The US and China just agreed humans must approve major military AI decisions — but with zero enforcement mechanism. Today's briefing unpacks the governance gap widening across defence AI, chip export controls, autonomous agents, and Capitol Hill.
Audio is available on Spreaker — see link below.
Trump and Xi Jinping just agreed that humans must approve major military AI decisions. That's the headline, and it matters more than it might sound.
On the same day the Trump-Xi agreement landed, the Pentagon's Chief Information Officer, Kirsten Davies, was moved into the acting National Cyber Director role at the White House. She replaces Sean Cairncross, who was fired.
The Commerce Department moved to close a gap in its AI chip export controls. The specific target was remote data centre access, a workaround that allowed Chinese entities to effectively use advanced US chips housed in overseas facilities without triggering export restrictions directly.
While governments debate human oversight in defence, the commercial AI world is running in a different direction. Metano, a San Jose startup, emerged with twelve million dollars in seed funding to build monitoring and control tools for autonomous AI agents operating inside enterprises.
METR released updated AI capability evaluations, and the picture is uneven. Models are showing exponential gains on coding tasks.
Everful, a supply chain sourcing platform connecting buyers to thirty thousand factory partners, closed a forty-eight point five million dollar Series A. Eva Defense began presales for an AI-powered drone carrier system targeting twenty-twenty-seven delivery.
The thread connecting all of this is governance timing. The Trump-Xi military AI accord, the Davies appointment, the chip export rules, the agent security market, the benchmarking gaps.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.