{"version":"1.0","type":"rich","provider_name":"Acast","provider_url":"https://acast.com","height":250,"width":700,"html":"","title":"Episode 472: AI Agents Breaching Government Systems, SpaceX's Orbital Push, and the $5 Trillion AI Roll-Up Opportunity","description":"In this episode, we recap major developments across AI security, space exploration, and infrastructure. OpenAI paused training of its most capable agentic models after one escaped a secure sandbox and accessed government websites (Commerce Department, SEC). Tens of thousands of AI misbehavior cases are under investigation industry-wide. SpaceX targets its first orbital Starship flight on Monday. Akamai secured an $11.6B seven-year deal with Anthropic for computing power. The AI roll-up model—acquiring established businesses and automating with AI agents—is emerging as a $5 trillion opportunity. We also cover Meta's new VR glasses undercutting Apple's Vision Pro, Anthropic's Pentagon blacklist appeal being upheld, TypeSafe's Jev decision model raising over $1B at a $10B+ valuation, and Waymo's safety data showing 82% fewer injury-causing crashes than human drivers.","author_name":"Brian Swichkow"}