>>Yes, it’s technically running, but not in a way that would be useful by normal LLM standards.

What are the LLM standards?

Do you know how many people use perplexity? I know many people who are not software engineers or tech workers and have a LLM subscription for rewriting their stuff (non-native english speakers) in english. There are many use cases for running good models locally. Maybe not for you, but someone might find this beneficial.