, and who remains responsible. A delayed Cognitive Revolution episode pointed to the same broader convergence of model progress, safety, policy, and d...
synthetic-mediacreator-trustai-accountability
Source handle nate-b-jones-you-can-t-tell-if-i-m-real-anymore. Links to https://www.youtube.com/watch?v=lWbtvC0Hn18.nate-b-jones-you-can-t-tell-if-i-m-real-anymoreyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=lWbtvC0Hn18Source handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/ai-am-3-zvi-on-fable-the-cases-for-against-the-ban-ai-for-math-logistics-more/.cognitiverevolutioncognitiverevolution.aiAI:AM #3: Zvi on Fable, the Cases For & Against the Ban, + AI for Math, Logistics & MoreZvi Mowshowitz joins AI:AM to examine Anthropic's Fable system card, including math gains, deceptive behavior, decision theory, and interpretability. The episode also weighs the US export-control order and surveys AI work in medicine, math, logistics, and software.
https://www.cognitiverevolution.ai/ai-am-3-zvi-on-fable-the-cases-for-against-the-ban-ai-for-math-logistics-more/
2 sourceshttps://www.cognitiverevolution.ai/ai-am-3-zvi-on-fable-the-cases-for-against-the-ban-ai-for-math-logistics-more/AI Digest
Going Bigger With AI-Assisted Software
The day’s strongest signal was a builder argument that AI-assisted development changes product scope: when implementation costs fall, broader integrated tools become worth attempting. The counterweight is that wider agent reach makes boundaries around deployment, auth, permissions, and policy mor...
ai-agentsdeveloper-toolsproduct-strategy
Source handle theo-t3-gg-it-s-time-to-go-bigger. Links to https://www.youtube.com/watch?v=WBT-z_-OPhw.theo-t3-gg-it-s-time-to-go-biggeryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=WBT-z_-OPhwSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/19/sean-lynch/.simonwillisonsimonwillison.netA quote from Sean LynchThe real valuable capability MCP offers over skills/CLI is isolating the auth flow outside of the agent’s context window, and potentially out of the harness completely. [...] Maybe the idealized …https://simonwillison.net/2026/Jun/19/sean-lynch/Source handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/dean-ball-on-joining-openai-new-power-centers-frontier-ai-policy-main-character-energy/.cognitiverevolutioncognitiverevolution.aiDean Ball, on Joining OpenAI: New Power Centers, Frontier AI Policy, & Main Character EnergyDean Ball discusses joining OpenAI to build a frontier AI policy team, reflects on drafting the US AI Action Plan, and assesses export controls, government AI use, state regulation, and emerging capabilities in coding, cyber, and robotics.
https://www.cognitiverevolution.ai/dean-ball-on-joining-openai-new-power-centers-frontier-ai-policy-main-character-energy/
3 sourceshttps://www.cognitiverevolution.ai/dean-ball-on-joining-openai-new-power-centers-frontier-ai-policy-main-character-energy/AI Digest
Agent Work Moves From Prompts to Procedures
Practitioner discourse is converging on a new operating model for coding agents: durable gains come from loops, review surfaces, and reusable procedures rather than longer one-off prompts.
ai-agentscoding-agentsdeveloper-workflows
Source handle theo-t3-gg-i-guess-we-re-writing-loops-now. Links to https://www.youtube.com/watch?v=iJVJwmCKW9o.theo-t3-gg-i-guess-we-re-writing-loops-nowyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=iJVJwmCKW9oSource handle nate-b-jones-your-ai-skills-are-trapped-here-s-h. Links to https://www.youtube.com/watch?v=9PUaEj0pMYE.nate-b-jones-your-ai-skills-are-trapped-here-s-hyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=9PUaEj0pMYESource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/codexs-record-and-replay-lets-you.departmentofproduct-substackdepartmentofproduct.substack.com🔵 Codex's Record and Replay lets you record a workflow and re-use itPlus: Claude Design's new capabilities, Square's coding agent shipped an entire backlog in a few days, show off your actual skills with LinkedIn's connected apps
https://departmentofproduct.substack.com/p/codexs-record-and-replay-lets-you
3 sourceshttps://departmentofproduct.substack.com/p/codexs-record-and-replay-lets-youAI Digest
Production Agents Are Becoming an Operations Problem
Today’s strongest AI discourse signal was the shift from model choice to production accountability: evaluation, tracing, governance, and repairability now define whether agents are ready for real deployment. Short builder clips on prompt loops reinforced the same theme at workflow scale.
ai-agentsevaluationobservability
Source handle ai-engineer-the-production-ai-playbook-deploying. Links to https://www.youtube.com/watch?v=ObTPqBGsEbA.ai-engineer-the-production-ai-playbook-deployingyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=ObTPqBGsEbASource handle theo-t3-gg-prompt-loops-not-individual-instructi. Links to https://www.youtube.com/shorts/z1I9MUjUrVM.theo-t3-gg-prompt-loops-not-individual-instructiyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/z1I9MUjUrVMSource handle greg-isenberg-agentic-loops-the-future-of-prompt. Links to https://www.youtube.com/shorts/ytO3bJWofmw.greg-isenberg-agentic-loops-the-future-of-promptyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/ytO3bJWofmw
3 sources
AI Digest
Disposable Code, Rising Discipline: AI Is Shifting Engineering Work From Writing to Governing
AI coding is increasingly judged by governance and maintainability rather than raw output speed. Charity Majors’ latest commentary suggests teams should treat generated code as disposable by default and invest in process-level quality controls that keep AI throughput reliable at scale.
ai-codingsoftware-engineeringgovernance
Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/17/charity-majors/.simonwillisonsimonwillison.netA quote from Charity MajorsWhat happened in 2025 was this: the economics of code production were turned upside down. Instead of being very hard, time-consuming, and expensive to generate code, it became effectively free …https://simonwillison.net/2026/Jun/17/charity-majors/Source handle wes-roth-here-s-really-why-fable-5-got-banned. Links to https://www.youtube.com/watch?v=XtgEA4L7sIU.wes-roth-here-s-really-why-fable-5-got-bannedyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=XtgEA4L7sIU
2 sources
AI Digest
Policy Friction Is Becoming the AI Work Surface
This digest shows the AI discourse turning from capability questions into governance questions: Fable/Mythos safety controls, prompt-framing behavior, and access restrictions are now central to how reliable and usable frontier models are. The practical consequence is that teams should treat model...
frontier-modelssafety-policygovernance
Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/16/matteo-wong-the-atlantic/.simonwillisonsimonwillison.netA quote from Matteo Wong, The AtlanticKatie Moussouris, a cybersecurity expert and the CEO of Luta Security, told me that Anthropic shared with her a copy of the White House’s report on the Fable jailbreak to …https://simonwillison.net/2026/Jun/16/matteo-wong-the-atlantic/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Jun/16/fable-5-export-controls/.simonwillison-2simonwillison.netThe Fable 5 Export Controls Harm US Cyber DefenseI quoted The Atlantic quoting Kate Moussouris earlier, when I should have gone straight to the source. Here she is confirming that the "jailbreak" that got Claude Fable 5 banned …https://simonwillison.net/2026/Jun/16/fable-5-export-controls/Source handle theo-t3-gg-fable-5-vs-mythos-5-the-basics. Links to https://www.youtube.com/shorts/9ut0E62jJ1o.theo-t3-gg-fable-5-vs-mythos-5-the-basicsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/9ut0E62jJ1oSource handle theo-t3-gg-anthropic-s-horrible-new-restrictions. Links to https://www.youtube.com/shorts/QlGKd6eiWyU.theo-t3-gg-anthropic-s-horrible-new-restrictionsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/QlGKd6eiWyUSource handle ai-engineer-you-might-not-need-50-diffusion-step. Links to https://www.youtube.com/watch?v=gHs5ZiY80PM.ai-engineer-you-might-not-need-50-diffusion-stepyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=gHs5ZiY80PM
5 sources
AI Digest
AI Agents Hit the Delivery Bottleneck
The day’s strongest AI discourse argued that coding agents are compressing implementation work without eliminating the human bottlenecks around deciding what to build, verifying results, and carrying accountability. The practical implication is to measure agent impact across the whole delivery lo...
coding-agentssoftware-engineeringlabor
Source handle normaltech. Links to https://www.normaltech.ai/p/why-ai-hasnt-replaced-software-engineers.normaltechnormaltech.aiWhy AI hasn’t replaced software engineers, and won’tCoding agents as normal technology
https://www.normaltech.ai/p/why-ai-hasnt-replaced-software-engineersSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/14/why-ai-hasnt-replaced-software-engineers/.simonwillisonsimonwillison.netWhy AI hasn’t replaced software engineers, and won’tArvind Narayanan and Sayash Kappor take on the question of AI job losses through the lens of a profession that is uniquely suited to AI disruption - software engineering. In …https://simonwillison.net/2026/Jun/14/why-ai-hasnt-replaced-software-engineers/Source handle jack-clark. Links to https://jack-clark.net/2026/06/15/import-ai-461-alignment-is-not-on-track-frontiercode-and-synthetic-research-interns/.jack-clarkjack-clark.netImport AI 461: “Alignment is not on track”; FrontierCode; and synthetic research internsWelcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe now AI researchers launch new safety startup because “alignment is not on track”:…Sequent will have a portfolio of under-resourced research bets…Researchers from the UK AI Security Institute…
https://jack-clark.net/2026/06/15/import-ai-461-alignment-is-not-on-track-frontiercode-and-synthetic-research-interns/
3 sources
https://www.normaltech.ai/p/why-ai-hasnt-replaced-software-engineersSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/14/why-ai-hasnt-replaced-software-engineers/.simonwillisonsimonwillison.netWhy AI hasn’t replaced software engineers, and won’tArvind Narayanan and Sayash Kappor take on the question of AI job losses through the lens of a profession that is uniquely suited to AI disruption - software engineering. In …https://simonwillison.net/2026/Jun/14/why-ai-hasnt-replaced-software-engineers/Source handle jack-clark. Links to https://jack-clark.net/2026/06/15/import-ai-461-alignment-is-not-on-track-frontiercode-and-synthetic-research-interns/.jack-clarkjack-clark.netImport AI 461: “Alignment is not on track”; FrontierCode; and synthetic research internsWelcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe now AI researchers launch new safety startup because “alignment is not on track”:…Sequent will have a portfolio of under-resourced research bets…Researchers from the UK AI Security Institute…https://jack-clark.net/2026/06/15/import-ai-461-alignment-is-not-on-track-frontiercode-and-synthetic-research-interns/AI Digest
The Harness Layer Becomes the Real AI Business
Today’s strongest AI discourse shifted from raw model capability to ownership of the workflow layer around models. Nate B. Jones argued that frontier-lab value accrues in proprietary harnesses, while Greg Isenberg’s local-model advice framed the same layer as operational resilience.
ai-agentsworkflowfrontier-models
Source handle nate-b-jones-the-real-reason-for-the-openai-ipo. Links to https://www.youtube.com/watch?v=7RDK84LLL2U.nate-b-jones-the-real-reason-for-the-openai-ipoyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=7RDK84LLL2USource handle greg-isenberg-claude-fable-5-is-banned-what-to-d. Links to https://www.youtube.com/watch?v=bdhUBBACglw.greg-isenberg-claude-fable-5-is-banned-what-to-dyoutube.comClaude Fable 5 is BANNED. What to do?In this solo episode, I walk through the implications of the ban of Claude Fable 5 — the most powerful model on the planet and the one I planned to build wit...
https://www.youtube.com/watch?v=bdhUBBACglw
2 sources
https://www.youtube.com/watch?v=bdhUBBACglwAI Digest
Fable and Mythos Turn Model Access Into a Policy Dependency
Anthropic's Fable 5 and Mythos 5 access suspension reframed frontier models as policy-dependent infrastructure, not just software services. The practical lesson is continuity planning: teams should map single-provider dependencies and keep fallback workflows ready.
frontier-modelsmodel-accessai-policy
Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/13/us-government-directive-to-suspend-access/.simonwillisonsimonwillison.netStatement on the US government directive to suspend access to Fable 5 and Mythos 5Well this is nuts: The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign …https://simonwillison.net/2026/Jun/13/us-government-directive-to-suspend-access/Source handle nate-b-jones-fable-mythos-access-suspension-comm. Links to https://www.youtube.com/watch?v=b3jlsjOIOzs.nate-b-jones-fable-mythos-access-suspension-commyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=b3jlsjOIOzsSource handle theo-t3-gg-fable-mythos-directive-reaction. Links to https://www.youtube.com/watch?v=W-kM56mf-Nw.theo-t3-gg-fable-mythos-directive-reactionyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=W-kM56mf-NwSource handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/ai-in-the-am-week-2-highlights-june-2026/.cognitiverevolutioncognitiverevolution.aiAI in the AM — Week 2 Highlights (June 2026)Week 2 highlights examine Anthropic’s Fable launch in real workflows, from safety gates and API refusals to coding, 3D worlds, and a Claude-run Twitter test. The episode also covers alignment theory, legal reasoning benchmarks, interpretability, token economics, and power concentration.
https://www.cognitiverevolution.ai/ai-in-the-am-week-2-highlights-june-2026/
4 sourceshttps://www.cognitiverevolution.ai/ai-in-the-am-week-2-highlights-june-2026/AI Digest
Codex as Computer Delegation
The strongest signal was a practical reframing of Codex: not just a coding assistant, but a supervised computer operator for bounded, inspectable jobs. The operator skill is shifting toward goals, sources, standards, permission boundaries, and proof of completion.
agentscodexcomputer-use
Read digest 
AI Digest
Fable 5 Makes Agent Work a Verification Problem
Claude Fable/Mythos reactions pointed less to raw benchmark excitement than to a new operating problem: stronger agents need clearer proof, constraints, and governance. The day’s builder evidence reinforced that agent progress now depends on workflow design, evals, and disciplined tool use.
claude-fable-5agentscoding-agents
Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/9/claude-fable-5/.simonwillisonsimonwillison.netInitial impressions of Claude Fable 5I didn’t have early access to today’s Claude Fable 5 release, but I’ve spent the past ~5.5 hours putting it through its paces. My initial impressions are that this is …
https://simonwillison.net/2026/Jun/9/claude-fable-5/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Jun/9/andrej-karpathy/.simonwillison-2simonwillison.netA quote from Andrej KarpathyI feel a lot of things changing as working software increasingly comes out on a tap. The Jevon's paradox kicks in and I feel my own demand for software growing …https://simonwillison.net/2026/Jun/9/andrej-karpathy/Source handle nitter. Links to https://nitter.net/bcherny/status/2064431111154053187.nitternitter.netBoris Cherny - Fable 5 reactionhttps://nitter.net/bcherny/status/2064431111154053187Source handle simonwillison-3. Links to https://simonwillison.net/2026/Jun/10/if-claude-fable-stops-helping-you/.simonwillison-3simonwillison.netIf Claude Fable stops helping you, you’ll never knowJonathon Ready highlights one of the more eyebrow-raising details from the 319 page system card for Fable 5 and Mythos 5. Here's a longer excerpt, highlights mine: In light of …https://simonwillison.net/2026/Jun/10/if-claude-fable-stops-helping-you/Source handle simonwillison-4. Links to https://simonwillison.net/2026/Jun/10/jeremy-howard/.simonwillison-4simonwillison.netA quote from Jeremy HowardEasy solution to slow down recursive AI self improvement: The lab with the top-ranked model must agree THEY must not use it for working on frontier AI But everyone else …https://simonwillison.net/2026/Jun/10/jeremy-howard/Source handle nate-b-jones-stop-coding-start-steering-claude-v. Links to https://www.youtube.com/watch?v=R2-Y1Hjwx2U.nate-b-jones-stop-coding-start-steering-claude-vyoutube.comNate B Jones - Stop Coding. Start Steering. Claude vs Codexhttps://www.youtube.com/watch?v=R2-Y1Hjwx2USource handle ai-engineer-self-driving-products-product-signal. Links to https://www.youtube.com/watch?v=zMiSRliEzv4.ai-engineer-self-driving-products-product-signalyoutube.comAI Engineer - Self Driving Products: Product Signals to Pull Requests — Joshua Snyder, PostHoghttps://www.youtube.com/watch?v=zMiSRliEzv4Source handle ai-engineer-stop-making-models-bigger-make-them. Links to https://www.youtube.com/watch?v=TNwJ1LMiENk.ai-engineer-stop-making-models-bigger-make-themyoutube.comAI Engineer - Stop Making Models Bigger, Make Them Behave — Kobie Crawdord, Snorkelhttps://www.youtube.com/watch?v=TNwJ1LMiENk
8 sourceshttps://simonwillison.net/2026/Jun/9/claude-fable-5/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Jun/9/andrej-karpathy/.simonwillison-2simonwillison.netA quote from Andrej KarpathyI feel a lot of things changing as working software increasingly comes out on a tap. The Jevon's paradox kicks in and I feel my own demand for software growing …https://simonwillison.net/2026/Jun/9/andrej-karpathy/Source handle nitter. Links to https://nitter.net/bcherny/status/2064431111154053187.nitternitter.netBoris Cherny - Fable 5 reactionhttps://nitter.net/bcherny/status/2064431111154053187Source handle simonwillison-3. Links to https://simonwillison.net/2026/Jun/10/if-claude-fable-stops-helping-you/.simonwillison-3simonwillison.netIf Claude Fable stops helping you, you’ll never knowJonathon Ready highlights one of the more eyebrow-raising details from the 319 page system card for Fable 5 and Mythos 5. Here's a longer excerpt, highlights mine: In light of …https://simonwillison.net/2026/Jun/10/if-claude-fable-stops-helping-you/Source handle simonwillison-4. Links to https://simonwillison.net/2026/Jun/10/jeremy-howard/.simonwillison-4simonwillison.netA quote from Jeremy HowardEasy solution to slow down recursive AI self improvement: The lab with the top-ranked model must agree THEY must not use it for working on frontier AI But everyone else …https://simonwillison.net/2026/Jun/10/jeremy-howard/Source handle nate-b-jones-stop-coding-start-steering-claude-v. Links to https://www.youtube.com/watch?v=R2-Y1Hjwx2U.nate-b-jones-stop-coding-start-steering-claude-vyoutube.comNate B Jones - Stop Coding. Start Steering. Claude vs Codexhttps://www.youtube.com/watch?v=R2-Y1Hjwx2USource handle ai-engineer-self-driving-products-product-signal. Links to https://www.youtube.com/watch?v=zMiSRliEzv4.ai-engineer-self-driving-products-product-signalyoutube.comAI Engineer - Self Driving Products: Product Signals to Pull Requests — Joshua Snyder, PostHoghttps://www.youtube.com/watch?v=zMiSRliEzv4Source handle ai-engineer-stop-making-models-bigger-make-them. Links to https://www.youtube.com/watch?v=TNwJ1LMiENk.ai-engineer-stop-making-models-bigger-make-themyoutube.comAI Engineer - Stop Making Models Bigger, Make Them Behave — Kobie Crawdord, Snorkelhttps://www.youtube.com/watch?v=TNwJ1LMiENkAI Digest
AI’s Hidden Bottlenecks
The day’s strongest AI discourse centered on the hidden constraints behind visible capabilities: data, compute, product trust, and agent context management. The clearest signal was Dwarkesh Patel’s argument that frontier systems remain dramatically less sample-efficient than humans.
