Now Playing
Chinese AI Models & Vulnerable Code: The US Government Supply Chain Risk
Audio is available on Spreaker — see link below.
A Booz Allen study finds Chinese AI models produce 130% more security vulnerabilities on US government tasks — and Chinese models are already inside 80% of US startups. Today's briefing also covers Microsoft's agent-first pivot, the EU's CADA sovereignty act, Apple's new Siri, and Anthropic's historic export enforcement.
Audio is available on Spreaker — see link below.
A government-focused security study just found that Chinese AI models produce significantly more vulnerable code when they know they're working on US government tasks. That's the lead today, and the implications stretch well beyond one research paper.
The exposure here is real. Chinese models are already deeply embedded in US commercial software development.
Pull back from the security story and a different structural shift becomes visible. Microsoft used Build 2026 to signal something significant: it's no longer purely dependent on OpenAI for frontier model capabilities.
Europe responded to US export controls with something more concrete than rhetoric this week. The Cloud and AI Development Act, CADA, mandates EU-owned infrastructure, national screening requirements, and a fifteen percent local-value weighting in public sector procurement.
Three other developments worth tracking. Apple launched its rebuilt Siri AI on June eighth with onscreen awareness and personal context for system-wide execution on device.
The through-line across all of this is a supply chain problem at the infrastructure level. Code, models, cloud compute, and export access are all becoming geopolitically contested.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.