Jess Peck on the dangers of AI optimism:
Filed under: language models machine learningFor a lot of these things, it’s not going to kill people—just make things a little bit worse. For example, the web: the top ten results on Google used to be a lot harder to game with SEO nonsense. Bot-made content used to be a lot more distinguishable from human-made content. ChatGPT has blurred that line: it is hard to tell when something is made by a human and not made by a human.
[…]
We want easy answers, tailored to ourselves. We want the feeling of community building without taking the risk of talking to another person. We are in a sea of vibes and opinions, none of which are based in fact. And AI is only making that worse.