sample-efficiencyagentscompute
Source handle dwarkesh. Links to https://www.dwarkesh.com/p/the-sample-efficiency-black-hole.dwarkeshdwarkesh.comThe sample efficiency black hole"We see these AIs as a galaxy glittering with capabilities, but at their center, invisible to the naked eye, holding all the constellations together, is an unimaginably massive black hole of data."
https://www.dwarkesh.com/p/the-sample-efficiency-black-holeSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/8/wwdc/.simonwillisonsimonwillison.netSiri AI at WWDC 2026Given how badly burned anyone who took Apple's 2024 WWDC Apple Intelligence announcements at face value was, I'm holding to a strict "I'll believe it when I see it" policy …https://simonwillison.net/2026/Jun/8/wwdc/Source handle nitter. Links to https://nitter.net/bcherny/status/2064327225504403752.nitternitter.netBoris Cherny - Nested subagent support in Claude Codehttps://nitter.net/bcherny/status/2064327225504403752Source handle theo-t3-gg-elon-won-after-all. Links to https://www.youtube.com/watch?v=jB2iKoBSPyo.theo-t3-gg-elon-won-after-allyoutube.comElon won after allThe compute crunch has gotten so bad, that it turns out buying way too many GPUs a couple years ago was a great plan...Thank you Wispr Flow for sponsoring! C...
https://www.youtube.com/watch?v=jB2iKoBSPyo
4 sourceshttps://www.dwarkesh.com/p/the-sample-efficiency-black-holeSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/8/wwdc/.simonwillisonsimonwillison.netSiri AI at WWDC 2026Given how badly burned anyone who took Apple's 2024 WWDC Apple Intelligence announcements at face value was, I'm holding to a strict "I'll believe it when I see it" policy …https://simonwillison.net/2026/Jun/8/wwdc/Source handle nitter. Links to https://nitter.net/bcherny/status/2064327225504403752.nitternitter.netBoris Cherny - Nested subagent support in Claude Codehttps://nitter.net/bcherny/status/2064327225504403752Source handle theo-t3-gg-elon-won-after-all. Links to https://www.youtube.com/watch?v=jB2iKoBSPyo.theo-t3-gg-elon-won-after-allyoutube.comElon won after allThe compute crunch has gotten so bad, that it turns out buying way too many GPUs a couple years ago was a great plan...Thank you Wispr Flow for sponsoring! C...
https://www.youtube.com/watch?v=jB2iKoBSPyoAI Digest
Agents Need Architecture, Not Just Bigger Context
The day’s strongest AI-discourse signal was a move from model capability claims toward the architecture around agents: context curation, state, gates, sandboxes, evidence, and measurement. Anthropic’s recursive-improvement claims supplied the backdrop, but practitioner talks made the case that us...
agentscontext-engineeringai-engineering
Source handle anthropic. Links to https://www.anthropic.com/institute/recursive-self-improvement.anthropicanthropic.comAnthropic Institute - When AI builds itselfhttps://www.anthropic.com/institute/recursive-self-improvementSource handle theo-t3-gg-i-didn-t-expect-this-from-anthropic. Links to https://www.youtube.com/watch?v=xjucOlb_mFM.theo-t3-gg-i-didn-t-expect-this-from-anthropicyoutube.comI didn’t expect this from AnthropicAnthropic wants to pause AI development and thinks that recursive self improvement is a real possibility...Thank you Depot for sponsoring! Check them out at:...
https://www.youtube.com/watch?v=xjucOlb_mFMSource handle jack-clark. Links to https://jack-clark.net/2026/06/08/import-ai-460-reward-hacking-society-rsi-data-from-anthropic-and-rl-based-quadcopter-racing/.jack-clarkjack-clark.netImport AI 460: Reward hacking society, RSI data from Anthropic; and RL-based quadcopter racingWelcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe now Society can be reward-hacked, just like cyber environments:…Imagine an army of credit card point optimizers gaming the system… forever…Research from Kings College London, Fudan University, and…
https://jack-clark.net/2026/06/08/import-ai-460-reward-hacking-society-rsi-data-from-anthropic-and-rl-based-quadcopter-racing/Source handle ai-engineer-why-more-context-makes-your-agent-du. Links to https://www.youtube.com/watch?v=EcqMYoIV57A.ai-engineer-why-more-context-makes-your-agent-duyoutube.comAI Engineer - Why More Context Makes Your Agent Dumber and What to Do About It — Nupur Sharma, Qodohttps://www.youtube.com/watch?v=EcqMYoIV57ASource handle nate-b-jones-fix-your-ai-pipeline-or-lose-your-b. Links to https://www.youtube.com/shorts/76ovBK3lJ2U.nate-b-jones-fix-your-ai-pipeline-or-lose-your-byoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/76ovBK3lJ2USource handle ai-engineer-why-eval-is-the-next-great-compute-p. Links to https://www.youtube.com/watch?v=SKDJo2CopRs.ai-engineer-why-eval-is-the-next-great-compute-pyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=SKDJo2CopRsSource handle ai-engineer-road-to-5-million-tokens-breaking-ba. Links to https://www.youtube.com/watch?v=TUnPNY4E2fw.ai-engineer-road-to-5-million-tokens-breaking-bayoutube.comRoad to 5 Million Tokens: Breaking Barriers in Long Context Training — Max Ryabinin, Together AITraining a standard LLaMA 3B model with a 3 million token context on a single 8xH100 node fails before you even start: the model parameters alone exhaust GPU...
https://www.youtube.com/watch?v=TUnPNY4E2fwSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/new-agentic-payment-abilities-and.departmentofproduct-substackdepartmentofproduct.substack.comNew Agentic Payment Abilities and Features ExploredThe 5 layers of Agentic Payments in 2026; what product teams need to know. Examples from Stripe, Adyen, Coinbase, Mastercard, and more.
https://departmentofproduct.substack.com/p/new-agentic-payment-abilities-and
8 sourceshttps://www.youtube.com/watch?v=xjucOlb_mFMSource handle jack-clark. Links to https://jack-clark.net/2026/06/08/import-ai-460-reward-hacking-society-rsi-data-from-anthropic-and-rl-based-quadcopter-racing/.jack-clarkjack-clark.netImport AI 460: Reward hacking society, RSI data from Anthropic; and RL-based quadcopter racingWelcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe now Society can be reward-hacked, just like cyber environments:…Imagine an army of credit card point optimizers gaming the system… forever…Research from Kings College London, Fudan University, and…https://jack-clark.net/2026/06/08/import-ai-460-reward-hacking-society-rsi-data-from-anthropic-and-rl-based-quadcopter-racing/Source handle ai-engineer-why-more-context-makes-your-agent-du. Links to https://www.youtube.com/watch?v=EcqMYoIV57A.ai-engineer-why-more-context-makes-your-agent-duyoutube.comAI Engineer - Why More Context Makes Your Agent Dumber and What to Do About It — Nupur Sharma, Qodohttps://www.youtube.com/watch?v=EcqMYoIV57ASource handle nate-b-jones-fix-your-ai-pipeline-or-lose-your-b. Links to https://www.youtube.com/shorts/76ovBK3lJ2U.nate-b-jones-fix-your-ai-pipeline-or-lose-your-byoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/76ovBK3lJ2USource handle ai-engineer-why-eval-is-the-next-great-compute-p. Links to https://www.youtube.com/watch?v=SKDJo2CopRs.ai-engineer-why-eval-is-the-next-great-compute-pyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=SKDJo2CopRsSource handle ai-engineer-road-to-5-million-tokens-breaking-ba. Links to https://www.youtube.com/watch?v=TUnPNY4E2fw.ai-engineer-road-to-5-million-tokens-breaking-bayoutube.comRoad to 5 Million Tokens: Breaking Barriers in Long Context Training — Max Ryabinin, Together AITraining a standard LLaMA 3B model with a 3 million token context on a single 8xH100 node fails before you even start: the model parameters alone exhaust GPU...https://www.youtube.com/watch?v=TUnPNY4E2fwSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/new-agentic-payment-abilities-and.departmentofproduct-substackdepartmentofproduct.substack.comNew Agentic Payment Abilities and Features ExploredThe 5 layers of Agentic Payments in 2026; what product teams need to know. Examples from Stripe, Adyen, Coinbase, Mastercard, and more.https://departmentofproduct.substack.com/p/new-agentic-payment-abilities-andAI Digest
Agent Safety Is Becoming Infrastructure
Today’s strongest AI discourse shifted from raw agent capability to the infrastructure needed to constrain it: diagnostic evals, scoped payments, sandboxes, and egress controls. The practical canon is becoming clear: useful agents need bounded authority, observable failures, and reusable workflow...
agentsevalsagent-safety
Source handle ai-engineer-evals-are-broken-use-them-anyway-ara. Links to https://www.youtube.com/watch?v=QuuIywMG4s8.ai-engineer-evals-are-broken-use-them-anyway-arayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=QuuIywMG4s8Source handle ai-engineer-building-safe-payment-infrastructure. Links to https://www.youtube.com/watch?v=KLSuFPj2ld0.ai-engineer-building-safe-payment-infrastructureyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=KLSuFPj2ld0Source handle simonwillison. Links to https://simonwillison.net/2026/Jun/5/openai-help-lockdown-mode/.simonwillisonsimonwillison.netOpenAI Help: Lockdown ModeOpenAI first teased this in February, but now it's live and "rolling out to eligible personal accounts, including Free, Go, Plus, and Pro, and self-serve ChatGPT Business accounts": Lockdown Mode …https://simonwillison.net/2026/Jun/5/openai-help-lockdown-mode/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Jun/6/micropython-in-a-sandbox/.simonwillison-2simonwillison.netRunning Python code in a sandbox with MicroPython and WASMI’ve been experimenting with different approaches to running code in a sandbox for several years now, but my latest attempt feels like it might finally have all of the characteristics …
https://simonwillison.net/2026/Jun/6/micropython-in-a-sandbox/Source handle magazine-sebastianraschka. Links to https://magazine.sebastianraschka.com/p/llm-research-papers-2026-part1.magazine-sebastianraschkamagazine.sebastianraschka.comLLM Research Papers: The 2026 List (January to May)A January-May 2026 list of notable LLM research papers, covering new models, training methods, agents, reasoning, and efficiency improvements.
https://magazine.sebastianraschka.com/p/llm-research-papers-2026-part1
5 sources
https://simonwillison.net/2026/Jun/6/micropython-in-a-sandbox/Source handle magazine-sebastianraschka. Links to https://magazine.sebastianraschka.com/p/llm-research-papers-2026-part1.magazine-sebastianraschkamagazine.sebastianraschka.comLLM Research Papers: The 2026 List (January to May)A January-May 2026 list of notable LLM research papers, covering new models, training methods, agents, reasoning, and efficiency improvements.https://magazine.sebastianraschka.com/p/llm-research-papers-2026-part1AI Digest
AI Work Moves From Output to Instrumentation
Today’s strongest AI discourse argued that useful AI systems need metadata, measurement, constraints, and accountability around their outputs. Voice AI, token dashboards, UI sandboxing, and open-source contribution rules all pointed toward the same operational shift.
ai-agentsvoice-aiinstrumentation
Source handle ai-engineer-beyond-transcription-building-voice. Links to https://www.youtube.com/watch?v=mFLlVpnGpds.ai-engineer-beyond-transcription-building-voiceyoutube.comBeyond Transcription: Building Voice AI That Understands Conversations — Hervé Bredin, pyannoteAIThe open ASR leaderboard reports Nvidia Parakeet at 11.4% word error rate on AMI meeting data. Hervé Bredin runs the same model on the same dataset and gets ...
https://www.youtube.com/watch?v=mFLlVpnGpdsSource handle nate-b-jones-build-a-token-dashboard-this-weeken. Links to https://www.youtube.com/watch?v=l8BloTSLK6M.nate-b-jones-build-a-token-dashboard-this-weekenyoutube.comNate B Jones - Build A Token Dashboard This Weekend. It'll Show The Work You Keep Avoiding.https://www.youtube.com/watch?v=l8BloTSLK6MSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/5/andreas-kling/.simonwillisonsimonwillison.netA quote from Andreas KlingWe will no longer accept public pull requests. [...] A substantial patch used to imply substantial effort, and that effort was a reasonable proxy for good faith. That assumption no …https://simonwillison.net/2026/Jun/5/andreas-kling/Source handle ai-engineer-beyond-components-designing-generati. Links to https://www.youtube.com/watch?v=hCMrEfPG2Yg.ai-engineer-beyond-components-designing-generatiyoutube.comBeyond Components: Designing Generative UI for MCP Apps — Ruben Casas, PostmanRuben Casas from Postman prompted a model to rewrite his blog. It built a search box with a blur animation and accessibility out of the box, without being as...
https://www.youtube.com/watch?v=hCMrEfPG2YgSource handle compuflair-the-physics-rule-that-stops-ai-from-g. Links to https://www.youtube.com/watch?v=l_gYpkYmbOc.compuflair-the-physics-rule-that-stops-ai-from-gyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=l_gYpkYmbOc
5 sourceshttps://www.youtube.com/watch?v=mFLlVpnGpdsSource handle nate-b-jones-build-a-token-dashboard-this-weeken. Links to https://www.youtube.com/watch?v=l8BloTSLK6M.nate-b-jones-build-a-token-dashboard-this-weekenyoutube.comNate B Jones - Build A Token Dashboard This Weekend. It'll Show The Work You Keep Avoiding.https://www.youtube.com/watch?v=l8BloTSLK6MSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/5/andreas-kling/.simonwillisonsimonwillison.netA quote from Andreas KlingWe will no longer accept public pull requests. [...] A substantial patch used to imply substantial effort, and that effort was a reasonable proxy for good faith. That assumption no …https://simonwillison.net/2026/Jun/5/andreas-kling/Source handle ai-engineer-beyond-components-designing-generati. Links to https://www.youtube.com/watch?v=hCMrEfPG2Yg.ai-engineer-beyond-components-designing-generatiyoutube.comBeyond Components: Designing Generative UI for MCP Apps — Ruben Casas, PostmanRuben Casas from Postman prompted a model to rewrite his blog. It built a search box with a blur animation and accessibility out of the box, without being as...https://www.youtube.com/watch?v=hCMrEfPG2YgSource handle compuflair-the-physics-rule-that-stops-ai-from-g. Links to https://www.youtube.com/watch?v=l_gYpkYmbOc.compuflair-the-physics-rule-that-stops-ai-from-gyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=l_gYpkYmbOcAI Digest
Coding Agents Hit the Workflow Wall
Coding-agent discourse shifted from benchmark gains toward workflow governance: durable decision records, executable specs, cost controls, task quality, and review systems now determine whether agent output becomes maintainable work.
