genai / news / / WIRED
The company has convened an industry consortium to grapple with the cybersecurity implications.
Anthropic announced Mythos Preview and Project Glasswing, a consortium to address AI cybersecurity risks.
KEY POINTS
- Over 40 major tech and infrastructure organizations have private access to Mythos Preview for security testing.
- Mythos Preview has already uncovered thousands of critical vulnerabilities, including decades-old overlooked bugs.
- The model can perform advanced exploit development and binary analysis without source code access.
- Anthropic is using a staggered release to allow partners time to mitigate vulnerabilities before public launch.
COMPANIES
Summarized by Newsio from WIRED. How we summarize →