Researchers Push Back on METR's AI Acceleration Findings

Multiple researchers challenge the widely cited AIFP/METR AI acceleration findings: Haydn Belfield argues its three key assumptions may not hold, with acceleration possibly only 1.4x rather than 2x, while DeepMind's Ege Erdil says shortened AI timelines reflect social conformity rather than evidence.

2026-11-13 ~ 2026-11-13 · 4 related posts

• Researcher pushes back on METR's 2x AI acceleration claim, citing 1.4x figure — HaydnBelfield · 2026-11-13
• Wildeford pushes back on AI acceleration research: 1.4x vs 2x barely changes the picture — peterwildeford · 2026-11-13
• AI 2027's stock market predictions miss reality, as researchers weigh the new METR report — DavidSKrueger · 2026-11-13
• DeepMind researcher says short AI timelines are social contagion, not evidence — NathanpmYoung · 2026-11-13