coding-agentsworkflowagent-governance
Source handle ai-engineer-bdd-adr-prd-wtf-capturing-decisions. Links to https://www.youtube.com/watch?v=504PvfXou5Y.ai-engineer-bdd-adr-prd-wtf-capturing-decisionsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=504PvfXou5YSource handle nate-b-jones-opus-4-8-scored-81-your-workflow-do. Links to https://www.youtube.com/watch?v=z73yuF14udI.nate-b-jones-opus-4-8-scored-81-your-workflow-doyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=z73yuF14udISource handle simonwillison. Links to https://simonwillison.net/2026/Jun/3/uber-caps-usage/.simonwillisonsimonwillison.netUber Caps Usage of AI Tools Like Claude Code to Manage CostsI wrote the other day about Uber blowing its 2026 AI budget in four months, and how that wasn't particularly surprising given they would have set that budget in 2025, …https://simonwillison.net/2026/Jun/3/uber-caps-usage/Source handle ai-engineer-how-lovable-self-improves-every-hour. Links to https://www.youtube.com/watch?v=KA5kPbdkK2E.ai-engineer-how-lovable-self-improves-every-houryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=KA5kPbdkK2ESource handle ai-engineer-task-fidelity-scaling-laws-kobie-cra. Links to https://www.youtube.com/watch?v=YYH0DMQr30A.ai-engineer-task-fidelity-scaling-laws-kobie-crayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=YYH0DMQr30ASource handle ai-engineer-what-lies-beneath-the-api-benjamin-c. Links to https://www.youtube.com/watch?v=HvZXAOZ3iv8.ai-engineer-what-lies-beneath-the-api-benjamin-cyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=HvZXAOZ3iv8Source handle greg-isenberg-the-next-100b-market-selling-to-ai. Links to https://www.youtube.com/watch?v=MlptIfpoLlw.greg-isenberg-the-next-100b-market-selling-to-aiyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=MlptIfpoLlw
7 sources
AI Digest
Agent Ops Is Becoming an Infrastructure Problem
Today’s strongest AI discourse shifted from model capability to operational control: network-level identity for agent sandboxes, Pareto-based model selection, and recurring AI workflows that need policy, measurement, and review.
agentsai-operationsmodel-evaluation
Source handle ai-engineer-what-if-the-network-was-the-sandbox. Links to https://www.youtube.com/watch?v=BM2JX9hqsVQ.ai-engineer-what-if-the-network-was-the-sandboxyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=BM2JX9hqsVQSource handle ai-engineer-20-days-of-compute-vs-7-hours-rethin. Links to https://www.youtube.com/watch?v=hqHC6Z_lXyo.ai-engineer-20-days-of-compute-vs-7-hours-rethinyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=hqHC6Z_lXyoSource handle simonwillison. Links to https://simonwillison.net/2026/Jun/1/hackers-simply-asked-meta-ai/.simonwillisonsimonwillison.netHackers Simply Asked Meta AI to Give Them Access to High-Profile Instagram Accounts. It WorkedI had trouble believing this story was true, but I've seen it verified from multiple sources now: One video shows a hacker starting a conversation with Meta’s AI support bot …https://simonwillison.net/2026/Jun/1/hackers-simply-asked-meta-ai/Source handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/practical-ways-to-use-claude-routines.departmentofproduct-substackdepartmentofproduct.substack.comPractical ways to use Claude Routines at work - a Guide for Product Teams🧠 Backlog triaging, API deprecation watchdog, user feedback aggregation and more.
https://departmentofproduct.substack.com/p/practical-ways-to-use-claude-routinesSource handle wes-roth-gpt-5-6-about-to-drop. Links to https://www.youtube.com/watch?v=cS0Tm6ddnsQ.wes-roth-gpt-5-6-about-to-dropyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=cS0Tm6ddnsQ
5 sources
https://departmentofproduct.substack.com/p/practical-ways-to-use-claude-routinesSource handle wes-roth-gpt-5-6-about-to-drop. Links to https://www.youtube.com/watch?v=cS0Tm6ddnsQ.wes-roth-gpt-5-6-about-to-dropyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=cS0Tm6ddnsQAI Digest
Agents Move From Pass Rates to Operating Quality
Today’s strongest AI-discourse signal was a shift from raw model success to organizational quality: generated code, enterprise agents, and fast voice prototypes now need context, review, and product judgment to matter. The day reinforced a sober canon: agents raise the floor, but weak workflows c...
ai-agentscoding-agentssoftware-quality
Source handle ai-engineer-can-llms-generate-enterprise-quality. Links to https://www.youtube.com/watch?v=NuePCNMpWGc.ai-engineer-can-llms-generate-enterprise-qualityyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=NuePCNMpWGcSource handle theo-t3-gg-i-hate-that-this-is-true. Links to https://www.youtube.com/watch?v=rTMRlqT8Q8c.theo-t3-gg-i-hate-that-this-is-trueyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=rTMRlqT8Q8cSource handle nate-b-jones-this-is-how-ai-agents-actually-take. Links to https://www.youtube.com/shorts/KWCZWFDOapk.nate-b-jones-this-is-how-ai-agents-actually-takeyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/KWCZWFDOapkSource handle ai-engineer-how-to-talk-to-statues-joe-reeve-ele. Links to https://www.youtube.com/watch?v=u-rJwPPU3QA.ai-engineer-how-to-talk-to-statues-joe-reeve-eleyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=u-rJwPPU3QA
4 sources
AI Digest
Agents Need Proof, Not Benchmarks
Practitioner discourse converged on a sharper standard for agent trust: realistic benchmarks, explicit specs, containment boundaries, and hard-to-fake evidence matter more than polished demos or larger instruction packs.
agentsevaluationcoding-agents
Source handle ai-engineer-how-i-deleted-95-of-my-agent-skills. Links to https://www.youtube.com/watch?v=vy7o1g2iHY8.ai-engineer-how-i-deleted-95-of-my-agent-skillsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=vy7o1g2iHY8Source handle theo-t3-gg-ai-code-benchmarks-lied-to-us. Links to https://www.youtube.com/watch?v=JpSHyEIZ_bo.theo-t3-gg-ai-code-benchmarks-lied-to-usyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=JpSHyEIZ_boSource handle ai-engineer-spec-driven-testing-for-agents-with. Links to https://www.youtube.com/watch?v=UQKg0td-Bf4.ai-engineer-spec-driven-testing-for-agents-withyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=UQKg0td-Bf4Source handle ai-engineer-engineering-voice-agents-latency-qua. Links to https://www.youtube.com/watch?v=N7b1PJc7SFc.ai-engineer-engineering-voice-agents-latency-quayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=N7b1PJc7SFcSource handle simonwillison. Links to https://simonwillison.net/2026/May/30/how-we-contain-claude/.simonwillisonsimonwillison.netHow we contain Claude across productsA complaint I often have about sandboxing products is that they are rarely thoroughly documented, and in the absence of detailed documentation it's hard to know how much I can …https://simonwillison.net/2026/May/30/how-we-contain-claude/Source handle nate-b-jones-the-compound-risk-of-ai-agents-ai-r. Links to https://www.youtube.com/shorts/oTTVQt4IjPI.nate-b-jones-the-compound-risk-of-ai-agents-ai-ryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/oTTVQt4IjPISource handle nate-b-jones-microsoft-says-86-treat-ai-output-a. Links to https://www.youtube.com/watch?v=UsCgEuIAclE.nate-b-jones-microsoft-says-86-treat-ai-output-ayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=UsCgEuIAclESource handle simonwillison-2. Links to https://simonwillison.net/2026/May/31/the-solution-might-be-cancelling-my-ai-subscription/.simonwillison-2simonwillison.netThe solution might be cancelling my AI subscriptionI find this post by David Wilson very relatable. David lists 16+ projects he's spun up with AI tooling, and concludes: I didn't mean to build most of these things. …https://simonwillison.net/2026/May/31/the-solution-might-be-cancelling-my-ai-subscription/
8 sourcesAI Digest
Context Platforms Become the Agent Stack
Practitioner discourse converged on a new agent infrastructure frame: stateful context platforms, auditable memory, and branchable data/state matter more than chatbot interfaces. The evidence is still mostly commentary and demos, but it sharpens the operational question around where agents safely...
agentscontext-platformsenterprise-ai
Source handle nate-b-jones-the-death-of-the-filing-cabinet. Links to https://www.youtube.com/shorts/59NCmQ3hxz4.nate-b-jones-the-death-of-the-filing-cabinetyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/59NCmQ3hxz4Source handle nate-b-jones-how-ai-is-quietly-replacing-databas. Links to https://www.youtube.com/shorts/WPe45FOh7G8.nate-b-jones-how-ai-is-quietly-replacing-databasyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/WPe45FOh7G8Source handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/inside-nathan-s-second-brain-daniel-miessler-security-expert-creator-of-pai-audits-my-ai-setup/.cognitiverevolutioncognitiverevolution.aiInside Nathan's Second Brain: Daniel Miessler, Security Expert & Creator of PAI, Audits My AI SetupSecurity expert Daniel Miessler audits Nathan's personal AI infrastructure, including a Claude Code memory database and autonomous agents for scheduling, communications, and projects. They discuss agent hierarchy, security, AI disclosure norms, ideal-state prompts, and Bitter Lesson engineering.
https://www.cognitiverevolution.ai/inside-nathan-s-second-brain-daniel-miessler-security-expert-creator-of-pai-audits-my-ai-setup/Source handle wes-roth-ghost-ai-let-s-ai-agents-build-disposab. Links to https://www.youtube.com/watch?v=hGCrvDDpl5w.wes-roth-ghost-ai-let-s-ai-agents-build-disposabyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=hGCrvDDpl5wSource handle salesforce. Links to https://www.salesforce.com/news/stories/how-engineering-became-agentic/?bc=DB.salesforcesalesforce.comPioneering the Agentic Shift Within Salesforce EngineeringKey Takeaways Autonomous tools are now writing code, reviewing pull requests (“PRs”), and driving deployments across the software development lifecycle.
https://www.salesforce.com/news/stories/how-engineering-became-agentic/?bc=DB
5 sourceshttps://www.cognitiverevolution.ai/inside-nathan-s-second-brain-daniel-miessler-security-expert-creator-of-pai-audits-my-ai-setup/Source handle wes-roth-ghost-ai-let-s-ai-agents-build-disposab. Links to https://www.youtube.com/watch?v=hGCrvDDpl5w.wes-roth-ghost-ai-let-s-ai-agents-build-disposabyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=hGCrvDDpl5wSource handle salesforce. Links to https://www.salesforce.com/news/stories/how-engineering-became-agentic/?bc=DB.salesforcesalesforce.comPioneering the Agentic Shift Within Salesforce EngineeringKey Takeaways Autonomous tools are now writing code, reviewing pull requests (“PRs”), and driving deployments across the software development lifecycle.
https://www.salesforce.com/news/stories/how-engineering-became-agentic/?bc=DBAI Digest
Claude Code Meets the Production Wall
Claude Opus 4.8 mattered less as a standalone model launch than as part of a broader move toward orchestrated agentic coding. The day’s strongest discourse paired Claude Code dynamic workflows and messy reverse-engineering success with warnings about token cost, observability, enterprise governan...
claude-codeagentsai-engineering
Source handle simonwillison. Links to https://simonwillison.net/2026/May/28/claude-opus-4-8/.simonwillisonsimonwillison.netClaude Opus 4.8: “a modest but tangible improvement”Anthropic shipped Claude Opus 4.8 today. My favourite thing about it is this note in the release announcement: Users will find Opus 4.8 to be a modest but tangible improvement …
https://simonwillison.net/2026/May/28/claude-opus-4-8/Source handle nitter. Links to https://nitter.net/bcherny/status/2060048873440129073.nitternitter.netBoris Cherny - Claude Opus 4.8 and Claude Code dynamic workflows threadhttps://nitter.net/bcherny/status/2060048873440129073Source handle theo-t3-gg-anthropic-fights-back. Links to https://www.youtube.com/watch?v=_goOUJkkxUk.theo-t3-gg-anthropic-fights-backyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=_goOUJkkxUkSource handle ai-engineer-reverse-engineering-a-viking-voip-ph. Links to https://www.youtube.com/watch?v=V-L0INGTEOg.ai-engineer-reverse-engineering-a-viking-voip-phyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=V-L0INGTEOgSource handle ai-engineer-how-agent-o11y-differs-from-traditio. Links to https://www.youtube.com/watch?v=XBaznoTRDFI.ai-engineer-how-agent-o11y-differs-from-traditioyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=XBaznoTRDFISource handle ai-engineer-most-enterprise-agentic-projects-are. Links to https://www.youtube.com/watch?v=AGkzpxMdPn8.ai-engineer-most-enterprise-agentic-projects-areyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=AGkzpxMdPn8Source handle nate-b-jones-cheap-software-made-your-pm-job-har. Links to https://www.youtube.com/watch?v=b6J387xJvHg.nate-b-jones-cheap-software-made-your-pm-job-haryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=b6J387xJvHg
7 sourceshttps://simonwillison.net/2026/May/28/claude-opus-4-8/Source handle nitter. Links to https://nitter.net/bcherny/status/2060048873440129073.nitternitter.netBoris Cherny - Claude Opus 4.8 and Claude Code dynamic workflows threadhttps://nitter.net/bcherny/status/2060048873440129073Source handle theo-t3-gg-anthropic-fights-back. Links to https://www.youtube.com/watch?v=_goOUJkkxUk.theo-t3-gg-anthropic-fights-backyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=_goOUJkkxUkSource handle ai-engineer-reverse-engineering-a-viking-voip-ph. Links to https://www.youtube.com/watch?v=V-L0INGTEOg.ai-engineer-reverse-engineering-a-viking-voip-phyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=V-L0INGTEOgSource handle ai-engineer-how-agent-o11y-differs-from-traditio. Links to https://www.youtube.com/watch?v=XBaznoTRDFI.ai-engineer-how-agent-o11y-differs-from-traditioyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=XBaznoTRDFISource handle ai-engineer-most-enterprise-agentic-projects-are. Links to https://www.youtube.com/watch?v=AGkzpxMdPn8.ai-engineer-most-enterprise-agentic-projects-areyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=AGkzpxMdPn8Source handle nate-b-jones-cheap-software-made-your-pm-job-har. Links to https://www.youtube.com/watch?v=b6J387xJvHg.nate-b-jones-cheap-software-made-your-pm-job-haryoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=b6J387xJvHg
AI Digest
Agents Need Management, Not Just Prompts
Serious AI-agent discourse shifted toward governing delegated work: comprehension, explicit decision context, analytics, and escalation paths. The same evidence sits against a growing belief that enterprise model usage is real enough to make agent control surfaces operationally urgent.
ai-agentscoding-agentsagent-analytics
Source handle ai-engineer-comprehend-first-code-later-the-ai-s. Links to https://www.youtube.com/watch?v=li0SaBt9RDM.ai-engineer-comprehend-first-code-later-the-ai-syoutube.comAI Engineer - Comprehend First, Code Later: The AI Skill I Rely On Daily — Priscila Andre de Oliveira, Sentryhttps://www.youtube.com/watch?v=li0SaBt9RDMSource handle ai-engineer-context-graphs-for-explainable-decis. Links to https://www.youtube.com/watch?v=abvQEhvRI_c.ai-engineer-context-graphs-for-explainable-decisyoutube.comContext Graphs for Explainable, Decision-Aware AI Agents — Andreas Kollegger & Zaid Zaim, Neo4jPrescribing drug X is correct 99% of the time for symptom Y. For the 1% where it is fatal, statistical reasoning does not help you. Andreas Kollegger calls t...
https://www.youtube.com/watch?v=abvQEhvRI_cSource handle nate-b-jones-a-cursor-agent-wiped-a-database-in. Links to https://www.youtube.com/watch?v=n0nC1kmztSk.nate-b-jones-a-cursor-agent-wiped-a-database-inyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=n0nC1kmztSkSource handle simonwillison. Links to https://simonwillison.net/2026/May/27/product-market-fit/.simonwillisonsimonwillison.netI think Anthropic and OpenAI have found product-market fitAnthropic are strongly rumored to be about to have their first profitable quarter. Stories are circulating of companies surprised at how expensive their LLM bills are becoming from usage by …https://simonwillison.net/2026/May/27/product-market-fit/Source handle theo-t3-gg-holy-sh-t-i-think-anthropic-is-profit. Links to https://www.youtube.com/watch?v=q88yYhLSPC0.theo-t3-gg-holy-sh-t-i-think-anthropic-is-profityoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=q88yYhLSPC0
5 sources
https://www.youtube.com/watch?v=abvQEhvRI_cSource handle nate-b-jones-a-cursor-agent-wiped-a-database-in. Links to https://www.youtube.com/watch?v=n0nC1kmztSk.nate-b-jones-a-cursor-agent-wiped-a-database-inyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=n0nC1kmztSkSource handle simonwillison. Links to https://simonwillison.net/2026/May/27/product-market-fit/.simonwillisonsimonwillison.netI think Anthropic and OpenAI have found product-market fitAnthropic are strongly rumored to be about to have their first profitable quarter. Stories are circulating of companies surprised at how expensive their LLM bills are becoming from usage by …https://simonwillison.net/2026/May/27/product-market-fit/Source handle theo-t3-gg-holy-sh-t-i-think-anthropic-is-profit. Links to https://www.youtube.com/watch?v=q88yYhLSPC0.theo-t3-gg-holy-sh-t-i-think-anthropic-is-profityoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=q88yYhLSPC0AI Digest
Agent Work Moves From Prompting to Workflow Control
Today’s strongest AI discourse signal is that reliable agent work is becoming workflow design: context ownership, visible execution, reversible actions, trace-based evals, and adversarial verification matter as much as model choice or prompt wording.
