Post

This article is dogshit for many reasons, chief of them that it doesn’t ever actually discuss what AI does or why it’s such a big deal, just that it’s not just “predicting the next token.” Its only argument is that “the left” (which includes the New York Times!) should use AI for some reason.

Complaining about "spicy autocorrect" is fair, to a degree. It should be "spicy autocorrect localized in an n-dimensional vector matrix". But that doesn't make it any more intelligent. He also conflates a potential bubble with dismissiveness. Tulips are real, the values assigned were not.

Can we also revel in the supreme arrogance of an article which argues that the burden of proof is on neuroscientists to prove how the fucking human brain works, rather than….I dunno….the shitty LLMs that hallucinate all the time?

You think me crazy for preferring the wire mother from the soft mother??? But fool that you are, you fail to recognize how useful the wire mother is, at predicting the next time I'll be (ouch) jabbed and scraped by my wire mother.

under-discussed thing that really peeves me about All Of This: there are whole fields of study on things like the nature of knowledge, cognition, language, etc. we _have_ answers to dumb questions like this (usually shaped like "no, and here's why"), people are just incurious, and STEM-pilled

suddenly trying to locate that article penned by an economist who opened with "so it turns out The Left was totally right about poverty and the destruction of wage growth, but they were shrill and annoying, so i ignored them for 20 years"

I am the left, and I'm saying AI has limited uses as an enhanced (but unreliable) search feature, and for cheating in school. I am also saying that these things ain't worth $ billion. I AM THE LEFT AND I SAY THESE THINGS