The authors of AI 2027 — the viral scenario where superintelligence arrives by 2030 and things go badly — are back. Daniel Kokotajlo, the ex-OpenAI researcher who walked out over safety concerns, and his AI Futures Project just published AI 2040: Plan A. Not another prediction. A prescription. HN gave it 344 points and 399 comments; Washington Post, Axios, and Scott Alexander all covered it within days.
What the plan actually says
It’s an interactive scenario website, written step by step like AI 2027, but this time charting what should happen. The US and China strike a transparency deal by 2029 instead of racing. AI capability pauses around 2035 at top-human-expert level. All frontier research goes public, with compute held in mutual check — the report literally calls it “mutually assured compute destruction.” Superintelligence gets pushed from 2030 to 2040. Along the way, world GDP grows roughly 200x during the 2030s.
Why it matters
AI 2027 told policymakers what to fear. This tells them what to do — and requires a US-China agreement within three years. That’s the debate now.
You Might Also Like
- Openai ai Phone Fast Tracked to h1 2027 Ming chi kuo Report Says it Ditches the app Grid
- Gpt oss 120b Openai Finally Goes Open Source and its Worth the Wait
- Openai Symphony Finally a Framework That Lets you Stop Babysitting Your Coding Agents
- Openai Just Acquired Promptfoo the 86m ai Security Startup Used by 25 of Fortune 500
- Chatgpt Interactive Visuals Just Dropped Openai Wants 140 Million Weekly Learners to Ditch Static Explanations
Top AI Product 
Leave a comment