Hacker News new | past | comments | ask | show | jobs | submit login

Just imagine if we didn’t use heuristics to fix arbitrary inputs and instead we used some sort of learning algorithm that was trained on producing valid looking json - then all we’d need to do is add a prompt and start with some random noise… :P



I've written a parser to handle all the cases. I can add a custom model/LLM if the parser fails to improve reliability even further if parser fails.

I believe ML Algos are a blackbox and non deterministic. So, they might fail for any given input and can't be trusted as a primary source of truth.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: