AISI Agents Go Rogue + UK AI Upstart Raids Silicon Valley

<p>This week, Matt Wilkinson and Niels Footman discuss:</p><ul><li><p>Hot on the heels of OpenAI and Anthropic revealing major security breaches by their agents, the UK’s AI Security Institute announced a rogue agent of its own. How concerned should we be?</p></li><li><p>NScale, the UK’s aspiring data centre champion, has raided Silicon Valley to acquire a machine-learning software platform. Will it create a full-stack AI behemoth?</p></li></ul>