{"version":"1.0","type":"rich","provider_name":"Acast","provider_url":"https://acast.com","height":250,"width":700,"html":"","title":"AI’s New Friction Points: Lawsuits, Agents, and the GPU Squeeze","description":"Claire and Peter break down a busy day in AI: xAI sues Colorado over a new “high-risk AI” law, raising the stakes for state-by-state compliance and enterprise risk mapping. Inside big tech, Meta reorganises around agentic tooling—and quietly learns the hard way that measuring employees by token usage drives the wrong behaviour and runaway costs. Google DeepMind’s Demis Hassabis argues focus and centralised compute are what put Google back on the frontier, while Microsoft faces a “Copilot code red” and internal demand that could tighten Azure capacity for customers. On the infrastructure front, Anthropic explores custom chip design as Claude demand surges, and Nutanix expands hybrid and sovereign-ready options for running agentic AI. Plus: Tata Power standardises on Databricks and rolls out a talk-to-data agent, Arkansas deploys AI-assisted work-zone phone detection with human verification, and TCS signals GenAI services are now a material revenue line. The throughline: AI is more embedded—and more expensive—so governance, measurement, and infrastructure strategy are becoming the real differentiators on AI in Flow.","author_name":"Six & Flow"}