It's shocking how many people didn't know LLMs can find novel real-world vulnerabilities. AIxCC deserved more attention