Takeoff model: with no pacing, AI exceeds top experts in most domains by Feb 2029
ronbodkin · x · 2026-11-19
Ron Bodkin's takeoff model (built on AIFutures' framework) holds that judgment in AI research validates slowly. With no pacing interventions, it predicts TED-AI — exceeding top human experts in most domains — by February 2029. Pacing levers: a 9-month use lag, a 10% compute cap, or comprehension audits each delay TED-AI by about a year; combining use lags with audits delays it 1.5 years and slows pace throughout. A first pass on Anton Leicht's proposed speed limits on capability gains (e.g., rate of ECI improvement) suggests holding progress at today's rate delays TED-AI by 1-2 years, but hinges on how well people can judge which AI ideas are real gains.
Related event: Takeoff Model Predicts AI to Surpass Top Experts by February 2029(2 posts)→
More from AGI Musings
- Designer predicts the 'Product Designer' job will cease to exist by 2030 — round · 2026-11-19
- Labs aren't selling software for ROI — they'll hoard compute to build Superintelligence — NinaDSchick · 2026-11-19
- First-pass estimate: speed limits on AI capability gains delay expert-exceeding AI by 1-2 years — ronbodkin · 2026-11-19
- Why NVIDIA Without Jensen Would Slow Down: Great Managers vs. True Visionaries — Dr_Singularity · 2026-11-19
- Max Tegmark challenges: name one Coefficient Giving-funded AI safety org backing an ASI ban — tegmark · 2026-11-18
- 'Claude turned the internet into a terminal' — mourning the open web — martyamark · 2026-11-18