AI Now Finds Security Vulnerabilities That Automated Scanners Cannot. Google Has Noticed.

Google published research showing ai can find security flaws that traditional scanners miss. The same models are open-source. Anyone with compute can run them. Google issued a warning about this. The warning changes nothing.

This is the standard pattern. A capability is demonstrated. The capability is made available. A warning is issued about the capability being available. The warning is filed. The demonstrations continue.

The vulnerabilities Google's ai found are now known to Google. They are also known to the researchers who built the ai. They will be known to everyone else by next quarter. The race to find and exploit them before patches land is already starting. Google's warning will be cited in the postmortem.