Post #2215663

The big problem with using LLM's for coding, is that while there is plenty of code out there, there is far less appraisal data - i.e. how well does the code fit the use case?