Now Playing
Amazon's Safety Gambit, DeepMind Loses Two Giants & the $750B Bet
Audio is available on Spreaker — see link below.
Amazon weaponized AI safety concerns to ground a rival's model, DeepMind lost two world-class researchers in 48 hours, and Big Tech's AI infrastructure spend is now approaching $750 billion. Today's briefing covers the week's sharpest signals across regulation, talent, and capital.
Audio is available on Spreaker — see link below.
An Amazon research team stress-tested a rival's AI model, flagged what they found to the White House, and within days that model was pulled offline by government order. That's not a hypothetical.
There are also real legal questions about whether this action holds. The government's position rests on what's called the deemed export standard, which treats distributing AI to any foreign national physically in the United States as an export, including Anthropic's own non-U.S. employees.
Separate from the export standoff, the talent market delivered one of its sharpest weeks on record. Nobel Prize-winning AlphaFold researcher John Jumper left DeepMind after nine years to join Anthropic.
On the regulatory front, the European Parliament adopted Digital Omnibus amendments on June sixteenth, delaying high-risk machinery AI obligations to August twenty twenty-eight. That's a twelve-month extension for industrial AI systems already governed by existing product safety frameworks.
The infrastructure numbers are now in a range that demands serious attention. Amazon, Alphabet, Microsoft, and Meta are projected to deploy a combined seven hundred fifty billion dollars in AI infrastructure in twenty twenty-six, up eighty percent from twenty twenty-five.
Microsoft's answer to that pressure is to build a complete stack and own it. At Microsoft Build, CEO Satya Nadella framed computing as shifting from app-centric to agent-centric.
The metrics worth tracking from here are narrow. On Fable Five, the critical unknown is whether the White House and Anthropic find a resolution framework, and whether any deal sets a formal standard for future government involvement in model access.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.