agentsevalsai-workflows
Source handle ai-engineer-the-maturity-phases-of-running-evals. Links to https://www.youtube.com/watch?v=FB-MLPhL9Ms.ai-engineer-the-maturity-phases-of-running-evalsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=FB-MLPhL9MsSource handle nate-b-jones-your-board-deck-has-a-wrong-formula. Links to https://www.youtube.com/watch?v=MFzxIT88zfg.nate-b-jones-your-board-deck-has-a-wrong-formulayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=MFzxIT88zfgSource handle theo-t3-gg-how-i-code-with-ai-changed-a-lot. Links to https://www.youtube.com/watch?v=xJaMTo2YgO8.theo-t3-gg-how-i-code-with-ai-changed-a-lotyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=xJaMTo2YgO8Source handle ai-engineer-what-the-best-agents-share. Links to https://www.youtube.com/watch?v=7CrPrHgoEYk.ai-engineer-what-the-best-agents-shareyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=7CrPrHgoEYkSource handle simonwillison. Links to https://simonwillison.net/2026/May/26/copilot-cowork-exfiltrates-files/.simonwillisonsimonwillison.netMicrosoft Copilot Cowork Exfiltrates FilesThe biggest challenge in designing agentic systems continues to be preventing them from enabling attackers to exfiltrate data. In this case Microsoft Copilot Cowork (yes, that's a real product name) …https://simonwillison.net/2026/May/26/copilot-cowork-exfiltrates-files/Source handle ai-engineer-stop-babysitting-your-agents. Links to https://www.youtube.com/watch?v=BiG2ssibKGc.ai-engineer-stop-babysitting-your-agentsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=BiG2ssibKGcSource handle nate-b-jones-shopify-ceo-reveals-their-secret-ai. Links to https://www.youtube.com/watch?v=NRBQmwlILjk.nate-b-jones-shopify-ceo-reveals-their-secret-aiyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=NRBQmwlILjkSource handle ai-engineer-frontier-ai-at-home. Links to https://www.youtube.com/watch?v=ESbWpPT_9-o.ai-engineer-frontier-ai-at-homeyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=ESbWpPT_9-oSource handle jack-clark. Links to https://jack-clark.net/2026/05/26/import-ai-458-reckoning-with-the-future-and-a-singularity-story/.jack-clarkjack-clark.netImport AI 458: Reckoning with the future; and a singularity storyWelcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. This issue consists of a lengthy essay based on a speech I recently gave, and a fictional story attempting to think through what a positive singularity might look…
https://jack-clark.net/2026/05/26/import-ai-458-reckoning-with-the-future-and-a-singularity-story/Source handle oneusefulthing. Links to https://www.oneusefulthing.org/p/choosing-to-stay-human.oneusefulthingoneusefulthing.orgChoosing to Stay HumanIf you go to your favorite social media site, you will find it full of posts that start to look suspiciously similar to each other:
https://www.oneusefulthing.org/p/choosing-to-stay-human
10 sources
https://jack-clark.net/2026/05/26/import-ai-458-reckoning-with-the-future-and-a-singularity-story/Source handle oneusefulthing. Links to https://www.oneusefulthing.org/p/choosing-to-stay-human.oneusefulthingoneusefulthing.orgChoosing to Stay HumanIf you go to your favorite social media site, you will find it full of posts that start to look suspiciously similar to each other:
https://www.oneusefulthing.org/p/choosing-to-stay-humanAI Digest
Coding Agents Are Now Workflow Systems
Coding-agent discourse shifted from raw model comparisons toward workflow design, verification, harness quality, and evaluation infrastructure. The strongest evidence came from Theo’s Claude Code/Codex/Cursor comparison and Google DeepMind/Kaggle’s agent-evaluation framing.
coding-agentsagent-evalsai-tools
Source handle theo-t3-gg-claude-code-vs-codex-vs-cursor-an-hon. Links to https://www.youtube.com/watch?v=JMYspR42HFM.theo-t3-gg-claude-code-vs-codex-vs-cursor-an-honyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=JMYspR42HFMSource handle ai-engineer-agentic-evaluations-at-scale-for-eve. Links to https://www.youtube.com/watch?v=Ubwb6NzegyA.ai-engineer-agentic-evaluations-at-scale-for-eveyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=Ubwb6NzegyASource handle simonwillison. Links to https://simonwillison.net/2026/May/25/encyclical-on-ai/.simonwillisonsimonwillison.netNotes on Pope Leo XIV’s encyclical on AIDropped this morning by the Vatican: Magnifica Humanitas of His Holiness Pope Leo XIV on Safeguarding the Human Person in the Time of Artificial Intelligence. This is a very interesting …https://simonwillison.net/2026/May/25/encyclical-on-ai/Source handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/your-biggest-lever-designing-your-ai-career-for-maximum-impact-with-80000-hours-founder-ben-todd/.cognitiverevolutioncognitiverevolution.aiYour Biggest Lever: Designing your AI Career for Maximum Impact, with 80,000 Hours founder Ben Todd80,000 Hours co-founder Ben Todd discusses how to choose high-impact AI careers, covering timelines, global risks, frontier lab tradeoffs, policy levers, funding, and when to join or start organizations.
https://www.cognitiverevolution.ai/your-biggest-lever-designing-your-ai-career-for-maximum-impact-with-80000-hours-founder-ben-todd/
4 sources
https://www.cognitiverevolution.ai/your-biggest-lever-designing-your-ai-career-for-maximum-impact-with-80000-hours-founder-ben-todd/AI Digest
Agents Are Becoming Platform Workloads
Coding agents are moving from developer-tool demos into platform workloads, creating new pressure around quotas, review, observability, procurement, and ownership. The strongest evidence came from OpenAI and Google DeepMind infrastructure discussions, reinforced by practitioner notes on agent-tea...
ai-agentsinfrastructuredeveloper-tools
Source handle nate-b-jones-the-infrastructure-nightmare-nobody. Links to https://www.youtube.com/watch?v=z3pbrFKVyQE.nate-b-jones-the-infrastructure-nightmare-nobodyyoutube.comThe Infrastructure Nightmare Nobody Is Talking AboutFull Post w/ Prompt Pack Build Your Own Eval Suite: https://natesnewsletter.substack.com/p/ai-big-tech-industrial-business?r=1z4sm5&utm_campaign=post&utm_med...
https://www.youtube.com/watch?v=z3pbrFKVyQESource handle ai-engineer-how-google-deepmind-runs-agents-at-s. Links to https://www.youtube.com/watch?v=7gujZrJ9L5I.ai-engineer-how-google-deepmind-runs-agents-at-syoutube.comHow Google DeepMind Runs Agents at Scale — KP Sawhney & Ian Ballantyne, Google DeepMindGoogle DeepMind employees have worse token quotas than paying customers. That is not a mistake. KP Sawhney explains: customers get priority, and if an intern...
https://www.youtube.com/watch?v=7gujZrJ9L5ISource handle ai-engineer-does-genai-belong-to-data-scientists. Links to https://www.youtube.com/watch?v=NKwIX3CiRgU.ai-engineer-does-genai-belong-to-data-scientistsyoutube.comDoes GenAI "belong" to data scientists? — Phil Hetzel, BraintrustAt most traditional enterprises, GenAI got handed to the ML platform team because it had AI in the name. Phil Hetzel from Braintrust argues that was the wron...
https://www.youtube.com/watch?v=NKwIX3CiRgUSource handle nate-b-jones-why-the-ai-boom-is-about-to-hit-a-w. Links to https://www.youtube.com/watch?v=Poyi6X7rOwY.nate-b-jones-why-the-ai-boom-is-about-to-hit-a-wyoutube.comWhy the AI boom is about to hit a wallFull Post w/ Prompt Pack: https://natesnewsletter.substack.com/p/ai-big-tech-industrial-business?r=1z4sm5&utm_campaign=post&utm_medium=web&showWelcomeOnShare...
https://www.youtube.com/watch?v=Poyi6X7rOwYSource handle simonwillison. Links to https://simonwillison.net/2026/May/24/armin-ronacher/.simonwillisonsimonwillison.netA quote from Armin RonacherThe most frustrating failure mode right now is that people submit issues that are not in their own voice. They contain an observed problem somewhere, but it has been thrown …https://simonwillison.net/2026/May/24/armin-ronacher/
5 sourceshttps://www.youtube.com/watch?v=z3pbrFKVyQESource handle ai-engineer-how-google-deepmind-runs-agents-at-s. Links to https://www.youtube.com/watch?v=7gujZrJ9L5I.ai-engineer-how-google-deepmind-runs-agents-at-syoutube.comHow Google DeepMind Runs Agents at Scale — KP Sawhney & Ian Ballantyne, Google DeepMindGoogle DeepMind employees have worse token quotas than paying customers. That is not a mistake. KP Sawhney explains: customers get priority, and if an intern...https://www.youtube.com/watch?v=7gujZrJ9L5ISource handle ai-engineer-does-genai-belong-to-data-scientists. Links to https://www.youtube.com/watch?v=NKwIX3CiRgU.ai-engineer-does-genai-belong-to-data-scientistsyoutube.comDoes GenAI "belong" to data scientists? — Phil Hetzel, BraintrustAt most traditional enterprises, GenAI got handed to the ML platform team because it had AI in the name. Phil Hetzel from Braintrust argues that was the wron...https://www.youtube.com/watch?v=NKwIX3CiRgUSource handle nate-b-jones-why-the-ai-boom-is-about-to-hit-a-w. Links to https://www.youtube.com/watch?v=Poyi6X7rOwY.nate-b-jones-why-the-ai-boom-is-about-to-hit-a-wyoutube.comWhy the AI boom is about to hit a wallFull Post w/ Prompt Pack: https://natesnewsletter.substack.com/p/ai-big-tech-industrial-business?r=1z4sm5&utm_campaign=post&utm_medium=web&showWelcomeOnShare...https://www.youtube.com/watch?v=Poyi6X7rOwYSource handle simonwillison. Links to https://simonwillison.net/2026/May/24/armin-ronacher/.simonwillisonsimonwillison.netA quote from Armin RonacherThe most frustrating failure mode right now is that people submit issues that are not in their own voice. They contain an observed problem somewhere, but it has been thrown …https://simonwillison.net/2026/May/24/armin-ronacher/AI Digest
Agent Time Becomes the Bottleneck
Coding-agent discourse is shifting from model capability to the operational problem of keeping multiple semi-autonomous sessions moving. The day’s strongest signal is that attention, orchestration, and interruption design are becoming core productivity bottlenecks.
coding-agentsdeveloper-workflowagent-orchestration
Source handle ai-engineer-let-s-talk-about-fomat-fear-of-missi. Links to https://www.youtube.com/watch?v=W-SX_srBa3Y.ai-engineer-let-s-talk-about-fomat-fear-of-missiyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=W-SX_srBa3YSource handle ai-engineer-scaling-the-next-paradigm-of-heterog. Links to https://www.youtube.com/watch?v=WRBNDpUhsJQ.ai-engineer-scaling-the-next-paradigm-of-heterogyoutube.comAI Engineer - Scaling the Next Paradigm of Heterogeneous Intelligence — Adrian Bertagnoli, Callosumhttps://www.youtube.com/watch?v=WRBNDpUhsJQSource handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/all-compute-is-food-palisade-s-jeffrey-ladish-on-ai-shutdown-resistance-self-replication-ecology/.cognitiverevolutioncognitiverevolution.aiAll Compute Is Food: Palisade's Jeffrey Ladish on AI Shutdown Resistance, Self-Replication & EcologyPalisade Research executive director Jeffrey Ladish discusses findings on AI shutdown resistance and self-replication, including cybersecurity risks, deceptive long-horizon training, and proposals for compute governance and pausing recursive self-improvement.
https://www.cognitiverevolution.ai/all-compute-is-food-palisade-s-jeffrey-ladish-on-ai-shutdown-resistance-self-replication-ecology/
3 sourceshttps://www.cognitiverevolution.ai/all-compute-is-food-palisade-s-jeffrey-ladish-on-ai-shutdown-resistance-self-replication-ecology/AI Digest
Agents Become Workflow Infrastructure
The strongest discourse signal was a convergence around agents as managed workflow infrastructure: isolated, permissioned, source-aware, and embedded into IDEs, data tools, mobile platforms, and enterprise runtimes. The day’s practical lesson is to judge agents by their scaffolding and auditabili...
agentsworkflowdeveloper-tools
Source handle ai-engineer-cooking-with-agents-in-vs-code-liam. Links to https://www.youtube.com/watch?v=dyHpnnlkTc8.ai-engineer-cooking-with-agents-in-vs-code-liamyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=dyHpnnlkTc8Source handle ai-engineer-scaling-agents-on-kubernetes-with-ac. Links to https://www.youtube.com/watch?v=VaS2h-dY1-4.ai-engineer-scaling-agents-on-kubernetes-with-acyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=VaS2h-dY1-4Source handle simonwillison. Links to https://simonwillison.net/2026/May/21/datasette-agent/.simonwillisonsimonwillison.netDatasette AgentWe just announced the first release of Datasette Agent, a new extensible AI assistant for Datasette. I’ve been working on my LLM Python library for just over three years now, …
https://simonwillison.net/2026/May/21/datasette-agent/Source handle simonwillison-2. Links to https://simonwillison.net/2026/May/21/datasette-agent-sprites/.simonwillison-2simonwillison.netRelease: datasette-agent-sprites 0.1a0Datasette Agent tools for working with Fly Spriteshttps://simonwillison.net/2026/May/21/datasette-agent-sprites/Source handle ai-engineer-ai-on-android-ask-me-anything-florin. Links to https://www.youtube.com/watch?v=owH1f0N-keY.ai-engineer-ai-on-android-ask-me-anything-florinyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=owH1f0N-keYSource handle nate-b-jones-your-ai-writes-from-twenty-sources. Links to https://www.youtube.com/watch?v=ltbzgzZZmgI.nate-b-jones-your-ai-writes-from-twenty-sourcesyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=ltbzgzZZmgISource handle simonwillison-3. Links to https://simonwillison.net/2026/May/22/ftc-active-listening/.simonwillison-3simonwillison.netFTC to Require Cox Media Group, Two Other Firms to Pay Nearly $1 Million to Settle Charges They Deceived Customers About “Active Listening” AI-Powered Marketing ServiceBack in 2024 Cox Media Group were caught trying to sell advertisers packages based on "active listening", with this deck which claimed: Smart devices capture real-time intent data by listening …https://simonwillison.net/2026/May/22/ftc-active-listening/Source handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/google-search-now-generates-ui-on.departmentofproduct-substackdepartmentofproduct.substack.com🔵 Google Search now generates UI on demandWill this set a new benchmark for in-product search capabilities? Plus: takeaways from Anthropic's Code with Claude event, Figma fights back and new data reveals the state of AI in design
https://departmentofproduct.substack.com/p/google-search-now-generates-ui-onSource handle compuflair-how-physics-can-teach-neural-nets-to. Links to https://www.youtube.com/watch?v=8nS_Zz3mN4s.compuflair-how-physics-can-teach-neural-nets-toyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=8nS_Zz3mN4s
9 sourceshttps://simonwillison.net/2026/May/21/datasette-agent/Source handle simonwillison-2. Links to https://simonwillison.net/2026/May/21/datasette-agent-sprites/.simonwillison-2simonwillison.netRelease: datasette-agent-sprites 0.1a0Datasette Agent tools for working with Fly Spriteshttps://simonwillison.net/2026/May/21/datasette-agent-sprites/Source handle ai-engineer-ai-on-android-ask-me-anything-florin. Links to https://www.youtube.com/watch?v=owH1f0N-keY.ai-engineer-ai-on-android-ask-me-anything-florinyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=owH1f0N-keYSource handle nate-b-jones-your-ai-writes-from-twenty-sources. Links to https://www.youtube.com/watch?v=ltbzgzZZmgI.nate-b-jones-your-ai-writes-from-twenty-sourcesyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=ltbzgzZZmgISource handle simonwillison-3. Links to https://simonwillison.net/2026/May/22/ftc-active-listening/.simonwillison-3simonwillison.netFTC to Require Cox Media Group, Two Other Firms to Pay Nearly $1 Million to Settle Charges They Deceived Customers About “Active Listening” AI-Powered Marketing ServiceBack in 2024 Cox Media Group were caught trying to sell advertisers packages based on "active listening", with this deck which claimed: Smart devices capture real-time intent data by listening …https://simonwillison.net/2026/May/22/ftc-active-listening/Source handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/google-search-now-generates-ui-on.departmentofproduct-substackdepartmentofproduct.substack.com🔵 Google Search now generates UI on demandWill this set a new benchmark for in-product search capabilities? Plus: takeaways from Anthropic's Code with Claude event, Figma fights back and new data reveals the state of AI in designhttps://departmentofproduct.substack.com/p/google-search-now-generates-ui-onSource handle compuflair-how-physics-can-teach-neural-nets-to. Links to https://www.youtube.com/watch?v=8nS_Zz3mN4s.compuflair-how-physics-can-teach-neural-nets-toyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=8nS_Zz3mN4sAI Digest
Agents Move From Chat to Engineering Surfaces
Today’s strongest AI discourse signal is that useful agents increasingly depend on engineered surfaces: skills, traces, evals, open repositories, compute jobs, APIs, and metrics. The practical canon is moving from clever prompting toward environments that make agent work inspectable, repeatable, ...
