In other words, when you ask for example “do vaccines cause autism?” you are looking for information about whether a thing in the world (a vaccination) produces an outcome (autism). But the LLM can only answer “what words could sound right after this?”

The thing to understand is that an LLM is not in any sense reading and understanding your words. It is not interpreting and producing “factual claims”; it is interpreting and producing “completions that are possible.”

I want to continue to push back on the reigning “hey I’m just being reasonable” stance on AI. I “understand” it being useful for coding & emails too. I also understand why beans make us fart. The adult issue is, why are we doing this and is the cost worth it? There’s only ONE reasonable answer

The inability to just tell the right wing to eat shit, fuck off, die mad about it, is based on a perverse belief that their incessant whining pissy moaning is something that deserves any kind of respect or civility.