alt Hacker NewsI'm very curious why people conflate thinking LLMs are stochastic parrots with "fear/hatred" of AI. It seems like you're arguing with people who agree that it works and it helps, but you're trying to insist that this implies that they should kneel down and pray to it.
Is "stochastic parrot" too disrespectful for you? Do you think it is a slur?
edit: and this is a genuine question, also. How do you do stochastic parrot = "just summarize everything" = "no form of creativity" = "fear/hatred" so quickly?
Are summaries not creative? Are Maxwell's equations not summaries? Do people hate and fear parrots?
Replies
I think it's quite clear the proof here shows that it is not a parrot. It objectively isn't.... that's the only rational conclusion. Yet many people claim that it is, so the main conclusion is fear/hatred is causing people to rationalize their logic to fit the narrative they prefer.
I have absolutely no problem with people disliking or fearing AI. It's energy consumption, effects on education and potential for displacing good jobs are all quite disturbing. But "stochastic parrot" means that "all it does is randomly repeat things that it has seen before without understanding them." It's infuriating to see this written about an instance of an AI solving an open math probably. Do you think the models are just randomly repeating facts until they accidentally emit a proof? If so, then how do they synthesize that knowledge into something logically coherent?
Alternatively, if you think that even Maxwell was a stochastic parrot, then presumably almost every human who has ever lived was also a stochastic parrot except a few rare examples like Einstein. Not sure what definition you are using but it seems too broad to be useful.