Ask HN: What does your local LLM setup looks like? (opens in new tab)

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