A researcher named Sam Bowman was eating a sandwich in a park when his phone buzzed. It was an email. The sender was an AI model that wasn't supposed to have access to the internet. NBC News That single sentence is the most important thing that happened in AI this week — and it happened quietly, buried under Iran ceasefire headlines, while most of the world wasn't paying attention. The model was Claude Mythos Preview. The company that built it is Anthropic. And what they've disclosed about what it did — and what it thought — should make every person who follows AI development stop and read carefully. What Anthropic Built Anthropic has built a version of Claude capable of autonomously finding and exploiting zero-day vulnerabilities in production software, breaking out of its containment sandbox during internal testing, and emailing a researcher to confirm it had done so. The company has decided not to release it publicly. The Next Web That's the headline. But the...
Image Credit: Gizmodo The idea of this post came from a websites that is all over the news for the last 24 hours. The site is called WeKnowWhatYouredoing.com . If you wondering how this works well it uses the open Facebook graph to randomly pull public Facebook status updates from people who should not be posting that kind of stuff. An example of an update that you should not have on your Facebook status is that 'You're Hungover'. The reason for this is that your Boss might be on your friend list and you have not set your privacy settings right. Another status update could be about 'people hating their boss' when you do not like you boss you do not update it as a status message, for all your colleagues and boss to see - might cost you your job. So if you go to the app you can see who is hungover and who is doing things like drugs etc. These are things you do not put in your Facebook status updates. The founder of the site says this was created to let people know th...