[dingnuts]

> think of tech that would actually improve my life I think about things like a clothes-folding machine,

I think about this constantly, and it's a funny example because apparently the math involved for figuring out how to fold clothes is basically an unsolved problem. As I understand it the robotics -- also a very hard unsolved problem -- are the easy part, which should emphasize how horribly hard this problem is.

Several other things: 1) I recently saw a LinkedIn lunatic claim laundry was already automated, because laundry machines exist (way to out yourself as someone who never helps with laundry)

> mediocre poetry

2) I refuse to call language models "intelligence" until they can at least meaningfully contribute to the design of novel devices like the robots we are discussing. If it can't invent, it's not intelligent! Intelligent things are capable of novelty, not only regurgitating existing witnessed patterns.