Toby Ord: METR's 4-month doubling implies training compute must 4x every 4 months
tobyordoxford · x · 2026-11-09
In a discussion with @BasilHalperin and @ramez, Oxford researcher Toby Ord works through the compute implications of METR's task-length doubling time: 7 months pre-2024, then 4 months. If compute is the sole driver (with data/algorithm mix held constant), training compute would need to quadruple every 7 months up to 2024, then every 4 months after.
He finds this worse than expected and notes that, since it assumes no gains from algorithmic or data-quality improvements, it serves as an upper bound on compute requirements.
More from AGI Musings
- LLMs will scale apocryphal attribution losses across the humanities — StewartalsopIII · 2026-11-09
- 7 Mid-Career Tech Workers Pivoted to AI Safety Full-Time — Almost None Found a Job — AaronBergman18 · 2026-11-09
- Should PhD Programs Be Smaller but Better Paid? Academia Debate — miniapeur · 2026-11-09
- Dan Faggella: AGI should be a participant in life's process, not a runaway prosthetic — danfaggella · 2026-11-09
- Caring more about who made it than if it's good: the AI authorship obsession — VraserX · 2026-11-09
- Paper in Phil Trans B: human-AI agent interactions are bidirectional and reshape who we are — AnnaCiaunica · 2026-11-09