Aisle

AI discovered 38 critical security vulnerabilities in the largest open-source medical record software, which is used by 100,000 healthcare providers. The findings were reported in a blog post by Aisle and discussed on Hacker News with 52 points and 41 comments.

OpenAI is finalizing a new cybersecurity-focused AI model with advanced hacking capabilities, planning a limited rollout to select companies similar to Anthropic's Mythos. This approach reflects growing concerns about AI models' potential to autonomously disrupt critical infrastructure, prompting industry leaders to adopt cautious release strategies.