Just the rumour of a bug is enough to find an exploit these days

A security researcher discovered that automated AI-powered exploit systems can find vulnerabilities within minutes of a bug fix being publicly disclosed, with attackers probing his server just ten minutes after opening a pull request. Traditional security embargoes are no longer effective since AI agents need only vague hints about a vulnerability's nature to independently discover and exploit flaws, meaning exploitation now often precedes the public patch release. This shift requires a fundamental rethinking of how the open-source community handles security responses.