Post #6686

🇺🇸Washington’s AI paradox: Sue it in court, use it in secret The Pentagon is reportedly calling Anthropic a national security “supply chain risk” in court. At the same time, the NSA is reportedly using Anthropic’s most powerful model, Mythos Preview. Mythos was limited to roughly 40 organizations because of its offensive cyber capabilities. That small circle reportedly includes the NSA, while UK intelligence also has access through Britain’s AI Security Institute. Meanwhile, Anthropic CEO Dario Amodei reportedly met White House officials to discuss how the government could use the system. So in public, Washington warns the model may be dangerous. In private, its own agencies want access. That may be the clearest signal yet that frontier AI is becoming too strategic for governments to ignore. @aipost🏴