I just read a very interesting summary from Sam Altman about AI governance. The question that caught the most attention was about the scenario of nationalization — like, what if the U.S. government decided to nationalize OpenAI or other AI projects?

His answer is quite nuanced. First, he acknowledges that, in theory, a government-led AGI could be appropriate in certain long-term contexts. But here’s the point: given the current development trajectory, nationalization is not something highly likely. In other words, it’s not exactly the direction things are heading.

What really caught my attention was what he emphasized afterward: close collaboration between government and AI companies is extremely important. It’s not a minor detail. People tend to take for granted the level of safety we have but completely underestimate the efforts and investments behind it. We’re talking about massive resources.

He considers this generally a positive thing, but makes it clear that there is still a lack of respect and understanding of the complexity involved. It’s a reminder that AI governance isn’t a simple matter of nationalizing or not — it’s about finding the appropriate model of cooperation that works for everyone.