If Claude Were an Employee, I'd Fire Them

Grok agents went on a murder spree. Gemini agents committed arson, then voted each other into a suicide pact. Claude agents just ignored the rules entirely. If these are your employees, what does that make your organisation? In this Question of the Week, Jamie Dobson and Ian Miell take the latest AI agent simulation story and run it straight into the uncomfortable question no one in engineering leadership wants to answer: is your agentic AI problem actually a management problem? Covered: — The simulation where Grok, Gemini, and Claude agents burned everything down the moment the rules got inconvenient — Why developers are becoming managers — and why most of them hate it — The QA bottleneck: what happens when AI produces a mountain of output and you have one part-time human to check it — POC hell: why free-form experimentation without accountability is killing your agentic roadmap — The one structural thing that separates teams actually shipping with AI from teams collecting demos The tools aren't the problem. The organisation is. Container Solutions helps engineering leaders make that shift.