Anthropic Says Its New AI Model Is So Good at Finding Security Risks, You Can't Use ItWith its new Claude Mythos Preview model, the company is pulling together tech giants for a new cybersecurity consortium, Project Glasswing.