Summary
Nate B. Jones interviews Alvin Grlin — a researcher at Stanford’s Human-Centered AI Institute, a DC think tank advisor on AI and technology policy, and an industry veteran who has worked across chip design, devices, models, and apps on both sides of the Pacific since the early 1990s — for a contrarian take on the US-China AI competition narrative.
Grlin argues that the arms race framing is largely manufactured: the AI industry has borrowed the military-industrial complex’s playbook of identifying a large, scary adversary to shed regulatory constraints and attract capital. His core claim is that there is no finish line, no winner-takes-all outcome, and no way to prevent software — especially open-weight models small enough to run on a Mac Studio — from distributing globally, much as no nation “won” the electricity race. The conversation covers how open-weight quantized models are commoditizing high-quality intelligence, and why bifurcating global AI systems creates dangerous blind spots, particularly for biosecurity: if US and Chinese AI systems stop talking to each other, neither can flag novel pathogen sequences the other detects.
Grlin also addresses dual-use risks directly, citing research showing that small, publicly available chemical synthesis models can generate weapon precursor routes on a laptop in hours, and arguing that the answer is international regulatory coordination — analogous to existing DNA synthesis screening consortia — rather than decoupling. His background spans Intel’s MMX SIMD instruction set, research at University of Washington and MIT, and current teaching at UDub on AI and technology policy.
📺 Source: AI News & Strategy Daily | Nate B Jones · Published September 13, 2026
🏷️ Format: Interview
