Anthropic's 'Claude Mythos' leak sends software names sharply lower
8 hours ago
- #AI
- #Cryptocurrency
- #Cybersecurity
- Anthropic is testing a new AI model called 'Claude Mythos', described as a major step up in capability from existing systems.
- Leaked internal documents warn that 'Claude Mythos' could heighten cybersecurity risks by rapidly finding and exploiting software vulnerabilities, potentially accelerating a cyber arms race.
- The iShares Expanded Tech-Software Sector ETF (IGV) is down nearly 3%, contributing to bitcoin's drop to $66,000.
- Palo Alto Networks (PANW), Crowdstrike (CRWD), and Fortinet (FTNT) stocks are down 4%-6% following the news.
- Details about 'Claude Mythos' surfaced after internal materials were accidentally exposed in a publicly accessible data store, including draft announcements and internal content.
- Anthropic is developing a new tier called 'Capybara', which would be larger and more intelligent than its current most advanced model, Opus.
- Publicly listed bitcoin miners are pivoting to AI and high-performance computing infrastructure due to unsustainable economics, with over $70 billion in contracts signed.