Am I crazy for thinking that this is a pretty big regression compared to past models? I remember being blown away by GPT 4.5, and I kept using it up until they decommisioned it. I think claude 3.7 sonnet was pretty good too. Gemini seems to be the best one right now for actually talking. Opus is top tier for code but when i talk to it I want to rip my hair out. GPT-5.6 is doing best for me right now among the powerful models.
I have had no issues with LLMs trying to force a language on me. I tried the whole snake game test with ChatGPT but Instead of using Python I asked it to use the nodejs bindings for raylib, which is rather unusual.
To be more honest, it did feel like if I just stuck with the standard library it was okay at generating a higher ratio of useful snippets. Once I introduced a library is where things fell apart.
