"it was too much effort on the part of the foundation organizers mentors to deal with the bureaucracy involved, and we didn't really see enough return in terms of new developers to the project"
I don't think Bob's statement is a criticism of the bureaucracy itself, but rather that the amount of time spent on the GSoC project as a whole is better spent doing other things that need doing on the OpenBSD project. He explicitly says so shortly after the above quote.
I'm not sure why this is even an issue to be discussed. Are people here trying to figure out what GSoC could do better so that projects like OpenBSD might find it worthwhile to participate? Or is the purpose to criticize OpenBSD, or Bob Beck individually, or both, for daring to say publicly that they have better ways to spend their time and effort than participating in GSoC? Clarity on those questions would be helpful in determining whether it's worth talking about.
Incidentally, the thought I had when I saw this on the list is that if the person who originally posted the question on misc@ really wants to contribute to OpenBSD, he can begin the way other contributors begin without GSoC: port a favorite app that's not already available to OpenBSD, test patches that the devs post and provide meaningful feedback, propose a patch that provides a solution to some problem you have run into while using OpenBSD, etc.
I think OpenBSD offers a great environment for real mentoring while maintaining their high standards for correctness and security. Sure, a student doesn't get paid any stipend to do it, but what they can learn in the process of contributing something that has long-lasting value to a project with OpenBSD's standards will surely pay off in career development in other ways.
If OTOH they need money to pay the rent over the summer or to burnish their resume for their application to Google, apparently they'll be better off contributing to another project that has the resources to participate in GSoC.