{"version":"1.0","type":"rich","provider_name":"Acast","provider_url":"https://acast.com","height":250,"width":700,"html":"<iframe src=\"https://embed.acast.com/$/68a43f4573bf5b62987006aa/69d7277634b90cef2bcd141e?\" frameBorder=\"0\" width=\"700\" height=\"250\"></iframe>","title":"AI Gets Embedded—and Complicated","description":"Claire and Peter break down today’s biggest moves shaping how AI lands inside real organisations. Anthropic withholds a powerful vulnerability-finding model (Claude Mythos) in favour of controlled access via Project Glasswing, while simultaneously pushing enterprise adoption with Claude Managed Agents for secure, long-running workflows. They also unpack the operational uncertainty created by Anthropic’s Pentagon blacklist dispute, what it means for federal procurement, and why multi-vendor strategies matter. Plus: Amazon’s massive AI capex debate, Canva’s shift toward an AI-led work platform, Taboola’s on-site answer engine momentum for publishers, DeepSeek’s delayed V4 and the hardware speculation behind it, China’s booming AI microdrama market and likeness-rights rulings, Utah’s AI sandbox as a regulatory template, and ABBYY/Tecala’s agent-driven document automation for AP and claims.","author_name":"Six & Flow"}