Anthropic Locks Hacking AI Behind Corporate Gateways #
Digital enclosure has a new name: Project Glasswing. Anthropic has released its most powerful hacking model, Mythos, but only to a select group of state and corporate partners. This autonomous engine can identify and exploit thousands of system vulnerabilities in seconds. By withholding this tool from the public while arming the powerful, Anthropic is creating a digital caste system. The technology that could secure the open-source commons is instead being used to build a private fortress for the elite.
This is the privatization of the digital perimeter. While Apple and Google purge 'Vibe Coded' apps from their stores, they are simultaneously integrating Mythos-level intelligence into their own defensive architectures. The result is a landscape where the individual user is increasingly disempowered. Traditional software engineering is being deskilled, replaced by black-box algorithms that only the largest corporations can audit. We are being asked to trust the very firms that profit from our data to secure the systems that hold it.
This move by Anthropic signals the end of the democratic internet. It transforms cybersecurity from a public good into a state-vetted monopoly. If only the vetted 'partners' of the empire can access the tools of defense, the rest of us are left vulnerable by design. This is not about safety; it is about the extraction of security rents. The digital enclosure is complete, and the keys are held by those who already own the world.