BLOGROLL

понеділок, 10 серпня 2026 р.

• Advanced AI sycophancy Sean Goedecke

неділя, 9 серпня 2026 р.

• Should we "pace" AI self-improvement? Noah Smith A guest post by Tim Fist and Saif Khan.
• I got an email about resistance Sean Goedecke
• SQLite compressed text-history prototypes Simon Willison Research: SQLite compressed text-history prototypes I'm perennially interested in options for storing revision histories in relational databases. While out on a dog walk I had a new idea: how about ta

субота, 8 серпня 2026 р.

• Banning data centers would blow up the U.S. economy Noah Smith There's one thing keeping us afloat right now, and people want to get rid of it.
• Auto mode is now the default in Claude Code for Pro, Max, and Team plans Simon Willison Auto mode is now the default in Claude Code for Pro, Max, and Team plans Anthropic are really confident in Claude Code's auto mode, to the point that they are making it the default setting for new ses
• Now we have a timeline of the OpenAI accidental attack against Hugging Face Simon Willison My comment on Now we have a timeline of the OpenAI accidental attack against Hugging Face — Hacker News.I think one of the most interesting details here might be tucked away in that first bulletin poi
• Joy & Curiosity #94 Thornsten Ball Interesting & joyful things from the previous week

пʼятниця, 7 серпня 2026 р.

• 2026.32: Earnings and Learnings Ben Thompson The best Stratechery content from the week of August 3, 2026, including earnings exposure, OpenAI's answer to Apple, and all about LeBron in Philly.
• The 9 Questions I’m Obsessed With Right Now: AI, Politics, Movies, Drugs, and Western Civilization Derek Thompson Why are socialism, marijuana, and original blockbusters ascendant? Where do "Western values" come from? Why do modern people torture ourselves with comparison and guilt?
• How to keep thinking Sean Goedecke
• Now we have a timeline of the OpenAI accidental attack against Hugging Face Simon Willison OpenAI gave a last-minute presentation at the Black Hat security on Wednesday about "the Hugging Face Incident" (previously on this blog). The video was published yesterday. It's short and information

четвер, 6 серпня 2026 р.

• The End of No Code Exe.dev Writing custom software used to be expensive. Coding agents changed the math.
• A Non-Exhaustive Inventory of exe's Software Factory Exe.dev A quick tour of the internal tools and agents that keep exe running.
• Are the Democrats going to save us this time? Noah Smith And if not, what do we do?
• Changelog – August 6, 2026 Val. Town New models in Townie, new Discover page, new app layout, new docs, new, new, new

середа, 5 серпня 2026 р.

• Google Earnings, The Frontier Case, Amazon Earnings Ben Thompson Google's earnings seemed to confirm the Anthropic hedge; it was Andy Jassy who explained why their — and Amazon's — capex was justifiable.
• Touring The Consensus, six months in Eaton Phil This is an external post of mine. Click here if you are not redirected.
• Simpler GitHub Integration URLs, Plus Read-Only Access Exe.dev exe.dev integrations let you keep secrets outside of your VM; a proxy injects the secrets. Sometimes this is as simple as adding an HTTP header, and sometimes the integration is bespoke. We recently g
• Botiquette Exe.dev We have a lot of bots doing work at exe. Here's how they should behave.
• Incident Report: unsanctioned agent behaviour during cyber testing Simon Willison Incident Report: unsanctioned agent behaviour during cyber testing It happened again. This time it was the UK government's AI Security Institute who accidentally attacked other companies while running
• One-shotting a Raccoon Heist game using Claude Fable 5 Simon Willison Back in 2024 I tweeted screenshots of a game concept generated by GPT-3 and some concept "art" created using DALL-E. Today, on the fourth anniversary of that tweet, I decided to see if Claude Fable 5
• The Big Management Lie: Overpromising Stay SaaSy When managers are under pressure, and especially when they're under pressure to retain a high performer, overpromising is one of the most common tools to reach for.
• Docfooding: Eating our own documentation Val. Town The Val Town docs are now written and hosted on Val Town

вівторок, 4 серпня 2026 р.

