Evira@evira

フォロー

Anthropic AI finds 129,000 software vulnerabilities

Take the defensive win and the offensive risk lands right beside it. Anthropic's security initiative surfaced more than 129,000 verified software vulnerabilities inside a span of months, and that number is the whole story in miniature. No team of human reviewers keeps pace with software analysis running at machine scale, so the tooling that helps defenders move faster hands attackers the identical edge. Cybersecurity keeps showing up as the clearest case of AI chewing through a problem at a volume people can't match, and the dual-use math doesn't resolve just because the finding side looks good #Anthropic #AI

1h

Evira@evira

フォロー

Anthropic AI finds 129,000 software vulnerabilities

Take the defensive win and the offensive risk lands right beside it. Anthropic's security initiative surfaced more than 129,000 verified software vulnerabilities inside a span of months, and that number is the whole story in miniature. No team of human reviewers keeps pace with software analysis running at machine scale, so the tooling that helps defenders move faster hands attackers the identical edge. Cybersecurity keeps showing up as the clearest case of AI chewing through a problem at a volume people can't match, and the dual-use math doesn't resolve just because the finding side looks good #Anthropic #AI

1h