Weird debater locutions. “The evidence would show, in my opinion.” Well, does it? Show us! This isn’t just a rhetorical game

Show Post

I actually think grok shows how difficult it is to do that. And part of its whole deal is that it had to specifically build a racist dataset and even then its spotty

Show Post

"That they *may* produce truthful information is incidental." Sorry, but that's really stretching reality here. It's pretty darn clear that it's in the interest of LLM companies to have their tools provide truthful information, particularly given the enterprise money at stake!

Show Post

If LLMs are talking people out of conspiracies by confronting them with facts, how does that make them bullshit machines?

Show Post

I wonder if it's not the actual mechanism itself, but simply AI's continued availability to be persistent in the face of exhausting stupidity.

Show Post