236
Anthropic says its latest AI model is too powerful for public release and that it broke containment during testing
(www.businessinsider.com)
This is a most excellent place for technology news and articles.
Translation: It is good at finding bugs that the NSA doesn't want people to know about.