FULL STORY
METR's "No Bottleneck" Claim Sparks Teknium Rebuttal
METR said it sees no convincing capability bottlenecks for the next 12 months, prompting open-source developer Teknium to publicly rebut the claim, citing hardware cost inflation and slowing funding as clear constraints.
2026-11-06 ~ 2026-11-06 · 2 episodes · 4 posts
Episode 1 · METR Sees No Clear Bottlenecks; AI Progress May Accelerate Over Next 12 Months (2026-11-06, 2 posts)
METR said it found no convincing evidence of major bottlenecks across algorithmic progress, training compute, experiment compute, and data over the next 12 months, suggesting capability progress may accelerate, while acknowledging its view may have missed bottlenecks.
- METR finds no convincing near-term blockers to AI capability growth within 12 months — METR_Evals · 2026-11-06
- METR: No clear bottlenecks seen, capability progress could accelerate further in 12 months — paulnovosad · 2026-11-06
Episode 2 · Teknium Pushes Back on METR, Citing Compute Inflation and Funding Slowdown (2026-11-06, 2 posts)
Open-model developer Teknium publicly disputed METR's no-bottleneck conclusion, pointing to roughly 20x inflation in compute and memory costs and a slowdown in AI industry fundraising as obvious constraints.