BREAKING NEWS: Anthropic's latest model will NOT help you if it thinks your ML research/ML engineering is interesting, and/or will secretly degrade its IQ so that the average engineer won't notice. We are already seeing Anthropic's latest model's moderation filters our GPU inference research and programming ๐ญ
@thsottiaux Not sure if itโs something Iโm missing but itโs really hard to get codex to work on long running tasks. Nothing crazy just want to work more than two hours at a time
Apart from less ideal prompt adherence, and some hallucination issues, Opus 4.7 actually has pretty good taste in terms of UI and product design, and is in fact much better than Opus 4.6 in coding imo. You just need to give it tasks in the right way.
@daniel_nguyenx Tauri is wildly unstable and full of visual bugs(on Mac at least). Weโve had a lot of issues with spawning correct looking windows with a fairly small project and it has been extremely frustrating.
Anyone notice Opus 4.7 is less willing to use sub-agents in Claude Code? I could be directly asking it to use sub-agents in the prompt, and it would still go on to edit the files by itself.