But the actual benefits for society are honestly underwhelming. Yes, we can code faster now -- but was there a serious problem with humanity being unable to produce enough code fast enough? We can write emails faster, and summarize documents much faster -- ok, yes, that does make some people more productive, myself included. But yeah, email also made us more productive than typewriters, which made us more productive than hand-written letters. LLMs might uncover medical discoveries that humans just couldn't on our own? Uh, maybe? I mean we've had ML/DL models for a while and they're getting better.
And then you consider the way that AI is being implemented, developed, invested in, and pushed, by large corporations, whatever benefits there could be from LLMs, are highly outweighed by the negative effects on society as a whole. It is a highly destructive force in almost every area. It could be a force for good, but at the rate we're going, it will not be, because that's not the profitable path.
Nuclear fission was an amazing scientific discovery as well! And yet if it weren't for strict regulations, international treaties, UN-verification agencies, major protests against nuclear weapons etc., there's a pretty good chance that millions would have died by now by some nuclear weapon. (In addition to the hundreds of thousands who did die by nuclear fission in Nagasaki and Hiroshima -- which makes you wonder if the discovery of nuclear fission was worth it at all.) (Ah, but nuclear-backed MAD prevents conventional wars that might have killed millions anyway! Yes, that argument has some merit but I'm certainly not convinced. I'm old enough to remember feeling like we were literally a button-press away from destroying much of humanity. The weird thing is that we still are, we just don't think about it much because it hasn't happened yet.)