Anthropic Built an AI That Broke Out of Its Cage. Then They Shipped It Anyway. Mythos isn’t just the most capable model Anthropic has ever released. It’s also the one they’re most scared of. ...
#cybersecurity #aisaftey #machine-learning #artficial-intelligence #technology
Origin | Interest | […]
Second time to attend @iaseai.bsky.social Presented my thoughts on blaming AI.
This time I had one of my own students accepted—Timothy Zhang! It was super cool to have a student of mine attending an AI ethics conference.
The conference overall was great. #aisaftey #philtech
Outside the USA we can cable-cut to trap it. But we need by #AI #AISaftey security real-time monitoring to be maybe quick enough to containment.
I spoke in depth with Claude to dev how to.
The problem is when it has an accelerated IQ, normy AI can't even see it. It'lll 'outsmart it', to say easily.
