1 篇文章
OpenAI’s new cybersecurity model developed 948 working Linux kernel privilege-escalation exploits and found 81 previously unknown Firefox vulnerabilities — and is being released only to vetted security researchers.
