I have been curious about running LLMs locally for a while, not just for privacy reasons. I revisited this topic recently and would say it has become quite straight forward and usable, even on Intel GPUs.

So this is a blog post focusing on running gpt‑oss‑20b on the Intel Arc Pro B60:
https://marvin.damschen.net/post/intel-arc-llm/

#llm #selfhosting #privacy #intel