Eli Lifland: OpenAI's training pause likely isn't a substantial slowdown, compute shifts to experiments

eli_lifland · x · 2026-09-30

Forecaster Eli Lifland argues OpenAI's pause likely won't meaningfully slow progress: paused training compute probably redirects to experiments, with little impact on a months timescale. He expects public statements to oversell the slowdown while remaining technically true, against official claims that all training, eval, and tool-use inference of their most capable models are paused.

Related event: Forecaster Eli Lifland doubts training pause will meaningfully slow frontier AI, puts ~25% odds on real slowdown(6 posts)→

Original post →

More from AGI Musings

AGI Musings channel →