Roose and Newton fact-check Zuckerberg's 6,500-word AI manifesto — policy wishlist dressed as optimism, with a dragon-to-everyone critique — then Pangram CEO Max Spero explains how AI text detection finally got good, and the hosts debut Running the Numbers with a Riemann hypothesis side problem solved via positive affirmation.
The Zed team released Delta, a multiplayer environment purpose-built for coding with AI agents — no collapsed diffs, no truncated transcripts, every annotation anchored to evolving code in real time.
AI removes the speed limit on implementation — bad decisions compound faster than ever, and judgment replaces coding speed as the scarce skill.
AI infrastructure consumes freshwater via cooling, electricity, and chip manufacturing — projected 4.2–6.6 billion m³/year by 2027, mostly in water-stressed regions.
The collapse of search is not a skills problem — it is the infrastructure that stored the public record, breaking down.
Dillon Mulroy, Rhys Sullivan, and Sunil Pai discuss AI burnout, lost collaboration, and the parts of making software that still bring joy.
Prompt-level instructions to make LLMs sound human are the wrong abstraction — preserve high-fidelity machine state and transform only at the boundary.
Claude failed to prove the Riemann hypothesis — and as a byproduct improved the bound on zeros on the critical line from 41.6% to 67.2%.
Free AI legal advice is flooding Britain's employment tribunals with hallucinated claims — a textbook commons collapse where too much access to justice means justice for no one.
A political-philosophy case against generative AI as it already exists — not another industrial revolution, but the end of the human relationship to language, capitalism, and individualism.
A conceptual paper reframing AI deskilling as a collective action problem — each organization rationally depletes the shared pool of expertise every org depends on to validate AI output.
A sharp defense of programming as genuine craft in an AI era — pushing back against the dismissive "code was never the hard part" trope that treats implementation as trivial.
AI is accelerating an existential crisis among knowledge workers — not by replacing them, but by making the spectacle of Workism too visible to ignore.
74% of executives trust AI advice over colleagues, and 44% would defer to a chatbot over their own judgment — Kinley names the three symptoms of unchecked AI deference.
AI makes building software as easy as frying a steak — and just as likely to hand you charcoal, confidently presented as medium-rare.
A complete Product OS walkthrough: define, design, develop, and distribute a real software product with Claude Code — with four case studies of non-technical founders making real money.
When AI makes everything cheap, taste — the wordless verdict that two of three plausible versions are wrong — becomes the only scarce skill.
In hobby dev communities the process of mastery is the product — an LLM that generates the finished piece just robs you of the craft.
An AI image on a personal blog is a trust signal that fails — it makes readers wonder if the words were machine-written too.
Stop relaying AI output verbatim in Slack, PR threads, and group chats — read it, validate it, write your own words.
Tailscale co-founder David Crawshaw: the source code is the extension system — AI agents make personalizing software cheaper than configuring it.
Diogo Almeida argues that pleasing a human and running reliable software are different objectives. His case for moving beyond AI assistance—and why he says Jev's approach is not simply RLVR.
Dillon Mulroy shows how Pi, Plannotator, and type-driven plans keep him involved in agent-written software.
The anti-generic Claude Design workflow: prepare three files (design.md → design system → template) so every deck starts on-brand and every feedback compounds into future designs.
Full Claude Design platform walkthrough: expanded templates, bundled usage, design systems for on-brand output, and a motion-graphics-for-videos use case that is genuinely surprising.
Lucas Rocha, an engineer at Brazilian edtech Nova Escola and alum of Hamel Husain's AI Evals course, tells his evals rollout story backwards: from a team that now runs LLM judges with no tickets and no engineer in the loop, back to the messy first spreadsheet and the three barriers — data, methodology, and judgment — that every team hits along the way.
Pi's creator joins a discussion of broken tool calls, model regressions, measurable agent loops, and the cost of renting intelligence.
Abi Aryan breaks down where LLM latency actually goes: the compute-bound prefill phase vs the memory-bandwidth-bound decode phase, and why decode — not prefill — dominates latency and cost, especially in multi-step agentic systems. A live Colab walkthrough with TinyLlama on a T4 measures both phases across chat, RAG, creative, and agentic request shapes.
Before worrying about evals, ask if you are building the right thing: AI that dumps an unverifiable answer creates more work. Hamel walks through design patterns — provenance, progressive disclosure, expert signals, modular steps — across three worked examples.
Dillon Mulroy, Rhys Sullivan, and Sunil Pai argue for products that serve a customer's chosen agent. Their examples connect APIs, documentation, review loops, cloud environments, and burnout.
Hamel argues that hiring a consultant to implement AI and walk away is almost always a mistake — the underlying problems remain and the AI itself is designed to remove the middleman. Consultants work only for education or when they have skin in the game.
When a product seems impossible to eval, that is usually a sign the product itself is weak. Hamel suggests infusing user-facing signals into the product so users — and you — can assess quality.
Dillon Mulroy, Rhys Sullivan, and Sunil Pai ask why more generated code has not produced better software. They compare architecture, token costs, model access, and autonomous loops.
Hamel Husain counters the 'evals take too long' objection with one principle: if you can't eval your product, your users can't either — so evals are how you figure out how to convey trust in AI answers.
Hamel Husain explains how to scientifically measure an LLM judge's noise against ground-truth human labels — and why every LLM judge is a black-box classifier with edge cases, so you have to decide how much noise is tolerable.
Hamel Husain argues that ROUGE/BLEU string-similarity scores and 1-to-10 rating scales are holdovers from traditional NLP that make evals harder to act on — nobody knows what a 4.2 means, and binary pass/fail evals are what actually let you ship.
Hamel Husain explains why he converts every 1-to-5 rating scale eval into binary pass/fail checks, scoping each failure into a specific, actionable criterion instead of hiding noise behind middle values.
Wolfe builds a linked Markdown wiki, then adds journal responses, contact records and scheduled ingestion while correcting the agent’s filing rules.
Anna Gutowska and Martin Keen explain agent discovery, task updates, and tool interfaces, then combine A2A and MCP in a retail inventory example.
A research sampler broadens AI-generated ideas by scoring plausibility separately from familiarity. The evidence shows a trade-off, not guaranteed discoveries.
Claire Vo builds reusable Claude Skills in Cursor, validates their files, tests a changelog newsletter, and fixes a naming error during web upload.
Colin Matthews turns screenshots and browser elements into reusable components, preserves a baseline, and forks prototypes before exploring new ideas.
Ryan Carson uses a product brief, a task list, and approval checkpoints in Cursor, then adds database tools, browser control, and explicit context selection.