Anthropic’s AI Flagged 29,000 Possible Software Vulnerabilities. Humans Checked Just 6,000

Anthropic's models flagged more possible software flaws than its reviewers could check. Its free new scanner changes who receives the unverified findings.

Anthropic's models flagged more possible software flaws than its reviewers could check. Its free new scanner changes who receives the unverified findings.

The conflict between Microsoft and a researcher known as Nightmare Eclipse intensifies, with threats of further zero-day exploit disclosures looming.