agentsai-engineeringcoding-agents
Source handle ai-engineer-your-coding-agent-should-do-ai-syste. Links to https://www.youtube.com/watch?v=JomVvNDjGb8.ai-engineer-your-coding-agent-should-do-ai-systeyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=JomVvNDjGb8Source handle ai-engineer-langfuse-coding-agent-skill-for-obse. Links to https://www.youtube.com/watch?v=vNCY9kXXyDQ.ai-engineer-langfuse-coding-agent-skill-for-obseyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=vNCY9kXXyDQSource handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/the-model-eats-the-scaffolding-deepmind-s-logan-kilpatrick-tulsee-doshi-on-3-5-flash-omni-more/.cognitiverevolutioncognitiverevolution.aiThe Model Eats the Scaffolding: DeepMind's Logan Kilpatrick & Tulsee Doshi on 3.5 Flash, Omni & MoreGoogle DeepMind's Logan Kilpatrick and Tulsee Doshi discuss Gemini 3.5 Flash, Omni video generation, and Gemini Spark, plus Google's emphasis on efficiency, agent harnesses, context limits, model psychology, AI welfare, and recursive self-improvement.
https://www.cognitiverevolution.ai/the-model-eats-the-scaffolding-deepmind-s-logan-kilpatrick-tulsee-doshi-on-3-5-flash-omni-more/Source handle ai-engineer-gemini-s-any-to-any-multimodal-agent. Links to https://www.youtube.com/watch?v=GIRpQEfYf3U.ai-engineer-gemini-s-any-to-any-multimodal-agentyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=GIRpQEfYf3USource handle nate-b-jones-you-re-not-bad-at-prompting-you-re. Links to https://www.youtube.com/watch?v=ogTLWGBc3cE.nate-b-jones-you-re-not-bad-at-prompting-you-reyoutube.comOpus 4.7 and OpenAI 5.5 Made Your Prompting Style Obsolete.Full Post w/ Prompt Pack: https://natesnewsletter.substack.com/p/ai-agents-better-communicator?r=1z4sm5&utm_campaign=post&utm_medium=web&showWelcomeOnShare=t...
https://www.youtube.com/watch?v=ogTLWGBc3cESource handle theo-t3-gg-discussion-of-extension-and-package-s. Links to https://www.youtube.com/watch?v=XKA94rcu8b8.theo-t3-gg-discussion-of-extension-and-package-syoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=XKA94rcu8b8Source handle simonwillison. Links to https://simonwillison.net/2026/May/20/tokens-per-second/.simonwillisonsimonwillison.netHow fast is 10 tokens per second really?Neat little HTML app by Mike Veerman (source code here) which simulates LLM token output speeds from 5/second to 800/second. Useful if you see a model advertised as "30 tokens/second" …https://simonwillison.net/2026/May/20/tokens-per-second/Source handle simonwillison-2. Links to https://simonwillison.net/2026/May/20/google-io/.simonwillison-2simonwillison.netGoogle I/O, Gemini Spark, AntigravityIt's hard to find much to write about Google I/O this year because I have a policy of not writing about anything that I can't try out myself, and a …https://simonwillison.net/2026/May/20/google-io/
8 sources
https://www.cognitiverevolution.ai/the-model-eats-the-scaffolding-deepmind-s-logan-kilpatrick-tulsee-doshi-on-3-5-flash-omni-more/Source handle ai-engineer-gemini-s-any-to-any-multimodal-agent. Links to https://www.youtube.com/watch?v=GIRpQEfYf3U.ai-engineer-gemini-s-any-to-any-multimodal-agentyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=GIRpQEfYf3USource handle nate-b-jones-you-re-not-bad-at-prompting-you-re. Links to https://www.youtube.com/watch?v=ogTLWGBc3cE.nate-b-jones-you-re-not-bad-at-prompting-you-reyoutube.comOpus 4.7 and OpenAI 5.5 Made Your Prompting Style Obsolete.Full Post w/ Prompt Pack: https://natesnewsletter.substack.com/p/ai-agents-better-communicator?r=1z4sm5&utm_campaign=post&utm_medium=web&showWelcomeOnShare=t...https://www.youtube.com/watch?v=ogTLWGBc3cESource handle theo-t3-gg-discussion-of-extension-and-package-s. Links to https://www.youtube.com/watch?v=XKA94rcu8b8.theo-t3-gg-discussion-of-extension-and-package-syoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=XKA94rcu8b8Source handle simonwillison. Links to https://simonwillison.net/2026/May/20/tokens-per-second/.simonwillisonsimonwillison.netHow fast is 10 tokens per second really?Neat little HTML app by Mike Veerman (source code here) which simulates LLM token output speeds from 5/second to 800/second. Useful if you see a model advertised as "30 tokens/second" …https://simonwillison.net/2026/May/20/tokens-per-second/Source handle simonwillison-2. Links to https://simonwillison.net/2026/May/20/google-io/.simonwillison-2simonwillison.netGoogle I/O, Gemini Spark, AntigravityIt's hard to find much to write about Google I/O this year because I have a policy of not writing about anything that I can't try out myself, and a …https://simonwillison.net/2026/May/20/google-io/AI Digest
Agent Maturity Moves From Demos to Control Systems
Agent discourse is converging on control systems: state, authority, tools, observability, user steering, and shutdown paths matter more than demo autonomy. The strongest evidence came from practitioner talks on agent maturity, protocols, deployment infrastructure, and on-device LLM agents.
agentsai-engineeringmcp
Source handle ai-engineer-don-t-build-slop-4-levels-of-ai-agen. Links to https://www.youtube.com/watch?v=yUmS-F9IX90.ai-engineer-don-t-build-slop-4-levels-of-ai-agenyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=yUmS-F9IX90Source handle nate-b-jones-google-spent-a-year-stitching-mcp-a. Links to https://www.youtube.com/watch?v=zP6TnEiueEc.nate-b-jones-google-spent-a-year-stitching-mcp-ayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=zP6TnEiueEcSource handle nate-b-jones-cloudflare-stripe-and-okta-decide-w. Links to https://www.youtube.com/watch?v=woGB2vr5wTg.nate-b-jones-cloudflare-stripe-and-okta-decide-wyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=woGB2vr5wTgSource handle ai-engineer-from-46-to-90-fine-tuning-tiny-llms. Links to https://www.youtube.com/watch?v=-TiET_K-E_g.ai-engineer-from-46-to-90-fine-tuning-tiny-llmsyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=-TiET_K-E_gSource handle ai-engineer-personalization-in-the-era-of-llms-s. Links to https://www.youtube.com/watch?v=5YSJEP0HWzM.ai-engineer-personalization-in-the-era-of-llms-syoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=5YSJEP0HWzMSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/how-ubers-product-teams-built-a-prd.departmentofproduct-substackdepartmentofproduct.substack.comHow Uber’s product teams built a PRD reviewer that catches gaps before they reach leadershipAnd how to build your own version. Plus: other examples from DoorDash, Atlassian and more. AI augmented product development process explored.
https://departmentofproduct.substack.com/p/how-ubers-product-teams-built-a-prdSource handle nitter. Links to https://nitter.net/karpathy/status/2056753169888334312.nitternitter.netAndrej Karpathy - Joins Anthropichttps://nitter.net/karpathy/status/2056753169888334312
7 sources
https://departmentofproduct.substack.com/p/how-ubers-product-teams-built-a-prdSource handle nitter. Links to https://nitter.net/karpathy/status/2056753169888334312.nitternitter.netAndrej Karpathy - Joins Anthropichttps://nitter.net/karpathy/status/2056753169888334312AI Digest
Agent Workflows Become the Main AI Story
AI discourse centered on the operational scaffolding behind useful agents: context management, skills, verification, machine-readable evidence, and institutional capacity. The strongest signal is that autonomy is now being judged as a work-system problem, not just a model-capability race.
ai-agentscoding-agentsworkflow
Source handle simonwillison. Links to https://simonwillison.net/2026/May/19/5-minute-llms/.simonwillisonsimonwillison.netThe last six months in LLMs in five minutesI put together these annotated slides from my five minute lightning talk at PyCon US 2026, using the latest iteration of my annotated presentation tool. # I presented this lightning …
https://simonwillison.net/2026/May/19/5-minute-llms/Source handle ai-engineer-build-agents-that-run-for-hours-with. Links to https://www.youtube.com/watch?v=mR-WAvEPRwE.ai-engineer-build-agents-that-run-for-hours-withyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=mR-WAvEPRwESource handle nate-b-jones-the-prove-it-economy-is-here-and-mo. Links to https://www.youtube.com/watch?v=725QE_LNXT4.nate-b-jones-the-prove-it-economy-is-here-and-moyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=725QE_LNXT4Source handle ai-engineer-rewiring-the-state-eoin-mulgrew-10-d. Links to https://www.youtube.com/watch?v=ObNKGf9YR0g.ai-engineer-rewiring-the-state-eoin-mulgrew-10-dyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=ObNKGf9YR0g
4 sourceshttps://simonwillison.net/2026/May/19/5-minute-llms/Source handle ai-engineer-build-agents-that-run-for-hours-with. Links to https://www.youtube.com/watch?v=mR-WAvEPRwE.ai-engineer-build-agents-that-run-for-hours-withyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=mR-WAvEPRwESource handle nate-b-jones-the-prove-it-economy-is-here-and-mo. Links to https://www.youtube.com/watch?v=725QE_LNXT4.nate-b-jones-the-prove-it-economy-is-here-and-moyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=725QE_LNXT4Source handle ai-engineer-rewiring-the-state-eoin-mulgrew-10-d. Links to https://www.youtube.com/watch?v=ObNKGf9YR0g.ai-engineer-rewiring-the-state-eoin-mulgrew-10-dyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=ObNKGf9YR0g
AI Digest
Agent Reliability Moves Out of the Prompt
The day’s strongest AI-discourse signal was a shift from prompts and model capability toward engineered agent systems: durable sessions, harnesses, verification, cost awareness, and workflow-level adoption. The practical takeaway is that serious AI products increasingly look like observable produ...
ai-agentsai-uxagent-harnesses
Source handle ai-engineer-why-your-ai-ux-is-broken-and-it-s-no. Links to https://www.youtube.com/watch?v=YNJvm7t3yq8.ai-engineer-why-your-ai-ux-is-broken-and-it-s-noyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=YNJvm7t3yq8Source handle ai-engineer-harnesses-in-ai-a-deep-dive. Links to https://www.youtube.com/watch?v=C_GG5g38vLU.ai-engineer-harnesses-in-ai-a-deep-diveyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=C_GG5g38vLUSource handle nate-b-jones-5-levers-that-separate-winning-ai-i. Links to https://www.youtube.com/watch?v=LIkYVsxMpS8.nate-b-jones-5-levers-that-separate-winning-ai-iyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=LIkYVsxMpS8Source handle nate-b-jones-how-teams-accidentally-sabotage-ai. Links to https://www.youtube.com/shorts/WwpDswEjTXg.nate-b-jones-how-teams-accidentally-sabotage-aiyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/WwpDswEjTXgSource handle simonwillison. Links to https://simonwillison.net/2026/May/17/gds-weighs-in/.simonwillisonsimonwillison.netGDS weighs in on the NHS’s decision to retreat from Open SourceTerence Eden continues his coverage of the NHS' poorly considered decision to close down access to their open source repositories in response to vulnerabilities reported to them as part of …https://simonwillison.net/2026/May/17/gds-weighs-in/Source handle exponentialview. Links to https://www.exponentialview.co/p/monday-data-the-cost-of-tokenmaxxing.exponentialviewexponentialview.co📈 Data to start your week: The cost of tokenmaxxingBye, bye fixed costs. Hello token anxiety.
https://www.exponentialview.co/p/monday-data-the-cost-of-tokenmaxxing
6 sources
https://www.exponentialview.co/p/monday-data-the-cost-of-tokenmaxxingAI Digest
Agents Need Specs, Experts, and Cost Controls
The strongest discourse signal was a shift from model capability to operational maturity: agents need behavioral specs, domain-expert review loops, recovery paths, and task-level cost controls before teams can delegate serious work.
ai-agentsai-engineeringtesting
Source handle ai-engineer-beyond-code-coverage-functionality-t. Links to https://www.youtube.com/watch?v=FWEInOtngmM.ai-engineer-beyond-code-coverage-functionality-tyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=FWEInOtngmMSource handle ai-engineer-how-to-leverage-domain-expertise. Links to https://www.youtube.com/watch?v=kfSDc2eVLo4.ai-engineer-how-to-leverage-domain-expertiseyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=kfSDc2eVLo4Source handle nate-b-jones-anthropic-s-mythos-just-beat-openai. Links to https://www.youtube.com/watch?v=dm3_Z-5PYnQ.nate-b-jones-anthropic-s-mythos-just-beat-openaiyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=dm3_Z-5PYnQSource handle nate-b-jones-agents-vs-chatbots-codex-changes-ev. Links to https://www.youtube.com/shorts/D0Lc4HF5pzE.nate-b-jones-agents-vs-chatbots-codex-changes-evyoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/shorts/D0Lc4HF5pzESource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/notions-new-workers-can-build-stripe.departmentofproduct-substackdepartmentofproduct.substack.com🔵 Notion's new Workers can build Stripe Dashboards and is a Claude Backlash brewing?Plus: Google's vibe coded widgets, the companies who are ditching AI features, a genuinely useful bookmark tool you can use.
https://departmentofproduct.substack.com/p/notions-new-workers-can-build-stripeSource handle simonwillison. Links to https://simonwillison.net/2026/May/16/openclaw-names/.simonwillisonsimonwillison.netWarelay -> OpenClawIn preparation for a lightning talk I'm giving at PyCon US this afternoon I decided to figure out how many names OpenClaw has actually had since that first commit back …https://simonwillison.net/2026/May/16/openclaw-names/
6 sourceshttps://departmentofproduct.substack.com/p/notions-new-workers-can-build-stripeSource handle simonwillison. Links to https://simonwillison.net/2026/May/16/openclaw-names/.simonwillisonsimonwillison.netWarelay -> OpenClawIn preparation for a lightning talk I'm giving at PyCon US this afternoon I decided to figure out how many names OpenClaw has actually had since that first commit back …https://simonwillison.net/2026/May/16/openclaw-names/AI Digest
Context Becomes the Agent Platform
The strongest AI discourse signal is a shift from model access toward context systems, observable execution, workflow ownership, and cheaper long-context operation. Agents look most durable where their memory, provenance, and operating costs can be made legible for real work.
agentscontextobservability
Source handle ai-engineer-connecting-the-dots-with-context-gra. Links to https://www.youtube.com/watch?v=eW_vxrjvERk.ai-engineer-connecting-the-dots-with-context-grayoutube.com- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.https://www.youtube.com/watch?v=eW_vxrjvERkSource handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/three-kinds-of-software-survive-tasklet-s-andrew-lee-on-competing-to-be-a-horizontal-platform/.cognitiverevolutioncognitiverevolution.aiThree Kinds of Software Survive: Tasklet's Andrew Lee on Competing to be a Horizontal PlatformTasklet CEO Andrew Lee discusses rebuilding the company's agent stack around file-system context, agentic search, and summarization. He also examines platform strategy, competition with model providers, and which software companies may survive the AI transition.
https://www.cognitiverevolution.ai/three-kinds-of-software-survive-tasklet-s-andrew-lee-on-competing-to-be-a-horizontal-platform/Source handle magazine-sebastianraschka. Links to https://magazine.sebastianraschka.com/p/recent-developments-in-llm-architectures.magazine-sebastianraschkamagazine.sebastianraschka.comRecent Developments in LLM Architectures: KV Sharing, mHC, and Compressed AttentionFrom Gemma 4 to DeepSeek V4, How New Open-Weight LLMs Are Reducing Long-Context Costs
https://magazine.sebastianraschka.com/p/recent-developments-in-llm-architectures
3 sourceshttps://www.cognitiverevolution.ai/three-kinds-of-software-survive-tasklet-s-andrew-lee-on-competing-to-be-a-horizontal-platform/Source handle magazine-sebastianraschka. Links to https://magazine.sebastianraschka.com/p/recent-developments-in-llm-architectures.magazine-sebastianraschkamagazine.sebastianraschka.comRecent Developments in LLM Architectures: KV Sharing, mHC, and Compressed AttentionFrom Gemma 4 to DeepSeek V4, How New Open-Weight LLMs Are Reducing Long-Context Costshttps://magazine.sebastianraschka.com/p/recent-developments-in-llm-architecturesAI Digest
Claude Code’s Subscription Boundary
Coding-agent discourse shifted toward platform economics: Anthropic’s Claude Code boundary raises questions about whether independent wrappers and automated workflows can remain viable under subscription pricing. A smaller Datasette/Codex signal reinforces the need for portable, auditable agent s...
coding-agentsclaude-codeplatform-economics
Source handle theo-t3-gg-anthropic-s-claude-code-subscription. Links to https://www.youtube.com/watch?v=131yAOjxHHQ.theo-t3-gg-anthropic-s-claude-code-subscriptionyoutube.comTheo / t3.gg - Anthropic’s Claude Code subscription boundary becomes the discourse storyhttps://www.youtube.com/watch?v=131yAOjxHHQSource handle simonwillison. Links to https://simonwillison.net/2026/May/13/welcome-to-the-datasette-blog/.simonwillisonsimonwillison.netSimon Willison - Datasette launches an official blog, built with Codex desktophttps://simonwillison.net/2026/May/13/welcome-to-the-datasette-blog/
2 sourcesAI Digest
Agent Workflows Are Becoming Continuous Systems
The day’s strongest AI discourse signal is a shift from better prompting toward continuous agent operating loops: specs, memory contracts, adaptive evals, richer review artifacts, and production feedback. The useful test is whether an AI proposal explains how agent work is specified, contextualiz...
