I'm trying to migrate from Pandas to Polars, but man it really sucks because ChatGPT is so weak for doing Polars stuff - it uses faulty or outdated API all the time. Feels like a fun challenge to load context from the Polars docs before doing API calls to ChatGPT.

I had some luck with copy pasting nativewind docs for react native since it didn’t have those libs yet and it worked but lost the context quickly

@backpack put a bank in your pocket and use your money whenever, wherever, however without permission

icebreaker.xyz/notdevin

Yeah feels like it requires new chat per query. I've heard sayings like "once an API is out there, it's permanent" (paraphrased), now with the information about APIs embedded directly into LLMs, feels like it's actually becoming harder to have breaking changes.

"Oh, breaking API changes? Now my ChatGPT code suggestions don't work."