...let me take the Devil's Advocate position. The more "general purpose"
our vehicle concept becomes, the better suited to manned operation.

The advantage to robotic missions is that they are one-shot, one-design, disposable systems optimized for a specific task. They only have to work once. This is what makes robots so convenient and economical. The high flexibility and adaptability applications start looking more suitable for manned crews, especially if you are trying to design for unanticipated future missions.

I guess the trick is to implement a program that blends both approaches in an optimal fashion, which argues that each mission be conceived as part of an overall general space strategy, a comprehensive space policy.
This is what NASA should be working on, not feel-good projects with big political payoffs and monster price tags.

This policy should take into account both long-term goals, opportunistic experimental packages, private enterprise participation, and the space programs of other countries. And planning should be on a long term basis, independent of the current administration's pet projects. I realize this is easier said than done, and I'm not the first to put these ideas forward, but it should always guide our thinking, even if opportunities or setbacks occasionally force us to improvise something.