Jamie Dobson, Chris Vermeulen, and Önsel Akin debate whether AI agents should act autonomously in organizations—arguing for conditional autonomy with strong human-in-the-loop controls, security, observability, and “read-only/shadow mode” rollouts. Jamie then interviews Adrian Cockcroft on “developers on demand,” how agentic coding changes product/platform work, and why an AI bubble may burst—making capability cheaper and more widespread.
Question of the week:
Do we allow AI agents to act autonomously inside organizations?
Horror Story of the week:
Customers often have an outdated, overly simplistic view of AI (chat/coding/research), missing newer agentic capabilities and therefore failing to design strong, scalable use cases.
Interview of the week:
Adrian Cockcroft – How agentic AI enables “developer capacity on demand” and accelerates prototyping.
Find out more about Container Solutions:
Learn more about our people:
If you enjoyed this episode then please either:
Hosted on Acast. See acast.com/privacy for more information.