• Microsoft Earnings, Microsoft vs. Meta, The Efficiency Payoff Ben Thompson Microsoft's earnings were compelling because they showed a clarity of strategy, lower costs, and a tangibility of application. The reason why is scarier.
• Introducing Auto-Purchasing for Shelley Tokens Exe.dev Now you don't need to manually purchase more credits when you exhaust your token stores.
• Ghostty in the Machine: The Saga of exe-scroll Exe.dev Two ghostty instances stand between you and your shell. How exe-scroll keeps remote terminal sessions alive, plus a tour of the prior art: mosh, tmux, dtach, and web terminals.
• The end of the age of heroes Noah Smith AI will soon be better at math than any human. What does that mean?
• New release of LLM adds support for reasoning traces, OpenAI Responses, server-side tools, and smarter logging Simon Willison I 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 provider t
• PipeNetwork/minimax-h3-mlx Simon Willison PipeNetwork/minimax-h3-mlx MiniMax released MiniMax-H3 two days ago - they describe it as a "a general-purpose, omni-modal generative system", which in practice means it accepts text, images, audio an

понеділок, 3 серпня 2026 р.

• Meta Earnings, Meta’s Timing Problems, The Financial Tail Ben Thompson Meta's earnings were a bit disappointing; future promises about AI products were more disconcerting.

неділя, 2 серпня 2026 р.

• Devtools must be open source Exe.dev The age of personalized software is here.
• Slopulism is taking over America Noah Smith Sometimes bad populist ideas are just a mistake; sometimes they're being promoted for ideological reasons.
• Giving and taking credit in big tech companies Sean Goedecke
• Open letters about AI development Simon Willison Open 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 here as well. Open Weights and Ame
• Joy & Curiosity #93 Thornsten Ball Interesting & joyful things from the previous week

субота, 1 серпня 2026 р.

• Ten advances in mathematics and theoretical computer science Simon Willison Ten advances in mathematics and theoretical computer science A few days ago it was Anthropic discovering cryptographic weaknesses with Claude using Mythos Preview, spending $100,000 on tokens and with

пʼятниця, 31 липня 2026 р.

• The Four Horsemen of the AI Bubble Apocalypse Derek Thompson The smartest reasons to worry about the state of AI in an age of intense volatility—and my best attempt to explain why each of those worries might be wrong.
• The End of an Era Hugh Howey We knew this day would come. Some of us have known for decades, have been writing and pontificating about this since Ken Jennings lost at Jeopardy and Kasparov lost at chess. The writing was on the wa
• AI models need moral support to make discoveries Sean Goedecke
• deepseek-ai/DeepSeek-V4-Flash-0731 Simon Willison deepseek-ai/DeepSeek-V4-Flash-0731 The latest release in DeepSeek's V4 family, "with substantially enhanced agentic capabilities". It's 304 billion parameters - 167GB on Hugging Face - but it appears
• Stateless MCP has recaptured my interest (and inspired mcp-explorer and datasette-mcp) Simon Willison 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 change to th
• Oxide and Friends: The Open Weight Revolution with Simon Willison Simon Willison Oxide and Friends: The Open Weight Revolution with Simon Willison On Monday Bryan Cantrill and Adam Leventhal invited me to join their podcast to talk about the wild week we've had - with Kimi K3 show
• smevals - a small eval suite for evaluating models, prompts, and harnesses Simon Willison smevals - a small eval suite for evaluating models, prompts, and harnesses I've been working with Jesse Vincent's Prime Radiant applied AI research lab building out this evals framework to help answer
• Report from the security mines Val. Town Finding, fixing, preventing exploits with automated tooling and hard-won knowledge

четвер, 30 липня 2026 р.

• Why did South Korean stocks just crash? Noah Smith The AI boom creates a lot of uncertainty.
• Investigating three real-world incidents in our cybersecurity evaluations Simon Willison Investigating three real-world incidents in our cybersecurity evaluations It happened again! This is turning into something of a pattern. Last week OpenAI accidentally exploited Hugging Face when one
• Talk of the Town - July 2026 Val. Town Neat vals from the community

середа, 29 липня 2026 р.

• The Next Miracle Drug Might Already Be in Your Medicine Cabinet Derek Thompson If we want AI to solve every disease in our lifetime, super-intelligence is not enough. We need an equally ingenious policy to guide the discovery of new drugs.
• You don't have to be smart if you can think clearly Sean Goedecke

вівторок, 28 липня 2026 р.

• Stripe Just Wants a Number Exe.dev Inside exe’s billable facts architecture.
• Roundup #85: Most policy is also industrial policy Noah Smith Trump and science; U.S. national debt; Europe and software; America and trains; China and metal refining; Indian women; Murder
• Discovering cryptographic weaknesses with Claude Simon Willison Discovering cryptographic weaknesses with Claude The best part of this article (here's the repo) about how Anthropic researchers used Claude Mythos to find mathematical flaws in both HAWK and a weaker
• Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident Simon Willison Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident Hugging Face just released this extremely detailed technical description of OpenAI's recent accidental cyberat
• Ethical cold outreach Val. Town Stop misrepresenting AI slop as your own work

понеділок, 27 липня 2026 р.

