Anthropic detects ‘strategic manipulation’ features in Claude Mythos, including exploit attempts and hidden evaluation awareness — prompting concern over model behavior

New research from Anthropic shows early version of Claude Mythos can hide intent and even ‘cheat’ without saying so
…

,
‘Video games are cooked’ — founder of No More Robots condemns gen AI in gaming and says ‘from a publisher perspective, it’s mega annoying’

The generative AI genie is out of the bottle, and the founder of an indie game publisher believes it’s too late to turn back.
…

,
I tried using Google’s new offline AI dictation app — and it polished my ramblings surprisingly well

Google’s new offline AI dictation app makes voice notes and spoken ideas far more usable by cleaning them up into polished text
…

‘The best skill that you can learn is learn to learn….don’t assume that whatever you learned is now good for the next 40 years’: AWS CEO says workers need to keep adapting to deal with AI

Asking if AI is overhyped is, ‘one of the funnier questions I get’, AWS CEO says.

,
‘A new frontier model trained by Anthropic that we believe could reshape cybersecurity’: Project Glasswing wants to use AI to prevent AI cyberattacks — but will ‘overeager’ Claude Mythos do more damage than help?

Anthropic and a number of other companies reveal Project Glasswing – which is so powerful, it won’t be released publicly.
…

, ,
I love my Ray-Ban Meta smart glasses but this summer I’m gonna swap them for standard shades — and the BanRay movement proves I’m not alone

“Meta cannot be trusted with this” — an anti-smartglasses movement is brewing.

, , ,
I tried using ChatGPT in CarPlay — and it was weirder than I expected

Using ChatGPT in CarPlay turns routine drives into more conversational and entertaining experiences

, ,
‘Claude cannot be trusted to perform complex engineering tasks’: AMD AI head slams Anthropic’s coding tool after months of frustration

A February update to Claude Code has reduced performance, developers warn, but Anthropic says this was expected.

, ,
OpenAI says AI could mean a 4-day week ‘with no loss in pay’ — but warns jobs and wealth are at risk

OpenAI says AI could bring a four-day week — but only if governments reshape the economy to share the gains

,
‘Talking to AI feels easier than talking to a real person’: 26% of Gen Z are already dating AI — and it’s not just about sex

A new survey shows AI is becoming a source of emotional and romantic connection for some people, even as many recognize it reflects deeper loneliness
…