{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreif7voz2623ehcwuozf3g7pe7rpe7kk4a5f6hn6vd4lfgz3jbm6sha",
    "uri": "at://did:plc:haakkg7y3xdghcdmprxeexso/app.bsky.feed.post/3mjrm2tiq4lb2"
  },
  "path": "/t/the-mythos-shock-deconstructing-project-glasswing-and-the-architecture-of-cyber-oligarchy/37236#post_3",
  "publishedAt": "2026-04-18T12:20:42.000Z",
  "site": "https://discuss.privacyguides.net",
  "textContent": "A large chunk of this seems AI generated, but let’s entertain it.\n\nWhen you “deconstruct the defensive narrative” what you really see is a corporation losing money and cutting costs, they cannot afford to give everyone access to mythos because they do not have the computational capacity for it. Smaller models are capable of finding most of the same exploits with a proper harness.\n\nNo, they cannot “hack everything”. This entire section is hyped up by AI, nothing really changes. There has always been a power-asymmetry because the bourgeoisie are the only ones capable of funding large-scale operations. What you are actually seeing is an irrational market and a US government run by corporations that desire nothing but profit.\n\nAnthropic have no moat, and OpenAI, even after their defense deal, is gaining goodwill again because anthropic keep cutting costs by lowering usage and quality. Local models have been getting better, the recent Qwen 3.6 release has been the moment for me where dealing with anthropic’s nonsense and cost cutting isn’t worth it anymore, this local model is good enough for me.",
  "title": "The Mythos Shock: Deconstructing Project Glasswing and the Architecture of Cyber-Oligarchy"
}