Toby Ord: lambda is task-dependent — the real question is lambda for AI R&D under high parallelism
tobyordoxford · x · 2026-09-30
In the same thread, Toby Ord adds a methodological caveat: λ is highly task-dependent, so the key question is what λ for AI would be on the task distribution of AI R&D in the limit of high parallelism.
He also suggests checking whether human λ and AI λ differ for the same type of task.
More from Research
- LLM-42 Paper at SOSP 2026 Brings Deterministic LLM Inference via Verified Speculation — tianyin_xu · 2026-09-30
- Frontier AI Is a Set, Not a Point: Jagged Capabilities May Be the Steady State — vsikka · 2026-09-30
- Open-source library adds reproducible seeded optimizers, recursive init, hard negative mining — leland_mcinnes · 2026-09-30
- Anthropic red team: GLM-5.3 pulls off full control flow hijacks in 4% of binary exploitation trials — Simon Willison · 2026-09-30
- Researcher loses confidence in AA benchmarks, calls them "very misleading" — tianyin_xu · 2026-09-30
- Adelaide's DCSD Decouples Credit Direction and Magnitude in Self-Distillation, Beats GRPO Across 11 Benchmarks — AdelaideUniversity · 2026-09-30