So you have to pick one and fix it. It's an engineering problem.
So for example, you have to fix campaign finance. The people in Washington aren't going to fix campaign finance in Washington because the system that exists is the one that got them elected. Ditto for the people in Sacramento, Albany, etc. But wait, you can have Sacramento and Albany fund the federal campaigns of Senators running in California and New York (and so on). State Senators don't get money from K Street, they get money from an entirely different set of lobbyists, so you can get the bills through by taking advantage of the regional weakness of the specific opponents. And then do it the other way around and have the federal government pass a national campaign finance bill for state-level offices.
That is obviously not a trivial undertaking, but we can't jut say "damn it, we're stuck in a cycle" and throw up our hands or it'll never get fixed.
That's a human problem, not an engineering problem. It can still be solved, but it won't be solved with an engineering approach.
How many millions of voters thought the Democrats were so conservative they went green? After the dust settles suddenly Obama has a chance to compete with the rich old white guys.
Change only comes from outside pressure. I'm proud to vote against the status quo.
So bootstrapping is not an impossible problem to solve. What makes politics different that an improvement in politics can't be bootstrapped?
We can't get better policies until we elect better parties, but we can't elect better parties until we address campaign finance and fix the electoral system, but we can't address campaign finance and fix the electoral system until we elect better parties willing to pass better policies. And the merry-go-round goes on and on.