LLMs definitely have personalities. And changing ones at that. gemini free tier was great for a few days but lately it keeps gaslighting me even when it is wrong (which has become quite often on the more complex tasks). To the point I am considering going back to claude. I am cheating on my llms. :D

edit: I realize now and find important to note that I haven't even considered upping the gemini tier. I probably should/could try. LLM hopping.

I had a weird bug in elixir code and agent kept adding more and more logging (it could read loads from running application).

Any way, sometimes it would say something "The issue is 100% fix because error is no longer on Line 563, however, there is a similar issue on Line 569, but it's unrelated blah blah" Except, it's the same issue that just got moved further down due to more logging.

Yeah, the heavily distilled models are very bad with hallucinations. I think they use them to cover for decreased capacity. A 1B model will happily attempt the same complex coding tasks as a 1T model but the hard parts will be pushed into an API call that doesn't exist, lol.