Can we all just pause and recognize that scaling models to TRILLIONS of parameters is absolutely bonkers.
Like how did we go from accumulators holding ~20 words on the original ENIAC computer to modern LLM’s having 10⁹–10¹² learned parameters.
It’s crazy because there are people who don’t even understand how a proof works. Yet they claim something is proof verified.
Brother I read papers all day and I don’t even fully know how a proof works. I don’t claim to. Nor should you. Unless you are a mathematician.
@IntenseRealist@matthew_petti The problem is it CAN be. Not that it is. Which means the mathematicians are the only ones who can validate. Why is an llm which can’t validate releasing something someone else must validate?
@Quasilocal Iv solved thousands of math problems. I can’t review them nor do I have the credentials to verify them. But you can verify them. Also we don’t need you because we solved it. Because we know it’s right. But you gotta tell us if we are right…