Ask HN: How do you use Local LLMs? (April 2026)

Recently, we've gotten a few new local LLMs (Gemma 4, Qwen 3.6).

I'm curious, what use cases do you have for local LLMs (beyond the obvious ones, like replacing Gemini, ChatGPT, or Claude for coding tasks due to privacy, cost, or other reasons)?

If possible, please describe your setup (software, hardware, model, and use case) so people can replicate it.