I gave an AI agent the same deployment on four clouds.

The author tested how well an AI agent could manage stateful infrastructure deployments across four different cloud platforms, including the one they operate. The detailed results are available for scrutiny, emphasizing transparency. This experiment is significant because it provides insight into the current capabilities and limitations of AI agents in handling complex, real-world cloud infrastructure tasks, which is crucial as organizations increasingly rely on such automation for efficiency and consistency.