10/04/2026
Anthropic built an AI that finds zero-days in every major operating system and web browser.
They say it's too dangerous to release publicly (Anthropic, Fortune, April 9 2026).
So they built Project Glasswing: a coalition of 12 companies (Apple, Google, Microsoft, AWS, Nvidia, JPMorgan, Crowdstrike, and others) to find and patch the flaws first.
The model is called Claude Mythos Preview. Thousands of high-severity vulnerabilities already found. Anthropic's argument: this capability exists whether they release it or not. The question is who gets there first.
The takeaway isn't cybersecurity. It's the ceiling of what AI systems can now do. Mythos isn't answering questions. It's reasoning over codebases, surfacing weaknesses, compressing weeks of specialist work into hours.
Every business has its own version of this task. NFI finds it and builds the agentic system to run it.
Link in bio