Thinking On Paper AI Strategy For The Department Of War: Signed, Pete Hegseth
Apr 3, 2026
A leaked Pentagon memo that made AI central to military strategy sparks a debate about automated targeting and the ethics of speed-over-safety. They unpack projects aimed at AI swarms and agent networks, and probe how bureaucratic barriers and vendor pushback were sidelined. The conversation centers on who controls lethal AI, supply chain labeling, and whether faster AI makes war more or less likely.
AI Snips
Chapters
Books
Transcript
Episode notes
AI Dominance Framed As A Strategic Race
- The memo frames US policy as sustaining AI dominance to reshape military affairs, prioritizing commercial AI innovation as the engine of advantage.
- Mark Fielding and Jeremy note the memo's race language implies China as the unstated adversary driving urgency and doctrinal overhaul.
Unleashing AI First Rewrites Military Rules
- The memo demands 'AI-first' reorganizations, urging experiments that reconceive legacy practices across the department.
- Jeremy warns experimentation is risky because modern LLMs' failure modes and hallucinations are poorly understood in life-and-death contexts.
Anthropic Labeled A Supply Chain Risk
- Jeremy recounts Anthropic's origin and its refusal to allow tech use for mass surveillance or full kill-chain automation.
- The hosts explain Anthropic was labeled a government supply chain risk after adding those user-protection clauses.
