Open source the layer people need to trust

The open-versus-proprietary discussion is often treated as a philosophical choice. I think it is more useful to ask which layer needs independent trust and which layer represents the service a company is actually building.

If software decides what may run on a personal machine, what data may leave, or what permission an agent receives, there is a strong argument that people should be able to inspect it.

The managed network, economics, and operational service may still be proprietary. The line should follow the trust boundary, not fashion.

Open source is most valuable where verification changes whether people can safely rely on the system.