alt Hacker NewsAnyone with evals picking up a real signal? I've learned not to trust "vibes".
The only one I'm familiar with testing Opus 5.5 specifically over time is https://github.com/ninjahawk/livenerf and it's not showing anything.
Replies
velcrovan • today at 3:08 PM
Seriously. "Something changed" is not a "nerf".
It's hard to tell, but I think they also pin the harness version? Updates to Claude Code can also have big effects on perf, and Claude Code updates a lot.