Researchers Debate Reliability of METR 2x Speedup Claim

Jacques Thibs and Jeff Ladish debated METR's 2x task-length-doubling claim, with Ladish arguing the figure may hold within error margins but questioning whether it relies on undisclosed private data and sound AECI analysis.

2026-11-08 ~ 2026-11-08 · 2 related posts

• Debate over METR benchmark fit: whether the 1.9x slope increase holds up — JacquesThibs · 2026-11-08
• Debate over METR's 2x speedup claim: does it lean on private information? — JacquesThibs · 2026-11-08