RE: From reactive politics to proactive politics (in blockchain governance)
What you seem to be missing is that a 'witness' slot does not have to be a person. We have witness slots on steem getting votes that represent open source projects, community initiatives, and businesses. The people who contribute to running those witnesses are often doing the job as a team, and/or getting assistance from advisors, etc.
It is a reasonable question whether witnesses should be individuals and if so then yes I would agree with you the role should probably be more narrow and friendly to specialization. But you somewhat neglect a third option which is to vote in teams or leaders who will build teams of specialized individuals (I've hired people to help with some of my witness efforts, though I do much of it myself for now).
I've always expected that should this system scale up, the witness role will require more than one person. For example, just being on call 24/7 with high expectations of response time, availability, and reliability is not always something that one person can do (e.g. illness, being occasionally unavoidably inaccessible, etc.). Indeed I would suggest that ability to lead and make use of specialization is something that voters should evaluate (though you make good points about voter apathy as well, and in that regard we are somewhat doomed with a voting-based governance system).
I think it's very probable that the system will evolve in this direction over time (if there are no other big changes to the governance).
But there is still the problem of reactivity instead of proactivity.
When witnesses are required to do everything, it becomes much more difficult for voters to see how they are doing their job. One witness may be great at one thing, but neglect something else. When things go wrong, it's hard to see whose fault that has been.
For example, just look at the current situation of SP for new accounts. Whose fault is that it's not taken care of proactively? It's too hard to say, so voters are not going to change their votes (this is one big reason for voter apathy).
If there was a separate role for somebody to take care of blockchain parameters, problems like this would make clear for everybody that the person/organization in that role isn't doing the job. It would be easy to demand stakeholders to vote for somebody else who is more capable of doing it.
Good point, but then there is a risk of apathy from another angle. With too many different things to vote on, voters can easily find that they don't have the time or interest to understand them and get involved with voting.
I'm not sure exactly what SP situation you are talking about but there are several different considerations related to the SP of new accounts and it isn't a simple problem. That's actually a good example of something where many users would not want to get into the weeds and really understand it.
None of these questions have simple answers though, everything is a tradeoff.