3 ms·

I would've said that understanding and fixing the data would be one of the best usecases for the AI.

• bugglebeetle 3y ago
  Unfortunately, there are still many ways to “fix” things that have a lot of trade-offs or downstream consequences for analysis. For most basic cleaning tasks, LLMs are also still way too slow.