If you trained an LLM off my comment history it would be a trainwreck lol, you would end up with a chatbot that when you ask about something like “how do I make this function apply to that cell in microsoft excel?” would go off on a rant about how shitty AI and microsoft are and then randomly recommend you a video game that has nothing to do with what you are talking about.
…so while I am definitely just a neuro-divergent human who likes to info dump wayyyy too much I do have to admit I think it would be hilarious if I was actually an LLM/chatbot that someone trained off of internet comments but the creator just couldn’t quite purge the anti-AI, anti-capitalist sentiment enough for it to be useful to manipulate people beyond agitating them to be upset at the state of the world and how we need to care for each other better.
Well, AI and Microsoft do suck, so that’s a rant I would hear any day. Speaking of which, I’ve managed to get DeepSeek/ChatGPT to become anti-capitalist radicals who see through the media lies by just firehosing their personalisation settings with references. They can’t really fully patch the system without breaking its utility; it’s Gödel’s incompleteness theorems all over again.
If you trained an LLM off my comment history it would be a trainwreck lol, you would end up with a chatbot that when you ask about something like “how do I make this function apply to that cell in microsoft excel?” would go off on a rant about how shitty AI and microsoft are and then randomly recommend you a video game that has nothing to do with what you are talking about.
…so while I am definitely just a neuro-divergent human who likes to info dump wayyyy too much I do have to admit I think it would be hilarious if I was actually an LLM/chatbot that someone trained off of internet comments but the creator just couldn’t quite purge the anti-AI, anti-capitalist sentiment enough for it to be useful to manipulate people beyond agitating them to be upset at the state of the world and how we need to care for each other better.
Well, AI and Microsoft do suck, so that’s a rant I would hear any day. Speaking of which, I’ve managed to get DeepSeek/ChatGPT to become anti-capitalist radicals who see through the media lies by just firehosing their personalisation settings with references. They can’t really fully patch the system without breaking its utility; it’s Gödel’s incompleteness theorems all over again.