Techstrong Gang

Meta launches Muse, an autonomous AI agent built to handle real-world tasks. Qualcomm locks in an AI chip deal with Amazon worth up to $60 billion. And Oracle has a rough week — enterprise spending cuts per the latest ETR survey, plus a House panel subpoena over a $17 billion VA health record overrun.

Congress is moving fast on AI agents — a new bipartisan House bill targets rogue AI agents in the wake of high-profile OpenAI breaches, and a proposed AI superintelligence ban would carry prison terms and corporate shutdowns for violators. Organizations are already struggling to detect and contain agents operating out of scope. Also on the show: whether copyright law loses to national security claims, and New York City’s move to expel AI from its public schools entirely.

OpenAI drops Astra, its first model to cross the “critical” cybersecurity risk threshold — and the company is already calling this the start of the AGI era, even as a $1B Daybreak initiative aims to protect frontline defenders and questions swirl about how tightly watchdogs are being kept on a short leash after OpenAI’s bots went rogue. Then: a mystery outage takes down ChatGPT, Claude, and Grok at the same time, right as GPT-6 launch rumors peak — raising uncomfortable questions about whether “multi-model redundancy” is real if every provider shares the same underlying infrastructure. Finally, the data center backlash goes to court: Kansas cities are using legal action to block local data center petitions, Trump is warning opponents they’re “backwards and poor,” a pro-AI group is running ads to make the public case for build-out, and the G20 has endorsed the U.S.-backed Carolina Principles for light-touch AI governance.

Host Mike Vizard is joined by Jack Poller, Tracy Ragan, and Jon Swartz.

JFrog’s CEO drops a bombshell claim: source code no longer matters as AI takes over engineering workflows — and JFrog is racing to secure agentic pipelines instead. CrowdStrike pushes AI-agent security to the endpoint, unveiling a new platform alongside its Project QuiltWorks AI initiative and a fresh approach to securing software supply chains. Plus, Broadcom unveils an AI Factory built on VMware Cloud Foundation and stakes its claim on Platform Engineering 2.0, while also shipping a new service to secure the Spring Framework.

Mike Vizard and Alan Shimel are joined by Jon Swartz, Jaime Hampton, and Sandy Carter to break down what it all means for engineering, security, and AI infrastructure teams.

Apple hands the CEO reins to John Ternus in a $4.5 trillion transition as OpenAI and Anthropic drive Mac hardware into short supply. Anthropic slashes enterprise AI context-caching costs 75% with Claude Fable 5.1 and Mythos 5.1, while startup Guickly emerges from stealth to help enterprises control ballooning AI spend. Plus, Meta Glasses face a criminal complaint in Germany from a digital rights group.

Mike Vizard and Alan Shimel are joined by Teri Robinson, Elizabeth Safran, and JP Morgenthal to break it all down.

A financial watchdog warns of catastrophic risk from advanced AI models, China lays down preconditions for U.S. AI safety talks while targeting Anthropic directly, Debian’s developers vote on how much accountability AI-assisted code needs, and Meta starts testing robots inside its AI data centers.

This episode covers hardware’s resurgence — including a new hardware fund from software’s biggest champion and DoJ antitrust scrutiny of Andreessen Horowitz — plus quantum computing’s gap between market hype and working machines, anchored by IBM’s HRL Laboratories deal. The panel closes with New York’s overtake of San Francisco as the top North American tech hub.

This episode covers Trump’s push for new semiconductor tariffs and the alarm it’s raising over U.S. AI growth, OpenAI and Anthropic’s warning that the window to defend against AI-driven cyberattacks is closing, and how autonomous killer drones — including Ukraine’s Delta battlefield AI — are becoming cheap and real.

NVIDIA is reportedly acquiring Hugging Face for $12.9 billion, Bill Gates is warning world leaders that AI-driven job disruption is coming faster than they’re prepared for, and Broadcom is lining up as much as $100 billion in financing for its AI chip build-out — raising the question of who’s really paying for the AI boom.

A look at whether the AI agent race is measuring the right progress, the EPA’s rollback of public notice rules for data center air pollution permits, and Meta’s landmark settlement of up to $16.7 billion over its handling of teen users — plus the platform changes coming to Instagram and Facebook.