The quantized model you can run locally works decently and they can’t read any of it, which is nice.

I use that one specifically https://huggingface.co/lmstudio-community/Meta-Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q4_K_M.gguf

If you’re looking for a relatively user-friendly software to use it, you can look at GPT4All (open source) or LM Studio.

It won’t fuck off even if you ask it how to disable it gives you bullshit answers. Some “AI”.