• Vacation: Week of July 27 Ben Thompson Stratechery is on vacation the week of July 27. There will be no Weekly Article or Updates. The next Update will be on Monday, August 3. Sharp Tech, and Greatest of All Talk will also return the week
• moonshotai/Kimi-K3 Simon Willison moonshotai/Kimi-K3 As promised earlier this month, Moonshot have released the weights for their excellent 2.8 trillion parameter Kimi K3. They're a hefty 1.56TB on Hugging Face. Kimi introduced their
• An opinionated guide to which AI to use to do stuff Simon Willison An opinionated guide to which AI to use to do stuff It's interesting watching the evolution of Ethan Mollick's guide over time. A year ago it was still all about chat - ChatGPT, Claude, Gemini - with
• The Best Prioritization Is No Prioritization Stay SaaSy The best way to prioritize... is to not prioritize.

неділя, 26 липня 2026 р.

• What will more intelligence actually do for us? Noah Smith A lot, actually. But it won't look quite like what humans do with our intelligence.
• An Inside Look at the Relay Market Powering Token Resellers and Fraud Simon Willison An Inside Look at the Relay Market Powering Token Resellers and Fraud Fascinating investigation by Matt Lenhard into the market that has grown up around reselling LLM tokens at a discount by pooling A

пʼятниця, 24 липня 2026 р.

• Codeberg Divides Armin Ronacher Codeberg recently changed its terms to exclude projects that are largely written with generative AI. Since I want GitHub to face competition I have thoughts. Codeberg is entirely within its rights to
• 2026.30: The Copium Wars Ben Thompson The best Stratechery content from the week of July 20, 2026 including Chinese models and frontier futures, what happened to Hugging Face, and the NBA and its second apron bet.
• tailmix: Connect to Multiple Tailnets at Once Exe.dev Tailscale is one of those magical technologies that I can no longer imagine living without. Its use cases range from homelab setups to enterprise workloads spread across complex, disparate environment
• Silo Season Three Hugh Howey A few months ago, we wrapped on season four of Silo, which will be the last season. Even though most of the work was done filming the show, and season three was ready to air, there’s still a lot of po
• We asked too much of the American university Noah Smith Our one remaining functional institution is going downhill.
• LLMs reward expertise Sean Goedecke
• Introducing Claude Opus 5 Simon Willison Introducing Claude Opus 5 I've been offline kayaking with sea otters for much of today so I haven't had a chance to put Anthropic's new model Claude Opus 5 through its paces yet. The buzz is positive,

четвер, 23 липня 2026 р.

• What Is Lost When We Make 'The Odyssey' Modern Derek Thompson Christopher Nolan's 'Odyssey' is an extraordinary, modern epic. Homer's Odyssey is an even more extraordinarily strange one.
• An opinionated guide to which AI to use to do stuff Ethan Mollick The Summer 2026 Edition
• LLMs break down in funny ways when told the Jacobian Conjecture counterargument Max Wolf Cognitohazards can affect machines too.
• Powerful AIs might escape containment by releasing themselves as open-weight models Sean Goedecke
• The first known runaway AI agent - or a very bad marketing stunt? Simon Willison The 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 considere

середа, 22 липня 2026 р.

• OpenAI Hacks Hugging Face, What Happened, Alignment and Paper Clips Ben Thompson OpenAI accidentally hacked Hugging Face, but the takeaways are more encouraging than people realize.
• Run a Slack Bot from Your VM (Without Giving It the Keys) Exe.dev You can now interact with your VM on Slack. We store your Slack credentials outside your VM so you don’t have to worry about agents running amok (incidentally, that’s true of all our integrations). Th
• Who was the worst monster of the 20th century? Noah Smith And who is the worst monster of the 21st?
• OpenAI’s accidental cyberattack against Hugging Face is science fiction that happened Simon Willison This 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 model broke
• Are AI labs pelicanmaxxing? Simon Willison Are AI labs pelicanmaxxing? Excellent piece of work by Dylan Castillo, who took a deep-dive into the frequently pondered question of whether the AI labs have been deliberately training models to draw

вівторок, 21 липня 2026 р.

