Ask HN: What does your local LLM setup looks like?

What models and hardware you are using? For what purpose? What were the challenges? Any tricks that helped you in doing this?

This might help new users like me setup theirs.