> If you are letting your LLM look at StackOverflow, you are doing it wrong
replySo you've evaluated all the sources that the model was trained on initially have you? How long did that take you?
> I'm shipping quality software and features to my customers at a pace I haven't been able to before.
I'm sorry are you agreeing with me or not? It sounds like you're agreeing with me.
I’m just saying that you can’t just let it rip based on its training alone, it needs to be grounded and harnessed in stack specific tooling.
reply