I don’t know what’s been happening, but for a few weeks now I’ve been running a prompt, and it simply generates the project with errors. Before this didn’t happen, before it generated the project and on the first run it worked without errors. Has anyone else noticed this? I’m using claude-sonnet 4.0
The only working model is sonnet 3.5 (at least for me) after I optimized and spent weeks perfecting and tweaking prompts/.cursor rules. If I use any other LLM or one that isn’t Claude 3.5 I just get error after error with nothing but linter errors and half baked code blocks that aren’t relevant to my prompts at all. Good luck!
Edit: Even when using the most expensive model, I just get literally only emojis
ok I’ll try here too