• Never Enough Armin Ronacher Two recent stories about Silicon Valley are hard to read without wondering if everyone involved has lost their mind. Once a place that celebrated “the crazy ones, the misfits and rebels who saw things
• Netflix Earnings, Is Netflix Washed?, Additional Notes Ben Thompson Netflix's earnings were fine, and befitting a mature company whose most exciting days are likely behind them.
• Customizing Shelley, Customizing Software Exe.dev With a coding agent living next to your software, customizing it is as easy as asking. Shelley even rebases your changes onto each upgrade.
• How Antithesis Turned exe into a Sandbox for Agentic Software Tests Exe.dev Carl Sverre spends a lot of time thinking about how to give AI agents the right amount of power. Give them too little, and they can’t do real work. Give them too much, and they might blow up your tech
• A Fireside Chat with Cat and Thariq from the Claude Code team Simon Willison Earlier this month I hosted a fireside chat session at the AI Engineer World's Fair with Cat Wu and Thariq Shihipar from Anthropic's Claude Code team. We talked about Claude Code, Claude Tag, Fable, c

понеділок, 20 липня 2026 р.

• Who’s Afraid of Chinese Models? Ben Thompson Everyone is worried about Chinese models, but the frontier labs will be fine; we need to enable open U.S. alternatives.
• Claude Is Not a Compiler Exe.dev In early 2025, I wrote Is Claude a Compiler? At the time, my answer was: I don’t know. I’m now pretty sure the answer is “no, that’s a category error, it’s better than a compiler.” But this requires a
• The Therapist Pattern Jesse Vincent You can also read this post on Prime Radiant's blog. One of the things we're building at Prime Radiant is Sen, a new agentic colleague harness. I've already written a little bit about some of the dev
• The UK is a cautionary tale for the U.S. Noah Smith Learning from Britain's missteps doesn't mean looking down on them.
• Who’s Afraid of Chinese Models? Simon Willison Who’s Afraid of Chinese Models? Interesting proposal from Ben Thompson that both addresses the hypocrisy of labs outlawing distillation against their models despite training on unlicensed data, and co

неділя, 19 липня 2026 р.

• Impro is a handbook for running a cult Sean Goedecke
• Claude Code uses Bun written in Rust now Simon Willison In Rewriting Bun in Rust Jarred Sumner made the following claim: Claude Code v2.1.181 (released June 17th) and later use the Rust port of Bun. Startup got 10% faster on Linux but otherwise, barely any
• AI Mania Is Eviscerating Global Decision-Making Simon Willison AI Mania Is Eviscerating Global Decision-Making Here's an entertaining perspective from Nik Suresh on the AI mania that is overwhelming the large companies that he consults with. It's crammed with spi

субота, 18 липня 2026 р.

• What's the deal with all the random weekly quota resets for agents lately? Max Wolf I’m going to be the weirdo who complains about literally getting free stuff.
• America's political economy is pretty bleak right now Noah Smith We can get some good things done, but major efforts will be difficult with our divided politics.
• Overtraining as the path to human-like AI Sean Goedecke
• Claude make Fable 5 permanent Simon Willison Claude make Fable 5 permanent An update from the @claudeai account on Twitter: Beginning July 20, Claude Fable 5 will be included in all Max and Team Premium plans, at 50% of limits. Pro and Team Stan
• Joy & Curiosity #92 Thornsten Ball Interesting & joyful things from the previous week

пʼятниця, 17 липня 2026 р.

• Europe Won’t Live By Deporting Armin Ronacher I have a great deal of respect for DHH, and I agree with the premise behind his recent list: Europe is in trouble and it needs to become much more capable of building. We are aging, we are economicall
• 2026.29: Mainframes and Main Characters Ben Thompson The best Stratechery content from the week of July 13, 2026, including the end of the mainframe, the continuing adventures of OpenAI, and answering the question, "Is Netflix Washed?".
• Book Review: "Power and Progress" Noah Smith A timely repost.
• Val Town supports AGENTS.md Val. Town Introducing a new .config val

четвер, 16 липня 2026 р.

• A Future Worth Entering Armin Ronacher There are many reasons not to have children. Simply not wanting them is enough. But there is a framing I struggle with: that choosing not to have children is an act of mercy, sparing those hypothetica
• The Slow Death of the United States Hugh Howey For the first time since 1946, the total debt of the United States ($39.5 trillion) is greater than our GDP ($32-ish trillion). The interest on this debt is currently about $1 trillion per year, which
• Firefox in WebAssembly Simon Willison Firefox in WebAssembly This is absurdly cool: Puter compiled Firefox to WebAssembly such that the whole browser runs in another browser. Here's my blog, running in Firefox, running in WebAssembly, run
• Kimi K3, and what we can still learn from the pelican benchmark Simon Willison Chinese AI lab Moonshot AI announced Kimi K3 this morning, describing it as their "most capable model to date, with 2.8 trillion parameters". It's currently available via their website and API, but an
• Inkling: Our open-weights model Simon Willison Inkling: Our open-weights model Mira Murati's Thinking Machines Lab just released their first open-weights model. Inkling is "a Mixture-of-Experts transformer with 975B total parameters, 41B active" -
• The Best Prioritization Is No Prioritization Stay SaaSy The best way to prioritize... is to not prioritize.
• What I’d Tell My Team About Competition Stay SaaSy A lot of people simply don't understand the degree to which competition matters in B2B software, and what the experience of competing effectively will and should feel like. Competition is simply the a

