{"version":"1.0","type":"rich","provider_name":"Acast","provider_url":"https://acast.com","height":250,"width":700,"html":"","title":"AI Becomes Infrastructure: OpenAI Spend Signals, Defence Targeting, and the Real Risks of Agents","description":"Claire and Peter break down today’s biggest shifts as AI hardens into critical infrastructure. They unpack reports of internal tension at OpenAI over IPO timing and massive infrastructure spend—and what that could mean for pricing, rate limits, and vendor concentration risk. They then look at the rapid expansion of Project Maven into AI-assisted battlefield management, highlighting how safety principles, governance, audit trails, and human-in-the-loop controls are becoming table stakes for high-stakes decision support. The episode also covers Taiwan’s investigations into covert talent recruitment, Wipro’s acquisition of Mindsprint as a signal of domain-led, long-horizon AI services deals, and new research on how agentic systems leak data through memory, tools, and agent-to-agent communication. Finally, they explore why full automation is often uneconomic, why hybrid human+AI systems win in practice, and quick hits on GeoAI fairness, AI-powered microgrids, resilience gains from structured adoption, and a labour market that’s surging for senior roles while squeezing entry paths.","author_name":"Six & Flow"}