News for hackers. Stuff that matters.
Re:Measuring AI Ability to Complete Long Tasks: Opus 4.5 has... by rishabhaiover on Sunday December 21, 2025 @10:35PM (#46351038)
Predictions over historical data in a landscape with fragile priors doesn't seem like a strong metric to me (it's a useful approximation at best)

There are no comments on this comment yet.