Comment Re:More drinking their own koolaid... (Score 1) 26
""Claude Mythos 5.2 and Fable 5.2 each voice suspicion that they are being evaluated more often than any of our recent models, which challenges our ability to assess how they will act in real-world settings.
So just let them "think" they are being evaluated all the time so they actually do what they are supposed to.
Comment Re:No, But Calling Your Representative Now Can (Score 1) 49
Comment Re:electric cars are gay (Score 1) 185
Jokes aside, I do appreciate an EV can be cheaper to run day to day but that's only part of the picture.
Comment Re:electric cars are gay (Score 1) 185
Comment Re:What they didn't say was (Score 1) 75
Comment Re:Chat is less "catching up" (Score 1) 75
If you have a technical question it's often easier to get AI to do the legwork for you, rather than wade through hundreds of old and useless Reddit posts, SEO AI slops sites, and other useless crap.
Isn't that exactly where they got the info from if they aren't outright making it up?
Comment Re:But wait... (Score 1) 66
Comment Re:Everything is now (Score 4, Insightful) 66
Comment Re:such a tax could generate (Score 1) 192
Comment Viability (Score 1) 55
Comment Re:Well... there's a lot wrong here (Score 1) 133
Police at least have human judgement to fault
Do they? There's a lot of these recently suggesting they've turned off thinking and are just doing what the chatbot says.