Big Technology Podcast Pentagon Insider: What's Next For Anthropic and The Department of War — With Michael Horowitz
110 snips
Mar 4, 2026 
Michael Horowitz, former Pentagon deputy assistant secretary and current UPenn professor, breaks down the Anthropic–Pentagon rupture and its policy stakes. He walks through how models like Claude are used in military workflows, why a contract fight over surveillance language blew up, and what the trust breakdown means for future AI-state partnerships.
AI Snips
Chapters
Transcript
Episode notes
Trust Breakdown Drove The Anthropic Pentagon Split
- The Anthropic–Pentagon rupture reflects a personality and politics clash more than a specific project dispute.
- The split began after Anthropic asked Palantir whether its tech was used in the Maduro operation, which offended the Pentagon and broke trust.
One Word Exposed Deeper Government Vendor Tension
- The core legal fight centered on a contract phrase about 'mass surveillance' versus 'all lawful uses,' revealing deeper trust and role disagreements.
- The Pentagon treats AI purchases like weapons procurement, resisting vendor constraints, while Anthropic sees AI as an evolving service needing deployment controls.
How Claude Plugged Into Pentagon Dashboards
- Claude was integrated as a backend decision aid inside Palantir's Maven Smart System to help commanders process public and classified feeds.
- Uses included querying news chatter and running simulations, always passing outputs through multiple human reviewers before action.