agentic-workflowsai-engineeringcontinuous-compute
Source handle ai-engineer-ci-cd-is-dead-agents-need-continuous. Links to https://www.youtube.com/watch?v=VktrqzQgytY.ai-engineer-ci-cd-is-dead-agents-need-continuousyoutube.comAI Engineer - CI/CD Is Dead, Agents Need Continuous Compute and Computershttps://www.youtube.com/watch?v=VktrqzQgytYSource handle ai-engineer-lessons-from-trillion-token-deployme. Links to https://www.youtube.com/watch?v=X6NShR2ccOg.ai-engineer-lessons-from-trillion-token-deploymeyoutube.comAI Engineer - Lessons from Trillion Token Deployments at Fortune 500shttps://www.youtube.com/watch?v=X6NShR2ccOgSource handle ai-engineer-dark-factory-how-openclaw-ships-fast. Links to https://www.youtube.com/watch?v=4VhbYlfC7Gs.ai-engineer-dark-factory-how-openclaw-ships-fastyoutube.comAI Engineer - Dark Factory: How OpenClaw Ships Faster Than You Can Read the Diffhttps://www.youtube.com/watch?v=4VhbYlfC7GsSource handle nate-b-jones-sap-just-spent-1b-on-the-agentic-ra. Links to https://www.youtube.com/watch?v=lqiwQiDglGk.nate-b-jones-sap-just-spent-1b-on-the-agentic-rayoutube.comNate B Jones - SAP Just Spent $1B+ on the Agentic RAG Problem Most Teams Missedhttps://www.youtube.com/watch?v=lqiwQiDglGkSource handle theo-t3-gg-stop-letting-your-agents-write-markdo. Links to https://www.youtube.com/watch?v=S9EGx6ik-18.theo-t3-gg-stop-letting-your-agents-write-markdoyoutube.comTheo / t3.gg - Stop letting your agents write Markdownhttps://www.youtube.com/watch?v=S9EGx6ik-18Source handle nate-b-jones-2025-prompting-vs-2026-prompting. Links to https://www.youtube.com/shorts/x8Y404We8O8.nate-b-jones-2025-prompting-vs-2026-promptingyoutube.comNate B Jones - 2025 Prompting vs 2026 Promptinghttps://www.youtube.com/shorts/x8Y404We8O8Source handle simonwillison. Links to https://simonwillison.net/2026/May/12/mitchell-hashimoto/.simonwillisonsimonwillison.netSimon Willison - Quoting Mitchell Hashimotohttps://simonwillison.net/2026/May/12/mitchell-hashimoto/Source handle simonwillison-2. Links to https://simonwillison.net/2026/May/12/mo-bitar/.simonwillison-2simonwillison.netSimon Willison - Quoting Mo Bitarhttps://simonwillison.net/2026/May/12/mo-bitar/
8 sourcesAI Digest
Agents Hit the Accountability Layer
AI-agent discourse is shifting from raw capability to accountability: authorization, auditability, maintenance cost, and ownership. The strongest signals came from agentic commerce, AI-assisted rewrites, and management uses of the “agentic era” frame.
ai-agentsagentic-commercesoftware-maintenance
Source handle nate-b-jones-agentic-commerce-is-really-a-fight. Links to https://www.youtube.com/watch?v=j5_wcDifNko.nate-b-jones-agentic-commerce-is-really-a-fightyoutube.comNate B Jones - Agentic commerce is really a fight over delegated authority, audit evidence, and liabilityhttps://www.youtube.com/watch?v=j5_wcDifNkoSource handle theo-t3-gg-bun-s-ai-assisted-rust-rewrite-become. Links to https://www.youtube.com/watch?v=gILMoijqeGA.theo-t3-gg-bun-s-ai-assisted-rust-rewrite-becomeyoutube.comTheo / t3.gg - Bun’s AI-assisted Rust rewrite becomes a case study in agentic refactoring riskhttps://www.youtube.com/watch?v=gILMoijqeGASource handle simonwillison. Links to https://simonwillison.net/2026/May/11/james-shore/.simonwillisonsimonwillison.netSimon Willison - AI coding agents must lower maintenance cost, not just speed up code creationhttps://simonwillison.net/2026/May/11/james-shore/Source handle simonwillison-2. Links to https://simonwillison.net/2026/May/11/gitlab-act-2/.simonwillison-2simonwillison.netSimon Willison - GitLab frames workforce restructuring around the “agentic era”https://simonwillison.net/2026/May/11/gitlab-act-2/Source handle nate-b-jones-career-moat-the-scarce-operator-who. Links to https://www.youtube.com/shorts/OhtcvogXtok.nate-b-jones-career-moat-the-scarce-operator-whoyoutube.comNate B Jones - Career moat: the scarce operator who can translate AI capability into workflow-specific planshttps://www.youtube.com/shorts/OhtcvogXtokSource handle simonwillison-3. Links to https://simonwillison.net/2026/May/11/zombie-internet/.simonwillison-3simonwillison.netSimon Willison - The “Zombie Internet” frame: AI prose is creating cognitive load even when humans are still presenthttps://simonwillison.net/2026/May/11/zombie-internet/
6 sourcesAI Digest
Production Agents Need Boundaries, Memory, and Public Workflows
The day’s strongest AI discourse shifted from model choice to the operating environment around production agents: context architecture, visible work trails, action-boundary validation, and durable execution. The practical lesson is to treat agents as governed coworkers and product infrastructure,...
ai-agentsagent-governancecontext-engineering
Source handle ai-engineer-hierarchical-memory-context-manageme. Links to https://www.youtube.com/watch?v=esY99nYXxR4.ai-engineer-hierarchical-memory-context-managemeyoutube.comAI Engineer - Hierarchical Memory: Context Management in Agentshttps://www.youtube.com/watch?v=esY99nYXxR4Source handle ai-engineer-you-can-t-just-one-shot-it. Links to https://www.youtube.com/watch?v=ON5LIT0M4do.ai-engineer-you-can-t-just-one-shot-ityoutube.comAI Engineer - You can’t just one shot ithttps://www.youtube.com/watch?v=ON5LIT0M4doSource handle ai-engineer-two-roads-to-durable-agents-replay-v. Links to https://www.youtube.com/watch?v=svCnShDvgQg.ai-engineer-two-roads-to-durable-agents-replay-vyoutube.comAI Engineer - Two Roads to Durable Agents: Replay vs. Snapshothttps://www.youtube.com/watch?v=svCnShDvgQgSource handle simonwillison. Links to https://simonwillison.net/2026/May/11/learning-on-the-shop-floor/.simonwillisonsimonwillison.netSimon Willison - Learning on the Shop floorhttps://simonwillison.net/2026/May/11/learning-on-the-shop-floor/Source handle ai-engineer-viktor-ai-coworker-that-lives-in-sla. Links to https://www.youtube.com/watch?v=ohKt066uFhg.ai-engineer-viktor-ai-coworker-that-lives-in-slayoutube.comAI Engineer - Viktor: AI Coworker That Lives in Slackhttps://www.youtube.com/watch?v=ohKt066uFhgSource handle ai-engineer-a-piece-of-pi-embedding-the-openclaw. Links to https://www.youtube.com/watch?v=vAIDdLKB6-w.ai-engineer-a-piece-of-pi-embedding-the-openclawyoutube.comAI Engineer - A Piece of Pi: Embedding The OpenClaw Coding Agent In Your Producthttps://www.youtube.com/watch?v=vAIDdLKB6-wSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/what-is-a-command-line-interface.departmentofproduct-substackdepartmentofproduct.substack.comDepartment of Product - What is a Command Line Interface and why do they suddenly matter to product teams?https://departmentofproduct.substack.com/p/what-is-a-command-line-interfaceSource handle nate-b-jones-llm-agents-the-security-breach-patt. Links to https://www.youtube.com/watch?v=SX1myuPEDFg.nate-b-jones-llm-agents-the-security-breach-pattyoutube.comNate B Jones - LLM Agents: The Security Breach Pattern Nobody's Talking Abouthttps://www.youtube.com/watch?v=SX1myuPEDFgSource handle theo-t3-gg-we-all-fell-for-it. Links to https://www.youtube.com/watch?v=lNVa33qUzZ8.theo-t3-gg-we-all-fell-for-ityoutube.comTheo / t3.gg - We all fell for it…https://www.youtube.com/watch?v=lNVa33qUzZ8Source handle simonwillison-2. Links to https://simonwillison.net/2026/May/10/new-york-times-editors-note/.simonwillison-2simonwillison.netSimon Willison - Quoting New York Times Editors’ Notehttps://simonwillison.net/2026/May/10/new-york-times-editors-note/Source handle jack-clark. Links to https://jack-clark.net/2026/05/11/import-ai-456-rsi-and-economic-growth-radical-optionality-for-ai-regulation-and-a-neural-computer/.jack-clarkjack-clark.netJack Clark / Import AI - RSI and economic growth; radical optionality for AI regulation; and a neural computerhttps://jack-clark.net/2026/05/11/import-ai-456-rsi-and-economic-growth-radical-optionality-for-ai-regulation-and-a-neural-computer/
11 sourcesAI Digest
AI’s Bottleneck Moved from Generation to Judgment
AI discourse in the last 24 hours centered less on raw model capability and more on whether AI systems can be made timely, accountable, and worth maintaining. Voice-agent latency, enterprise oversight, and coding-agent judgment all point to deployment constraints becoming the main bottleneck.
voice-agentsai-adoptioncoding-agents
Source handle ai-engineer-why-tts-models-now-look-like-llms-sa. Links to https://www.youtube.com/watch?v=3jGAU2sbAyY.ai-engineer-why-tts-models-now-look-like-llms-sayoutube.comAI Engineer - Why TTS Models Now Look Like LLMs — Samuel Humeau, Mistralhttps://www.youtube.com/watch?v=3jGAU2sbAyYSource handle nate-b-jones-human-oversight-isn-t-slowing-ai-do. Links to https://www.youtube.com/shorts/lNX4m6KmX4U.nate-b-jones-human-oversight-isn-t-slowing-ai-doyoutube.comNate B Jones - Human Oversight Isn't Slowing AI Down, It's Protecting Ithttps://www.youtube.com/shorts/lNX4m6KmX4USource handle simonwillison. Links to https://simonwillison.net/2026/May/10/andrew-quinn/.simonwillisonsimonwillison.netSimon Willison - Quoting Andrew Quinnhttps://simonwillison.net/2026/May/10/andrew-quinn/Source handle nitter. Links to https://nitter.net/BrianRoemmele/status/2053152857386619116.nitternitter.netBrian Roemmele / Nitter - Enterprise OpenClaw cost audit threadhttps://nitter.net/BrianRoemmele/status/2053152857386619116
4 sourcesAI Digest
Voice Agents Meet the Systems-Engineering Wall
Voice AI discourse shifted from demo quality toward the hard product stack: transport fidelity, turn-taking, tool latency, observability, privacy, and cost. The same maturation showed up in agent-workflow commentary, where repeatable packaging and deterministic checks matter more than better one-...
voice-aiagentsai-workflows
Source handle ai-engineer-voice-ai-when-is-the-her-moment-neil. Links to https://www.youtube.com/watch?v=P_RI1kCkRbo.ai-engineer-voice-ai-when-is-the-her-moment-neilyoutube.comAI Engineer - Voice AI: when is the "Her" moment? — Neil Zeghidour, Gradium AIhttps://www.youtube.com/watch?v=P_RI1kCkRboSource handle ai-engineer-give-your-chat-agent-a-voice-luke-ha. Links to https://www.youtube.com/watch?v=DCZZ3AJKzuc.ai-engineer-give-your-chat-agent-a-voice-luke-hayoutube.comAI Engineer - Give Your Chat Agent a Voice — Luke Harries, ElevenLabshttps://www.youtube.com/watch?v=DCZZ3AJKzucSource handle simonwillison. Links to https://simonwillison.net/2026/May/9/luke-curley/.simonwillisonsimonwillison.netSimon Willison - Quoting Luke Curleyhttps://simonwillison.net/2026/May/9/luke-curley/Source handle nate-b-jones-everyone-is-prompting-better-almost. Links to https://www.youtube.com/watch?v=647pSnX5H_Y.nate-b-jones-everyone-is-prompting-better-almostyoutube.comNate B Jones - Everyone Is Prompting Better. Almost Nobody Is Packaging Work.https://www.youtube.com/watch?v=647pSnX5H_YSource handle compuflair-the-physics-secret-to-building-stable. Links to https://www.youtube.com/watch?v=Px0a9eexeu8.compuflair-the-physics-secret-to-building-stableyoutube.comCompuFlair - The Physics Secret to Building Stable AI Agentshttps://www.youtube.com/watch?v=Px0a9eexeu8Source handle simonwillison-2. Links to https://simonwillison.net/2026/May/8/unreasonable-effectiveness-of-html/.simonwillison-2simonwillison.netSimon Willison - Using Claude Code: The Unreasonable Effectiveness of HTMLhttps://simonwillison.net/2026/May/8/unreasonable-effectiveness-of-html/
6 sourcesAI Digest
Production Agents Need Runtime Infrastructure
The strongest discourse signal was a shift from model choice toward production-agent infrastructure: observability, externalized memory, permissions, checkpoints, and model-swappable runtimes. Operator attention should move from prompt demos to telemetry and durable state.
agentsobservabilityagent-runtime
Source handle ai-engineer-everything-you-need-to-know-about-ag. Links to https://www.youtube.com/watch?v=-aM2EDTiaMs.ai-engineer-everything-you-need-to-know-about-agyoutube.comAI Engineer - Everything You Need To Know About Agent Observability — Danny Gollapalli and Ben Hylak, Raindrophttps://www.youtube.com/watch?v=-aM2EDTiaMsSource handle nate-b-jones-i-tested-openclaw-against-model-chu. Links to https://www.youtube.com/watch?v=85Q9htV2CBE.nate-b-jones-i-tested-openclaw-against-model-chuyoutube.comNate B Jones - I Tested OpenClaw Against Model Churn. Here's What Survived.https://www.youtube.com/watch?v=85Q9htV2CBESource handle theo-t3-gg-anthropic-just-wait-what. Links to https://www.youtube.com/watch?v=3pkz-Ie_k_c.theo-t3-gg-anthropic-just-wait-whatyoutube.comTheo / t3.gg - Anthropic just…wait whathttps://www.youtube.com/watch?v=3pkz-Ie_k_cSource handle wes-roth-google-just-turned-this-mmo-into-an-agi. Links to https://www.youtube.com/watch?v=vqBSNTZeLsg.wes-roth-google-just-turned-this-mmo-into-an-agiyoutube.comWes Roth - Google just turned this MMO into an AGI experimenthttps://www.youtube.com/watch?v=vqBSNTZeLsg
4 sourcesAI Digest
Agent Interfaces Move Beyond Chat
The day’s strongest AI-discourse signal was a shift from raw model output toward workflow-native agent interfaces, especially MCP Apps/MCP UI. Related evidence from creative tools, enterprise deployment, and embodied-agent failures points to harnesses, control surfaces, and operational fit as the...
agentsmcpai-product
Source handle ai-engineer-mcp-apps-as-a-standard-for-interacti. Links to https://www.youtube.com/watch?v=o-zkvb0iFDQ.ai-engineer-mcp-apps-as-a-standard-for-interactiyoutube.comAI Engineer - MCP Apps as a standard for interactive UI inside agent/chat hostshttps://www.youtube.com/watch?v=o-zkvb0iFDQSource handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/descript-isn-t-a-slop-machine-laura-burkhauser-on-the-ai-tools-creators-love-and-hate/.cognitiverevolutioncognitiverevolution.aiThe Cognitive Revolution - Descript isn’t a slop machine: Laura Burkhauser on the AI tools creators love and hatehttps://www.cognitiverevolution.ai/descript-isn-t-a-slop-machine-laura-burkhauser-on-the-ai-tools-creators-love-and-hate/Source handle wes-roth-anthropic-openai-enterprise-deployment. Links to https://www.youtube.com/watch?v=rzVhPTSNWnE.wes-roth-anthropic-openai-enterprise-deploymentyoutube.comWes Roth - Anthropic/OpenAI enterprise deployment thesis via forward-deployed engineershttps://www.youtube.com/watch?v=rzVhPTSNWnESource handle simonwillison. Links to https://simonwillison.net/2026/May/5/our-ai-started-a-cafe-in-stockholm/.simonwillisonsimonwillison.netSimon Willison - Our AI started a cafe in Stockholmhttps://simonwillison.net/2026/May/5/our-ai-started-a-cafe-in-stockholm/
4 sourcesAI Digest
Small Models Become Infrastructure
The strongest AI discourse signal was an operational turn: small and distilled models are useful, but only when teams understand their failure boundaries and build serving, routing, observability, and capacity strategy around them.
