Ryan Greenblatt predicts AI takeover could happen by 2029
mattturck · x · 2026-08-28
Matt Turck interviews Ryan Greenblatt of Redwood Research about AI timelines, Recursive Self-Improvement (RSI), and AI alignment. Greenblatt suggests planning "as though it happens in 2029," emphasizing that superintelligence poses a "dangerous" threat rather than just "bad." The discussion covers the AI 2040 Plan A, the concept of "mutually assured compute destruction" with China, and a sobering year-by-year timeline from today to potential AI takeover.
Related event: Redwood Researcher Predicts AI Could Take Over by 2029(2 posts)→
More from AGI Musings
- METR's OpenAI incident retrospective: AI had to read 1,300 transcripts because humans can't — alliekmiller · 2026-08-28
- Using framing theory to understand AI narratives: sci-fi and history metaphors shape public views — jacyanthis · 2026-08-28
- London Futurists Podcast: A guide to AI and future technologies — cccalum · 2026-08-28
- 'The Overshield Writes Itself': AI-Rated Surveillance Dystopia Arriving Ahead of Schedule — IridiumEagle · 2026-08-28
- Mouse Connectome Cost Fell From $10B to ~$150M; Human Brain Map Maybe $1-2B: Marblestone — juanbenet · 2026-08-28
- Timothy Lee bets against home humanoid robots before 2029, sees only teleoperated pilots — binarybits · 2026-08-28