CLAIM-2026-08-30-03
The human baseline was 30 ideas from 28 researchers with at least one year in technical AI safety, 2.5 years on average, each given up to eight hours.
Why it matters
It is the comparator that makes the 6.4-hour figure mean anything, and it is the number aggregator summaries got wrong.
Sources
- alignmentsc2026automated
- anthropic2026automated