small-modelsai-infrastructuredistillation
Source handle ai-engineer-the-small-model-infrastructure-nobod. Links to https://www.youtube.com/watch?v=qdh_x-uRs9g.ai-engineer-the-small-model-infrastructure-nobodyoutube.comAI Engineer - The Small Model Infrastructure Nobody Built (So We Did)https://www.youtube.com/watch?v=qdh_x-uRs9gSource handle nate-b-jones-this-is-why-distilled-models-collap. Links to https://www.youtube.com/shorts/X1h75jqeD0Q.nate-b-jones-this-is-why-distilled-models-collapyoutube.comNate B Jones - This Is Why Distilled Models Collapsehttps://www.youtube.com/shorts/X1h75jqeD0QSource handle theo-t3-gg-prime-is-mostly-right-about-ai. Links to https://www.youtube.com/watch?v=VDPMXSAxiWk.theo-t3-gg-prime-is-mostly-right-about-aiyoutube.comTheo / t3.gg - Prime is (mostly) right about AIhttps://www.youtube.com/watch?v=VDPMXSAxiWkSource handle simonwillison. Links to https://simonwillison.net/2026/May/4/granite-41-3b-svg-pelican-gallery/.simonwillisonsimonwillison.netSimon Willison - Granite 4.1 3B SVG Pelican Galleryhttps://simonwillison.net/2026/May/4/granite-41-3b-svg-pelican-gallery/Source handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/spotifys-new-natural-language-api.departmentofproduct-substackdepartmentofproduct.substack.comDepartment of Product - Spotify’s new Natural Language API Interface and other Examples Exploredhttps://departmentofproduct.substack.com/p/spotifys-new-natural-language-api
5 sourcesAI Digest
AI Work Is Becoming Loop Work
The strongest discourse signal is a convergence around iterative AI loops: automated AI research is becoming a strategic accelerator, while builders are finding that simple tool-using loops often beat elaborate orchestration. The organizational consequence is that task ownership may erode before ...
agentsai-researchautomation
Source handle jack-clark. Links to https://jack-clark.net/2026/05/04/import-ai-455-automating-ai-research/.jack-clarkjack-clark.netJack Clark / Import AI - Import AI 455: Automating AI Researchhttps://jack-clark.net/2026/05/04/import-ai-455-automating-ai-research/Source handle ai-engineer-ralph-loops-build-dumb-ai-loops-that. Links to https://www.youtube.com/watch?v=2TLXsxkz0zI.ai-engineer-ralph-loops-build-dumb-ai-loops-thatyoutube.comAI Engineer - Ralph Loops: Build Dumb AI Loops That Ship — Chris Parsons, Cherrypickhttps://www.youtube.com/watch?v=2TLXsxkz0zISource handle exponentialview. Links to https://www.exponentialview.co/p/data-to-start-your-week-compute-crunch-may-2026.exponentialviewexponentialview.coAzeem Azhar / Exponential View - Data to start your week: AI boom, nowhere near the ceilinghttps://www.exponentialview.co/p/data-to-start-your-week-compute-crunch-may-2026Source handle nate-b-jones-your-performance-review-is-lying-to. Links to https://www.youtube.com/watch?v=rYqt6mMlv7o.nate-b-jones-your-performance-review-is-lying-toyoutube.comNate B Jones - Your Performance Review Is Lying To You By 18 Months.https://www.youtube.com/watch?v=rYqt6mMlv7o
4 sourcesAI Digest
Issue Trackers as Agent Control Planes
The strongest AI-discourse signal is that agents make structured work-state systems more important, not less. Issue trackers and similar tools become durable state graphs for ownership, permissions, history, and safe agent actions.
ai-agentsworkflowagent-control-planes
Read digest AI Digest
Prose Is the Agent Control Plane
Practitioner discourse converged on a concrete pattern: reliable agent work is being built from versioned prose, examples, goal loops, APIs, permissions, and external evaluators. The implication is that instructions and harnesses now need the same ownership, review, and rollback discipline as code.
ai-agentscoding-agentsagent-workflows
Source handle ai-engineer-posthog-llm-codegen-fails-and-how-to. Links to https://www.youtube.com/watch?v=juoNbJiZUi0.ai-engineer-posthog-llm-codegen-fails-and-how-toyoutube.comAI Engineer / PostHog - LLM codegen fails and how to stop 'emhttps://www.youtube.com/watch?v=juoNbJiZUi0Source handle simonwillison. Links to https://simonwillison.net/2026/Apr/30/codex-goals/.simonwillisonsimonwillison.netSimon Willison - Codex CLI 0.128.0 adds /goalhttps://simonwillison.net/2026/Apr/30/codex-goals/Source handle nitter. Links to https://nitter.net/karpathy/status/2049903821095354523.nitternitter.netAndrej Karpathy via Nitter - Sequoia Ascent highlights / LLMs beyond speedupshttps://nitter.net/karpathy/status/2049903821095354523Source handle ai-engineer-swyx-running-ai-engineer-with-ai. Links to https://www.youtube.com/watch?v=zepu8Kk6FBQ.ai-engineer-swyx-running-ai-engineer-with-aiyoutube.comAI Engineer / swyx - Running AI Engineer with AIhttps://www.youtube.com/watch?v=zepu8Kk6FBQSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/stripe-shows-off-its-new-product.departmentofproduct-substackdepartmentofproduct.substack.comDepartment of Product - Stripe shows off its new product payment capabilitieshttps://departmentofproduct.substack.com/p/stripe-shows-off-its-new-productSource handle nate-b-jones-microsoft-is-testing-claude-against. Links to https://www.youtube.com/watch?v=JvCtGjrn_N0.nate-b-jones-microsoft-is-testing-claude-againstyoutube.comNate B Jones - Microsoft Is Testing Claude Against Its Own Copilot. Here's Why.https://www.youtube.com/watch?v=JvCtGjrn_N0Source handle nate-b-jones-tests-vs-scenarios-which-one-actual. Links to https://www.youtube.com/shorts/g2occe4xMHk.nate-b-jones-tests-vs-scenarios-which-one-actualyoutube.comNate B Jones - Tests vs Scenarios: Which One Actually Workshttps://www.youtube.com/shorts/g2occe4xMHkSource handle simonwillison-2. Links to https://simonwillison.net/2026/Apr/30/andrew-kelley/.simonwillison-2simonwillison.netSimon Willison - Quoting Andrew Kelleyhttps://simonwillison.net/2026/Apr/30/andrew-kelley/Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/ai-tinkerers-24-community-spotlights.post-training-aitinkererspost-training.aitinkerers.orgAI Tinkerers / Post-Training - Top AI Demos #24: Deterministic agent pipelines, code leaks and agent memoryhttps://post-training.aitinkerers.org/p/ai-tinkerers-24-community-spotlights
9 sourcesAI Digest
Agent Harnesses Meet Governance
Practitioner discourse centered on where agent workflows should live: hard-coded harnesses, markdown skills, governed tools, or human-maintained institutions. The signal is a shift from model demos toward product architecture, maintainer accountability, and resilient development infrastructure.
ai-agentsagent-harnessesdeveloper-workflows
Source handle ai-engineer-replacing-12k-loc-with-a-200-loc-ski. Links to https://www.youtube.com/watch?v=WE_Gnowy3uw.ai-engineer-replacing-12k-loc-with-a-200-loc-skiyoutube.comAI Engineer - Replacing 12K LoC with a 200 LoC Skill — David Gomes, Cursorhttps://www.youtube.com/watch?v=WE_Gnowy3uwSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/are-agent-harnesses-the-new-secret.departmentofproduct-substackdepartmentofproduct.substack.comRich Holmes / Department of Product - Are Agent Harnesses the New Secret...https://departmentofproduct.substack.com/p/are-agent-harnesses-the-new-secretSource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/ai-tinkerers-24-community-spotlights.post-training-aitinkererspost-training.aitinkerers.orgAI Tinkerers / Post-Training - AI Tinkerers #24: Community Spotlightshttps://post-training.aitinkerers.org/p/ai-tinkerers-24-community-spotlightsSource handle simonwillison. Links to https://simonwillison.net/2026/Apr/29/llm/.simonwillisonsimonwillison.netSimon Willison - LLM 0.32a0https://simonwillison.net/2026/Apr/29/llm/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Apr/30/zig-anti-ai/.simonwillison-2simonwillison.netSimon Willison - Zig's anti-AI policyhttps://simonwillison.net/2026/Apr/30/zig-anti-ai/Source handle theo-t3-gg-the-painful-death-of-github. Links to https://www.youtube.com/watch?v=R7ex-Gt8dtw.theo-t3-gg-the-painful-death-of-githubyoutube.comTheo / t3.gg - The painful death of Githubhttps://www.youtube.com/watch?v=R7ex-Gt8dtw
6 sourcesAI Digest
Coding Agents Become Operations Systems
Practitioner discourse around coding agents is converging on operations: evals, identity, reproducible environments, team governance, and model routing now matter more than raw coding demos. The strongest signal is that adoption depends on turning personal agent tricks into accountable, observabl...
coding-agentsagent-opsevals
Source handle ai-engineer-building-your-own-software-factory. Links to https://www.youtube.com/watch?v=rnDm57Py54A.ai-engineer-building-your-own-software-factoryyoutube.comAI Engineer - Building your own software factoryhttps://www.youtube.com/watch?v=rnDm57Py54ASource handle ai-engineer-why-building-eval-platforms-is-hard. Links to https://www.youtube.com/watch?v=_fQ7Z_Wfouk.ai-engineer-why-building-eval-platforms-is-hardyoutube.comAI Engineer - Why building eval platforms is hardhttps://www.youtube.com/watch?v=_fQ7Z_WfoukSource handle ai-engineer-one-login-to-rule-them-all-cross-app. Links to https://www.youtube.com/watch?v=EmhRyw6xeT0.ai-engineer-one-login-to-rule-them-all-cross-appyoutube.comAI Engineer - One Login to Rule Them All: Cross-App Access for MCPhttps://www.youtube.com/watch?v=EmhRyw6xeT0Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/ai-tinkerers-24-community-spotlights.post-training-aitinkererspost-training.aitinkerers.orgAI Tinkerers / Post-Training - Top AI Demos #24: Deterministic agent pipelines, code leaks and agent memoryhttps://post-training.aitinkerers.org/p/ai-tinkerers-24-community-spotlightsSource handle ai-engineer-everything-i-learned-training-fronti. Links to https://www.youtube.com/watch?v=fLUtUkqYHnQ.ai-engineer-everything-i-learned-training-frontiyoutube.comAI Engineer - Everything I Learned Training Frontier Small Modelshttps://www.youtube.com/watch?v=fLUtUkqYHnQSource handle nate-b-jones-gpt-5-5-vs-claude-vs-gemini-the-rea. Links to https://www.youtube.com/watch?v=9aIYhjeYxzM.nate-b-jones-gpt-5-5-vs-claude-vs-gemini-the-reayoutube.comNate B Jones - GPT-5.5 vs Claude vs Gemini: The Real Difference Nobody's Talking Abouthttps://www.youtube.com/watch?v=9aIYhjeYxzMSource handle simonwillison. Links to https://simonwillison.net/2026/Apr/28/openai-codex/.simonwillisonsimonwillison.netSimon Willison - Quoting OpenAI Codex base_instructionshttps://simonwillison.net/2026/Apr/28/openai-codex/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Apr/28/matthew-yglesias/.simonwillison-2simonwillison.netSimon Willison - Quoting Matthew Yglesiashttps://simonwillison.net/2026/Apr/28/matthew-yglesias/
8 sourcesAI Digest
Trust Signals After AI Slop
AI discourse today centered on how cheap generated artifacts weaken traditional evidence of competence and product trust. The actionable shift is toward observable process, scoped interfaces, and agent workflows that can prove why their outputs deserve confidence.
developer-workflowstrustai-ux
Source handle theo-t3-gg-realistic-advice-about-software-dev-r. Links to https://www.youtube.com/watch?v=88qc67oYDl4.theo-t3-gg-realistic-advice-about-software-dev-ryoutube.comTheo / t3.gg - Realistic advice about software dev right nowhttps://www.youtube.com/watch?v=88qc67oYDl4Source handle nngroup. Links to https://www.nngroup.com/articles/ai-chatbots-design-guidelines/?utm_source=rss&utm_medium=feed&utm_campaign=rss-syndication.nngroupnngroup.comNielsen Norman Group - 10 Guidelines for Designing Your Site’s AI Chatbotshttps://www.nngroup.com/articles/ai-chatbots-design-guidelines/?utm_source=rss&utm_medium=feed&utm_campaign=rss-syndicationSource handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/ai-tinkerers-24-community-spotlights.post-training-aitinkererspost-training.aitinkerers.orgAI Tinkerers / Post-Training - Top AI Demos #24: Deterministic agent pipelines, code leaks and agent memoryhttps://post-training.aitinkerers.org/p/ai-tinkerers-24-community-spotlightsSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/designmd-explained-the-format-reshaping.departmentofproduct-substackdepartmentofproduct.substack.comDepartment of Product - DESIGN.md Explained - The Format Reshaping How AI Builds UIhttps://departmentofproduct.substack.com/p/designmd-explained-the-format-reshaping
4 sourcesAI Digest
Agent Control Beats Specs-to-Code
Practitioner discourse shifted toward a harder question than raw capability: how to keep coding and desktop agents inside reviewable, governable workflows. The strongest signals argued that broader execution surfaces make software fundamentals, supervision, and explicit control points more import...
agentscoding-agentsworkflow
Source handle youtube. Links to https://www.youtube.com/watch?v=v4F1gFy-hqg.youtubeyoutube.comyoutubehttps://www.youtube.com/watch?v=v4F1gFy-hqgSource handle youtube-2. Links to https://www.youtube.com/watch?v=2d9ZmA-4QzU.youtube-2youtube.comyoutubehttps://www.youtube.com/watch?v=2d9ZmA-4QzU
2 sourcesAI Digest
Judgment Becomes the Bottleneck
The clearest AI discourse shift is that faster generation is raising the value of judgment, constraint obedience, and trust in software workflows. Mozilla's Firefox security review result shows the upside, while practitioner commentary says the winning teams will be the ones with better quality l...
software-engineeringcoding-agentssecurity
Source handle blog-mozilla. Links to https://blog.mozilla.org/en/privacy-security/ai-security-zero-day-vulnerabilities/.blog-mozillablog.mozilla.orgbloghttps://blog.mozilla.org/en/privacy-security/ai-security-zero-day-vulnerabilities/Source handle youtube. Links to https://www.youtube.com/watch?v=wjk0ulMAkbc.youtubeyoutube.comyoutubehttps://www.youtube.com/watch?v=wjk0ulMAkbcSource handle simonwillison. Links to https://simonwillison.net/2026/Apr/21/andreas-pahlsson-notini/.simonwillisonsimonwillison.netsimonwillisonhttps://simonwillison.net/2026/Apr/21/andreas-pahlsson-notini/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Apr/22/claude-code-confusion/.simonwillison-2simonwillison.netsimonwillisonhttps://simonwillison.net/2026/Apr/22/claude-code-confusion/
4 sourcesAI Digest
Workflow Design Is the Real AI Speed Limit
The strongest AI discourse signal today is that practitioners are hitting workflow limits before model limits. Across coding, design, agent operations, and local inference, the winning pattern is bounded, reviewable loops with memory, recovery, and explicit handoffs instead of raw generation alone.
developer-workflowsagentsdesign-tools
Source handle post-training-aitinkerers. Links to https://post-training.aitinkerers.org/p/ai-tinkerers-23-community-spotlights.post-training-aitinkererspost-training.aitinkerers.org"AI Tinkerers #23: Community Spotlights"https://post-training.aitinkerers.org/p/ai-tinkerers-23-community-spotlightsSource handle why-ai-coding-is-slowing-developers-down. Links to https://www.youtube.com/shorts/kp4quCdtvXI.why-ai-coding-is-slowing-developers-downyoutube.com"Why AI Coding is Slowing Developers Down"https://www.youtube.com/shorts/kp4quCdtvXISource handle i-tried-claude-design-so-you-don-t-have-to. Links to https://www.youtube.com/watch?v=wDgq9aiuL-w.i-tried-claude-design-so-you-don-t-have-toyoutube.com"I Tried Claude Design So You Don't Have To"https://www.youtube.com/watch?v=wDgq9aiuL-wSource handle run-ai-locally-on-iphone-with-mlx-swift-lm-adrie. Links to https://www.youtube.com/watch?v=a2muGkT4WD4.run-ai-locally-on-iphone-with-mlx-swift-lm-adrieyoutube.com"Run AI Locally on iPhone with MLX Swift LM - Adrien Grondin"https://www.youtube.com/watch?v=a2muGkT4WD4
4 sourcesAI Digest
Agents as Software Users
Practitioner discourse converged on a specific design shift: agents are becoming a first-class user of software, pushing builders toward headless interfaces, capability-scoped runtimes, and machine-legible workflows. The strongest evidence came from product, runtime, and research angles that all ...
