I think we sometimes treat "open" as automatically good without examining the tradeoffs.

You can easily sponsor Iran or Russia killing real people by doing such things.

Powerful tools, once released, can be used by anyone, including those with harmful intentions. And let's be honest: much of open source functions as a way for large companies to cut costs on essential but non-differentiating infrastructure. That's fine, but it complicates the idealistic narrative.

With generative AI, these questions matter more. Maybe it's time to revisit what open source should mean in this context.