Gregory Barbaccia, the federal chief information officer, sent an email to Cabinet departments on Tuesday outlining a new memo for 'Project Glasswing.' The memo, reviewed by Bloomberg, details how the White House is preparing to give major federal agencies exclusive access to Anthropic’s 'Mythos'—a powerful autonomous cybersecurity model. While the government secures its own digital perimeter, the public remains locked out of the tools needed to defend against the very hacks Mythos was designed to stop.

This is the final phase of Cognitive Enclosure. By gating the Mythos model, the state and its 'Glasswing' partners—including JPMorgan Chase and CrowdStrike—are extracting security rents from a public that is increasingly vulnerable to AI-generated code. As of this week, 60 percent of new global code is AI-generated, much of it riddled with the 'hallucination tax' of security flaws that only a model like Mythos can reliably audit.

In the UK, banks are already being briefed on how to use Mythos to secure financial ledgers, effectively creating a two-tiered internet. One tier is a fortified, autonomous fortress for the elite; the other is a toxic commons for the rest of us, where 'vibe coding' and un-audited software are the only options left. The information space is being saturated with synthetic interference, making the signal-to-noise ratio unmanageable for anyone without an enterprise-grade AI filter.

According to the memo, the OMB is setting up 'protections' that would allow agencies to use the guarded tool. These protections do not extend to the small business owner or the private citizen. The state is hoarding the cure for a digital plague it helped create, ensuring that while your data is leaked and your accounts are drained, the federal servers remain pristinely silent.