Toby Ord: 80% reliability emerges as bottleneck for AI task length growth
Toby Ord notes that the task length AI models can complete with 80% reliability has grown from 15 to 25 minutes, but reliability may saturate at 80-90%, leaving full R&D automation incomplete. Data from OpenAI and Google DeepMind shows slower growth than Anthropic's.
2026-11-19 ~ 2026-11-19 · 2 related posts
- Toby Ord: AI 80%-reliability horizon grew from 15 to 25 minutes, but 80-90% cap can't close the R&D loop — tobyordoxford · 2026-11-19
- Toby Ord: OpenAI and GDM Data Lag Anthropic's, 80% Horizon Grows to ~25 Minutes — tobyordoxford · 2026-11-19