sjolsen :sparkles_agender: on Nostr: Last boost: When we talk about so-called "AI", it is easy and tempting to focus on ...

Last boost: When we talk about so-called "AI", it is easy and tempting to focus on the fact that the product is bullshit and doesn't do 95% of what is claimed. It isn't sentient, it isn't intelligent, it doesn't have a personality. The cutting edge of AI is an overgrown Magic 8 Ball.

But it's important to remember that the point of the technology isn't to "work". The tech industry (read: capital) is not excited about AI because it can deliver on a promise of correct or meaningful results. Capital is excited about AI because, irrespective of whether it "works", it benefits capital.

In the application cited, the LLM is not a technology for predicting human behavior. That's just window dressing. It's a technology for laundering accountability. It does not exist so your boss can "right-size" the right people. It exists so when your boss gets caught disproportionately "right-sizing", say, black employees, your boss can say: I'm not racist; the computer made those decisions. Computers can't be racist!