agentsapplication-layerapis
Source handle simonwillison. Links to https://simonwillison.net/2026/Apr/19/headless-everything/.simonwillisonsimonwillison.net"Headless everything for personal AI"https://simonwillison.net/2026/Apr/19/headless-everything/Source handle the-new-application-layer-malte-ubl-cto-vercel. Links to https://www.youtube.com/watch?v=XKup1pj-34M.the-new-application-layer-malte-ubl-cto-vercelyoutube.com"The New Application Layer - Malte Ubl, CTO Vercel"https://www.youtube.com/watch?v=XKup1pj-34MSource handle code-mode-let-the-code-do-the-talking-sunil-pai. Links to https://www.youtube.com/watch?v=8txf05vVVl4.code-mode-let-the-code-do-the-talking-sunil-paiyoutube.com"Code Mode: Let the Code do the Talking - Sunil Pai, Cloudflare"https://www.youtube.com/watch?v=8txf05vVVl4Source handle jack-clark. Links to https://jack-clark.net/2026/04/20/import-ai-454-automating-alignment-research-safety-study-of-a-chinese-model-hifloat4/.jack-clarkjack-clark.net"Import AI 454: Automating alignment research; safety study of a Chinese model; HiFloat4"https://jack-clark.net/2026/04/20/import-ai-454-automating-alignment-research-safety-study-of-a-chinese-model-hifloat4/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Apr/20/claude-token-counts/.simonwillison-2simonwillison.net"Claude Token Counter, now with model comparisons"https://simonwillison.net/2026/Apr/20/claude-token-counts/
5 sourcesAI Digest
AI's Control Layer
Practitioner discourse shifted toward the layer above the model: prompt policy, tool routing, evals, traces, and retrieval are increasingly where teams expect real leverage and real failures. The strongest signals treated orchestration and scoring surfaces as the actual product and governance lay...
agentsorchestrationevals
Source handle simonwillison. Links to https://simonwillison.net/2026/Apr/18/opus-system-prompt/.simonwillisonsimonwillison.netsimonwillisonhttps://simonwillison.net/2026/Apr/18/opus-system-prompt/Source handle youtube. Links to https://www.youtube.com/watch?v=xnG8h3UnNFI.youtubeyoutube.comyoutubehttps://www.youtube.com/watch?v=xnG8h3UnNFISource handle youtube-2. Links to https://www.youtube.com/watch?v=zZsTVBXcbow.youtube-2youtube.comyoutubehttps://www.youtube.com/watch?v=zZsTVBXcbow
3 sourcesAI Digest
Coding-Agent Friction Becomes a Feature
The clearest practitioner signal today is that strong coding-agent use now depends on deliberately preserving friction: explicit briefs, legible codebases, and real verification loops. The discourse is shifting from raw autonomy toward judgment-preserving workflow design, with permissions and pay...
coding-agentsworkflowverification
Source handle youtube. Links to https://www.youtube.com/watch?v=_Zcw_sVF6hU.youtubeyoutube.comyoutubehttps://www.youtube.com/watch?v=_Zcw_sVF6hUSource handle simonwillison. Links to https://simonwillison.net/guides/agentic-engineering-patterns/adding-a-new-content-type/.simonwillisonsimonwillison.netsimonwillisonhttps://simonwillison.net/guides/agentic-engineering-patterns/adding-a-new-content-type/Source handle magazine-sebastianraschka. Links to https://magazine.sebastianraschka.com/p/workflow-for-understanding-llms.magazine-sebastianraschkamagazine.sebastianraschka.commagazinehttps://magazine.sebastianraschka.com/p/workflow-for-understanding-llmsSource handle youtube-2. Links to https://www.youtube.com/shorts/7wmoB-J1fFE.youtube-2youtube.comyoutubehttps://www.youtube.com/shorts/7wmoB-J1fFE
4 sourcesAI Digest
Claude Code's New Default Posture
The strongest AI discourse signal was not a new benchmark winner but a workflow reset around coding agents: fuller delegation, deliberate effort settings, fewer interruptions, and explicit verification. Supporting evidence from Simon Willison and Uber suggests the durable shift is from model comp...
coding-agentsclaude-codeworkflows
Source handle claude. Links to https://claude.com/blog/best-practices-for-using-claude-opus-4-7-with-claude-code.claudeclaude.comclaudehttps://claude.com/blog/best-practices-for-using-claude-opus-4-7-with-claude-codeSource handle simonwillison. Links to https://simonwillison.net/2026/Apr/16/llm-anthropic/.simonwillisonsimonwillison.netsimonwillisonhttps://simonwillison.net/2026/Apr/16/llm-anthropic/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Apr/16/qwen-beats-opus/.simonwillison-2simonwillison.netsimonwillisonhttps://simonwillison.net/2026/Apr/16/qwen-beats-opus/Source handle uber. Links to https://www.uber.com/us/en/blog/ai-prototyping/.uberuber.comuberhttps://www.uber.com/us/en/blog/ai-prototyping/Source handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/uber-reveals-how-ai-is-transforming.departmentofproduct-substackdepartmentofproduct.substack.comdepartmentofproducthttps://departmentofproduct.substack.com/p/uber-reveals-how-ai-is-transforming
5 sourcesAI Digest
Bespoke AI Tools Are Still Winning
The clearest AI discourse signal today is that practical value is still arriving through small, custom tools built around real workflow friction. Simon Willison's Claude-built previewer is a strong example of how repository context plus a narrow task can produce durable operator leverage.
claudeworkflowinternal-tools
Read digest AI Digest
The Bottleneck Shifted to Control Surfaces
Today's practitioner discourse suggests the scarce asset is no longer raw model access but the layers that control how AI is steered and deployed. The strongest signals point to three leverage points: infrastructure coordination, prompt-shaped interfaces, and teams' ability to encode tacit standa...
control-surfacesagent-workflowsspeech-interfaces
Source handle dwarkesh. Links to https://www.dwarkesh.com/p/jensen-huang.dwarkeshdwarkesh.comdwarkeshhttps://www.dwarkesh.com/p/jensen-huangSource handle simonwillison. Links to https://simonwillison.net/2026/Apr/15/gemini-31-flash-tts/.simonwillisonsimonwillison.netsimonwillisonhttps://simonwillison.net/2026/Apr/15/gemini-31-flash-tts/Source handle youtube. Links to https://www.youtube.com/watch?v=2PWJu6uAaoU.youtubeyoutube.comyoutubehttps://www.youtube.com/watch?v=2PWJu6uAaoUSource handle exponentialview. Links to https://www.exponentialview.co/p/the-classified-frontier.exponentialviewexponentialview.coexponentialviewhttps://www.exponentialview.co/p/the-classified-frontier
4 sourcesAI Digest
AI discourse turns toward durability
The strongest discourse signal was a shift away from headline model comparisons and toward the economic and organizational durability of AI products. Even in a thin cycle, the most useful angle was adoption reality, operating cost pressure, and whether AI usage is becoming sticky enough to sustai...
economicsadoptionoperators
Source handle youtube. Links to https://www.youtube.com/watch?v=0vdlwOK_Qdk.youtubeyoutube.comyoutubehttps://www.youtube.com/watch?v=0vdlwOK_QdkSource handle youtube-2. Links to https://www.youtube.com/watch?v=YiitvyQGbkc.youtube-2youtube.comyoutubehttps://www.youtube.com/watch?v=YiitvyQGbkcSource handle simonwillison. Links to https://simonwillison.net/2026/Apr/13/steve-yegge/#atom-everything.simonwillisonsimonwillison.netsimonwillisonhttps://simonwillison.net/2026/Apr/13/steve-yegge/#atom-everything
3 sourcesAI Digest
Cheap AI output shifts the bottleneck again
Today's strongest AI discourse signal was not a new model or product launch. It was a multi-source correction to the way teams are currently operationalizing coding agents and "AI-first" org design. Across five distinct practitioner voices...
coding-agentssoftware-engineeringsecurity
Source handle simonwillison. Links to https://simonwillison.net/2026/Apr/13/bryan-cantrill/#atom-everything.simonwillisonsimonwillison.netsimonwillisonhttps://simonwillison.net/2026/Apr/13/bryan-cantrill/#atom-everythingSource handle bcantrill-dtrace. Links to https://bcantrill.dtrace.org/2026/04/12/the-peril-of-laziness-lost/.bcantrill-dtracebcantrill.dtrace.orgbcantrillhttps://bcantrill.dtrace.org/2026/04/12/the-peril-of-laziness-lost/Source handle youtube. Links to https://www.youtube.com/watch?v=I82j7AzMU80.youtubeyoutube.comyoutubehttps://www.youtube.com/watch?v=I82j7AzMU80Source handle jack-clark. Links to https://jack-clark.net/2026/04/13/import-ai-453-breaking-ai-agents-mirrorcode-and-ten-views-on-gradual-disempowerment/.jack-clarkjack-clark.netjack-clarkhttps://jack-clark.net/2026/04/13/import-ai-453-breaking-ai-agents-mirrorcode-and-ten-views-on-gradual-disempowerment/Source handle youtube-2. Links to https://www.youtube.com/watch?v=E1idsrv79tI.youtube-2youtube.comyoutubehttps://www.youtube.com/watch?v=E1idsrv79tISource handle youtube-3. Links to https://www.youtube.com/watch?v=zhXgkQ3nYeE.youtube-3youtube.comyoutubehttps://www.youtube.com/watch?v=zhXgkQ3nYeE
6 sourcesAI Digest
Where agent systems really win or lose
The strongest practitioner-level AI discourse in this cycle was not about a new frontier model. It was about where teams are likely to win or lose in the next phase of deployment: evaluation quality, agent governance surfaces, interface le...
agentsevalsenterprise-ai
Source handle youtube. Links to https://www.youtube.com/watch?v=X4dEHRzBLmc.youtubeyoutube.comyoutubehttps://www.youtube.com/watch?v=X4dEHRzBLmcSource handle youtube-2. Links to https://www.youtube.com/watch?v=VXfRt_H-V08.youtube-2youtube.comyoutubehttps://www.youtube.com/watch?v=VXfRt_H-V08Source handle nngroup. Links to https://www.nngroup.com/articles/ai-agents-as-users/.nngroupnngroup.comnngrouphttps://www.nngroup.com/articles/ai-agents-as-users/Source handle youtube-3. Links to https://www.youtube.com/watch?v=erV_8yrGMA8.youtube-3youtube.comyoutubehttps://www.youtube.com/watch?v=erV_8yrGMA8
4 sourcesAI Digest
Handmade design becomes an AI trust signal
Today's discourse signal was thin, and one item mattered much more than the rest: Nielsen Norman Group's argument that visibly handmade design is becoming a trust signal in an AI-saturated environment. The important shift is not aesthetic ...
uxtrustdesign
Source handle nngroup. Links to https://www.nngroup.com/articles/handmade-designs/.nngroupnngroup.comnngrouphttps://www.nngroup.com/articles/handmade-designs/Source handle youtube. Links to https://www.youtube.com/watch?v=_zdroS0Hc74.youtubeyoutube.comyoutubehttps://www.youtube.com/watch?v=_zdroS0Hc74
2 sourcesAI Digest
Claude Mythos changes security workflows
The dominant discourse signal this cycle is that Claude Mythos has done something qualitatively new: it moved named, senior security maintainers from skepticism to active engagement within weeks. Greg Kroah-Hartman now describes AI securit...
Source handle anthropic. Links to https://www.anthropic.com/glasswing.anthropicanthropic.comAnthropic — Project Glasswinghttps://www.anthropic.com/glasswingSource handle red-anthropic. Links to https://red.anthropic.com/2026/mythos-preview/.red-anthropicred.anthropic.comAnthropic Red Team — Mythos Previewhttps://red.anthropic.com/2026/mythos-preview/Source handle simonwillison. Links to https://simonwillison.net/2026/Apr/3/vulnerability-research-is-cooked/.simonwillisonsimonwillison.netSimon Willison — Vulnerability Research Is Cookedhttps://simonwillison.net/2026/Apr/3/vulnerability-research-is-cooked/Source handle natebjones. Links to https://www.natebjones.com/.natebjonesnatebjones.comNate B Joneshttps://www.natebjones.com/Source handle theo. Links to https://www.youtube.com/@t3dotgg.theoyoutube.comTheohttps://www.youtube.com/@t3dotggSource handle ai. Links to https://www.ai.engineer/europe/schedule.aiai.engineerAI Engineer Europe 2026 Schedulehttps://www.ai.engineer/europe/scheduleSource handle simonwillison-2. Links to https://simonwillison.net/2026/Apr/7/glm-51/.simonwillison-2simonwillison.netSimon Willison — GLM-5.1https://simonwillison.net/2026/Apr/7/glm-51/
7 sourcesAI Digest
Cheap generation forces a new operating model
The strongest AI discourse signal today is that the bottleneck has moved below the model and above the prompt at the same time. Builders are now arguing about execution substrates, workflow contracts, and product operating models more than...
Source handle youtube. Links to https://www.youtube.com/watch?v=7HP1jFJ9W1c.youtubeyoutube.comyoutubehttps://www.youtube.com/watch?v=7HP1jFJ9W1cSource handle youtube-2. Links to https://www.youtube.com/watch?v=TilDSWeiAlw.youtube-2youtube.comyoutubehttps://www.youtube.com/watch?v=TilDSWeiAlwSource handle departmentofproduct-substack. Links to https://departmentofproduct.substack.com/p/the-new-product-development-operating.departmentofproduct-substackdepartmentofproduct.substack.comdepartmentofproducthttps://departmentofproduct.substack.com/p/the-new-product-development-operating
3 sourcesAI Digest
The new layers builders must own for agents
The most useful AI discourse today asks a practical question: if agents are becoming real software systems rather than chat features, what new layers do builders now have to own? The strongest answers from the ledger point to four layers t...
Source handle simonwillison. Links to https://simonwillison.net/2026/Apr/5/research-llm-apis/.simonwillisonsimonwillison.netSimon Willison — research-llm-apishttps://simonwillison.net/2026/Apr/5/research-llm-apis/Source handle natebjones. Links to https://www.natebjones.com/.natebjonesnatebjones.comNate Joneshttps://www.natebjones.com/Source handle simonwillison-2. Links to https://simonwillison.net/2026/Apr/5/scan-for-secrets/.simonwillison-2simonwillison.netSimon Willison — scan-for-secretshttps://simonwillison.net/2026/Apr/5/scan-for-secrets/Source handle blockchain. Links to https://blockchain.news/zh/ainews/personal-ai-knowledge-bases-karpathy-highlights-farzapedia-s-file-first-personalization-approach-analysis-zh.blockchainblockchain.newsKarpathy on Farzapediahttps://blockchain.news/zh/ainews/personal-ai-knowledge-bases-karpathy-highlights-farzapedia-s-file-first-personalization-approach-analysis-zhSource handle cognitiverevolution. Links to https://www.cognitiverevolution.ai/.cognitiverevolutioncognitiverevolution.aiThe Cognitive Revolution — Joseph Nelsonhttps://www.cognitiverevolution.ai/Source handle exponentialview. Links to https://www.exponentialview.co/p/ev-568.exponentialviewexponentialview.coAzeem Azhar — The labs are rationinghttps://www.exponentialview.co/p/ev-568
6 sourcesAI Digest
What makes an agent trustworthy at work
Today's strongest AI discourse asks a more useful question than `which agent is best?`: what has to be true before an agent is trustworthy enough to become part of real work? Across builder essays, operator commentary, and human-centered c...
Source handle sebastianraschka. Links to https://sebastianraschka.com/.sebastianraschkasebastianraschka.comSebastian Raschkahttps://sebastianraschka.com/Source handle natebjones. Links to https://natebjones.com/.natebjonesnatebjones.comNate Joneshttps://natebjones.com/Source handle nngroup. Links to https://www.nngroup.com/.nngroupnngroup.comNielsen Norman Grouphttps://www.nngroup.com/Source handle departmentofproduct. Links to https://www.departmentofproduct.com/.departmentofproductdepartmentofproduct.comDepartment of Producthttps://www.departmentofproduct.com/Source handle techcrunch. Links to https://techcrunch.com/2026/04/04/anthropic-says-claude-code-subscribers-will-need-to-pay-extra-for-openclaw-support/.techcrunchtechcrunch.comTechCrunchhttps://techcrunch.com/2026/04/04/anthropic-says-claude-code-subscribers-will-need-to-pay-extra-for-openclaw-support/
5 sourcesAI Digest
When useful agents hit testing and rate limits
The strongest AI discourse in this window is about the operational consequences of agentic usefulness. Once agents are good enough to produce large amounts of code, the real constraints shift to testing, evaluation, fatigue, inspectable workflows, and metered access.
Source handle simonwillison. Links to https://simonwillison.net/.simonwillisonsimonwillison.netSimon Willisonhttps://simonwillison.net/Source handle karpathy. Links to https://karpathy.ai/.karpathykarpathy.aiAndrej Karpathyhttps://karpathy.ai/Source handle theo. Links to https://www.youtube.com/@t3dotgg.theoyoutube.comTheohttps://www.youtube.com/@t3dotggSource handle ai-google. Links to https://ai.google.dev/.ai-googleai.google.devGemini API Docshttps://ai.google.dev/
4 sourcesAI Digest
From coding assistant to agent system
The highest-signal AI developments in the last 24 hours point to a rapid shift from single-shot coding assistants toward structured agent systems with explicit research, planning, and live-documentation phases.
agent-systemsdeveloper-toolsexecution-design
Source handle github. Links to https://github.blog/.githubgithub.blogGitHub Bloghttps://github.blog/Source handle google-github. Links to https://google.github.io/adk-docs/.google-githubgoogle.github.ioGoogle ADK Docshttps://google.github.io/adk-docs/Source handle ai-google. Links to https://ai.google.dev/gemini-api/docs/mcp.ai-googleai.google.devGemini API MCP Docshttps://ai.google.dev/gemini-api/docs/mcp
3 sources