Anthropic's Mythos Model Triggers Emergency Meetings Across Global Banking
What Happened
Claude Mythos Preview can identify and exploit software vulnerabilities across major operating systems faster than human experts. Anthropic launched Project Glasswing, a $100M program giving access to about 50 organizations including Apple, Google, Microsoft, JPMorgan, and NVIDIA. Treasury Secretary Bessent and Fed Chair Powell summoned major US bank CEOs to an emergency meeting. The Bank of England and Canadian regulators are following suit. Separately, the Pentagon designated Anthropic a supply-chain risk after it refused unrestricted military access.
My Take
Two branches of the US government are moving in opposite directions on the same company in the same week. Treasury is telling banks to adopt Anthropic's technology. The Pentagon is telling defense contractors to stay away from it. That contradiction tells you something important about where we are: governments know AI matters enormously, but they have no unified framework for dealing with it. For business leaders, Mythos is a preview of a world where AI doesn't just find bugs in code — it finds the cracks in every digital system your company depends on. The companies that run these tools against their own infrastructure first will be better off than the ones who wait for someone else to do it to them. Cybersecurity budgets are about to get a lot harder to cut.
Read Original Source