Jan Kulveit Bets Against AI 2027: Reality Tracks My Timeline, Not Kokotajlo's
jankulveit · x · 2026-10-14
AI risk researcher Jan Kulveit argues reality is tracking his timeline rather than Daniel Kokotajlo's "AI 2027" forecast, and he is increasingly confident he'll win their bet. He agrees with AI 2027's tech tree but not its human-response forecasts. His example of a wrong feedback loop: labs operate in a "risk homeostasis" regime where leaders' perceived sense of doom targets 10-20% — lower and they accelerate, much higher and they act. As more social systems get involved, technology becomes a smaller share of what forecasters must get right, so divergence with AI 27 will keep widening.
More from AGI Musings
- TIME Reveals the Secretive Anthropic-OpenAI Race to Build Self-Improving AI — jeremiecharris · 2026-10-14
- Lab insiders: recursive self-improvement may start within 3-12 months — jeremiecharris · 2026-10-14
- China's compute deficit delays payoff from stolen AI algorithm secrets — jeremiecharris · 2026-10-14
- Compute gap means stolen algorithmic insights pay off slowly for China — jeremiecharris · 2026-10-14
- No scalable metrics for alignment, unlike the capability road to RSI — jeremiecharris · 2026-10-14
- RSI is on the table if compute is big enough or great ideas are common — jeremiecharris · 2026-10-14