back
4 comments
Though an attorney (and the attorney he hired) ought to know better about...well, just about all of this, TechDirt makes a weak point here:

Meta’s AI chatbot has no access to Meta’s internal moderation decisions. It’s just a probabilistic stochastic parrot, repeating a probable sounding answer to users’ questions.

Yes, and that is clearly labelled, I'm sure, when one goes to use Meta's chatbot. Or are we just supposed to assume that an attorney-by-trade knows how LLMs work, rather than the attorney assuming that the chatbot supplies valid answers?

Is this article really suitable for sharing on HackerNews? It's a ridiculously unfair interpretation of the news. It's basically "fake news", as in it's so slanted that it has negative truth value and will leave readers misinformed.
TechDirt is reply-guy hot-take garbage, giving the most uncharitable take on the news as possible. I don't know why it gets posted here, either.
The brainworm really did a number on junior, didn't it?