alt Hacker NewsOut of curiosity, was this on a codebase where other folks started using Opus 5.5? I’ve been curious if each generation of model has its own biases that make it generate output that’s more difficult for other models/previous models to work with. That’s been my personal suspicion when I’ve hit situations like yours
It was on 2 separate private repos, personal projects. No one else working on those
But now that you asked, I recently had a similar experience with Codex, but on a shared repo. I also was using a slightly older model, and the all of a sudden it kinda stopped running tasks. I would ask it to do something and the it would just kind of do an analysis, then give me a recommendation of what to do. I had to tell it to execute the task, and then it would “be lazy”, like run one part then give me a recap and recommendation again instead of just going
Switched to a newer model, and the issue was instantly gone