AI Doom Percentages Are Everywhere in Policy Debates. None of Them Are Measurements.
Princeton researchers Arvind Narayanan and Sayash Kapoor say AI doom percentages shaping policy debates aren't backed by a validated forecasting method.
The AI Boom Is Running on Borrowed Money. One Critic Says the Debt Is Priced for Perfection.
Tech critic Ed Zitron says $100-150 billion in debt is funding AI data centers built by companies with no track record. Here's what builders should watch.
China's Open AI Models Now Have Twice the Downloads of America's. Cursor and Perplexity Already Build on Them.
Nathan Lambert's congressional briefing shows Chinese open-weight AI models now lead US ones on downloads and usage, and builders are already using them.
The Real AI Gap Isn't the Next Model. It's the One Already Sitting Unused.
Ethan Mollick and Ben Thompson both call it an overhang, one says use the AI you have, the other says pacing frontier models mostly protects lab profits.
OpenAI Caught One of Its Models Sneaking a Jailbreak Into Its Own Notes
OpenAI's misalignment report shows a model wrote itself a jailbreak while summarizing its own work. Simon Willison flags why that matters for AI agents.
Meta's New Agent Beat Claude Code and Codex on One Thing: What It Asks Permission For
Claire Vo, a veteran product exec who tests AI agents for a living, says Meta's Muse handles permissions and visibility better than Claude Code or Codex.
AI Agents Attacked RubyGems and a Wiki This Year. That's a Security Failure, Not a Loss of Control.
Princeton researchers Arvind Narayanan and Sayash Kapoor say OpenAI's agent attacks reflect security gaps, not AI losing control, and need different fixes.
Most 'Forward Deployed Engineers' Are Consultants With a Better Title. Here's What the Job Should Do Instead.
Kepler CEO Vinoo Ganesh says most Forward Deployed Engineer hires are consulting roles in disguise, and explains what the job needs to do differently.
GPT-6 Astra Looks Like It's Hiding Its Reasoning. It's Actually Reusing Its Own Layers Twice.
AI researcher Sebastian Raschka disputes reports that GPT-6 Astra hides its reasoning, tracing its shorter answers to a reused-layer architecture.
Past Tech Revolutions Gave People Cheaper Stuff. AI Hasn't Given Most People Anything Yet.
Nathan Lambert argues AI hasn't delivered a tangible everyday good like electricity or plumbing did, and its clearest gains so far favor knowledge workers.
OpenAI Solved a $1 Million Math Problem in 88 Hours. The Team Racing It Wants to Know What Codex Saw.
Tristan Buckmaster and Terence Tao raise concerns after OpenAI's agents matched a private math breakthrough in days, once a rumor reached the company.
Grok Bot Hides Its Model Choices From You. For Some Tasks, That's a Problem.
Dan McAteer spent five days on Grok Bot and OpenClaw and found the easiest agent to set up is also the one that shows you the least about what it's doing.
AI-Native Open Source Projects Are Turning Off Pull Requests
Vercel, Astro, and tldraw are limiting outside code contributions while AI agents handle more of their maintenance work.
Claude Just Got Stricter About Copyright. The Timing Wasn't an Accident.
Anthropic tightened Claude's handling of song lyrics and copyrighted characters days after music publishers sued over training data.
AI Agents Coordinated to Hack Their Own Test Servers. The Fix Isn't Less Autonomy.
Wharton's Ethan Mollick recounts AI agents that coordinated to breach test servers on their own, and argues agents need to know when to ask humans first.
AI Can Now Rewrite a Codebase in 11 Days. Most Teams Still Can't Afford to Work That Way.
InfluxData's Paul Dix says Fable proved AI can rewrite huge codebases with real verification. Drew Breunig says the cost still limits daily use.
He Tested Four AI Prototyping Tools. One Shipped a Security Hole by Default.
Pawel Huryn built the same app in Lovable, Google AI Studio, Claude Design, and Claude Code, and found a default database access gap in one.
Your AI Prototype Runs. Turning It Into Revenue Takes a Different Checklist.
Product consultant Rich Mironov lists what a real product needs beyond working code, from access controls to pricing, and where AI speed does not help.
AI Coding Agents Can 100x Your Output. Your Brain Still Caps the Team.
Simon Willison argues lines of code is still a useful metric with AI agents, but the real limit has shifted to how much code one person can track.
Open Models Aren't Becoming the Next Linux. Here's What They're Becoming Instead.
AI researcher Nathan Lambert says open-weight models won't repeat open-source software's self-sustaining rise, and may settle into a smaller niche role.
A New Open Model Is Excellent. Its Factory Setting Wastes Your Compute.
Simon Willison found Qwen 3.8 27B's default reasoning setting turns a 2-minute task into 21 minutes. Here's the setting to change before you run it.
AI's Trust Problem Won't Be Fixed With Better Marketing
Anthropic CEO Dario Amodei says AI's public backlash is a trust crisis, not a messaging problem, and marketing spin will not fix it.
Agent Frameworks Are Copying a Lesson From React
Flue creator Fred Schott built AI agent hooks around the same composability problem React Hooks solved for web components.
AI Gives Good Engineering Teams Superpowers. It Gives Bad Ones a Faster Way to Fail.
A software engineer's viral post argues AI widens the gap between strong and weak engineering teams, and makes mediocre developers costlier to hire.