scaling01: Single AECI Score Meaningless, Calls for Public Benchmarks

KOL scaling01 argues that a single AECI score is meaningless without public benchmarks and task details, calling the 200 ECI figure a "vibe estimate." He urges frontier labs to disclose internal capability levels, noting that non-disclosure itself poses a safety risk.

2026-11-06 ~ 2026-11-06 · 2 related posts

• scaling01: single AECI score useless, frontier labs should disclose full benchmarks — scaling01 · 2026-11-06
• scaling01: 200 ECI is just a vibe estimate without benchmark details — scaling01 · 2026-11-06