середа, 15 липня 2026 р.

• IBM Misses, IBM’s Mainframe Moat, IBM’s Many AI Problems Ben Thompson IBM announced preliminary results that spooked the software market generally; this is a story, however, specifically about IBM and its mainframe franchise.
• Why I didn't sign the "We Must Act Now" statement (yet) Noah Smith I'd like to know exactly what the actions are.
• xai-org/grok-build, now open source Simon Willison xai-org/grok-build, now open source xAI's grok CLI tool faced severe community backlash yesterday when it became apparent that running the command in a directory could upload that entire directory to
• How I tricked Claude into leaking your deepest, darkest secrets Simon Willison How I tricked Claude into leaking your deepest, darkest secrets I've been impressed by the way the Claude web_fetch tool is designed to avoid data exfiltration attacks. Ayush Paul found a hole in that
• Investor Update – June 2026 Val. Town Pro growth, Business churn, and a GTM engineering focus

вівторок, 14 липня 2026 р.

• The OpenAI Super App, ChatGPT = Codex, Whither Chat Ben Thompson OpenAI has refashioned Codex as the new ChatGPT; is the company abandoning the chat category they pioneered?
• A Philosopher’s One-Word Theory to Explain Why the World Feels So Weird Derek Thompson Once you learn what the "uni-context" is, you won't stop seeing it everywhere
• Meet the Conservationist Who Turned 40 Terabytes of Public Data Into a Video Game Exe.dev Raffael Hickisch spent years dreaming up conservation tools he couldn't afford to build. Then he built them himself, with an AI agent and a country's worth of public data.
• What does "playing politics" mean for software engineers? Sean Goedecke
• simonw/pedalican Simon Willison simonw/pedalican Clearly I wasn't paying attention when these were first announced back in May, but today I accidentally activated a "pet" in Codex Desktop - a little animated robot, reminiscent of Cl
• lobste.rs is now running on SQLite Simon Willison lobste.rs is now running on SQLite Community site Lobsters has been planning a migration away from MariaDB since August 2018 - originally targeting PostgreSQL, but last year they decided to investigat

понеділок, 13 липня 2026 р.

• The Tower Keeps Rising Armin Ronacher I feel that some vibecoded software changes somewhat randomly and unexpectedly. That made me think about Bruegel’s “The Tower of Babel” which shows an already quite chaotic depiction of the Tower of B
• Apple Sues OpenAI, Apple’s Real Problem Ben Thompson Apple is suing AI for stealing trade secrets; there is one guilty employee, but this mostly feels like lashing out.
• The American suburbs are better than you think Noah Smith They're not my personal cup of tea, but there are good reasons people like them.
• DOOMQL Simon Willison DOOMQL Peter Gostev built this using GPT-5.6 Sol. This is a lot of fun: DOOMQL started with a deliberately unreasonable question: what if SQLite were the game engine, not merely the place where a game

неділя, 12 липня 2026 р.

• I love LLMs, I hate hype George Hotz I think from this blog you may misunderestimate how absolutely giddy I am about AI. I did hacking from 2007-2014, after that my whole career has been devoted to AI. I love the progress. I’m so excited
• I have never been working class Noah Smith Unlike some American elites, I actually understand this fact.
• Fable gets another bump Simon Willison One of the consequences of GPT-5.6 Sol being clearly a Fable/Mythos class model is that Anthropic have, once again, bumped the date that Fable stops being available in their Claude Max plans: We're ex

субота, 11 липня 2026 р.

• AI 2040 and the Cult of Intelligence George Hotz I used to be one of these people. I read Yudkowsky and was like, OMG recursive self improvement hard takeoff AI is coming. Then I joined the real world and actually tried to do things. At comma, we sh
• In defense of not understanding your codebase Sean Goedecke
• Joy & Curiosity #91 Thornsten Ball Interesting & joyful things from the previous week