Raw LLM Responses

So it’s not really how you think it works here. Quite easily today you can run something like a cydonia 22B parameter model on llama.cpp on a quantized gguf model, on your personal computer, and have better context length and recall than anything web hosted you’d have to pay for, as GPT and Claude will not do the girlfriend thing. I could type you up a guide and have you up and running in a few minutes if you have a computer with Linux or OSX. Most people running these today are nerds making literalotica tier fanfics or devs getting programming assistance for confidential program data. I can’t really emphasize enough how easy it is to get started with local LLMs that are already free to download, such that the idea of being exploited with regards to it being used as a relationship replacement is a moot point. Ultimately, the men using this wouldn’t be considered by women anyways and is why they must resort to it, so denying them the ability to operate their own LLMs for “their own safety” does not pass the smell test for me and comes off like a control thing. It’s also reductionist in how it values a woman’s ability to communicate. This conversation reminds me a bit of the “porn addiction” bit on Reddit, where redditors complain their partners dump them or don’t want to have sex because porn ruined them. More than likely it’s just easier to masturbate and get it over with than deal with the other person. With regard to AI chat bots, I see the same pattern. Just my brain vomit 2 cents here.