202636 entries
October2 posts/1 link/1 release
Learning to let go in the age of AIslk 0.29.0A useful personal agentWhat would a useful agent be like?September1 post/2 links
Unbounded: Progress, Economic Growth, and SocietyThere usually isn't a correct answerWriting a Minecraft server in Clojure: it was funAugust2 posts/1 note/11 links/2 tools
AI "features" and integrations should remove friction, not add itPreviewing the Model Hardware StandardDebates over AI consciousness are a trapAI coding tools unlock small softwareTurn And Face The StrangeReading Code Considered HelpfulFeed PreviewAI is removing the middle class of software engineeringThere are no lossless transformations of natural-language textLet's stop talking about A.I.Design Ethical AI Rewards and Policies“Code was never the hard part” is an insult to all programmersIs it all just vapourware?Find accessible colours for the webHave the models come full circle?Software Should Work ConferenceJuly2 notes/7 links/1 quote/1 tool
Eval-driven development: Lessons from evaluating GenAI at scaleLearning new things is hardAI is ruining self-learningSecurity incident disclosure — July 2026Annotation toolIntelligence isn't trustAI Mania Is Eviscerating Global Decision-MakingLinus Torvalds on LLM usage in kernel development12-Factor Agents - Principles for building reliable LLM applicationsAgentic Continuous Delivery (ACD)Zig Creator Calls Spade a Spade, Anthropic Blows SmokeMay1 post
Accepting my AI-centric fateJanuary1 link
Spec-ulation