You're totally right. But let me flip the question around a little bit. What would our financial system look like if we didn't have to place so much trust in our intermediaries?
It's not that banks aren't trustworthy. They are. They're extremely good stewards of the public trust, for the most part. But the fact that we place so much trust in them has systemic effects that are stifling to innovation. And I think that's what blockchains may allow us to overcome.
I think risk is associated with stores of value. "I rob banks because that's where the money is." So we move the stores of value from banks to blockchain exchanges and hot wallets, and the risk moves accordingly. Now, instead of the risk being borne by organizations with decades or centuries of experience mitigating it, backed by the Fed, it's borne by people who barely understand the wallet software they downloaded, and Magic: The Gathering card traders who aren't as smart as they thought they were, and people who are absolutely definitely sure that smart contra--oops, let's just "fix" that.
So what you describe as centralized stores of trust stifling innovation can also be described as centralized stores of risk mitigating loss. And I think that's pretty much what we've seen so far.
Large banks have a lot of money to buy influence; that's the source of their systemic risk.
But you can do wire transfers at small banks and credit unions.
I think at that point it becomes fair to ask: what is the likelihood that I am going to lose money unexpectedly to my credit union, or my recipient to her bank? And then by comparison, what is the likelihood that I or my recipient are going to lose money to coin theft, exchange failures, fluctuation in valuation, or other things inherent to dealing with blockchain transfers?
Rightly or wrongly, I think many people would rate potential losses higher with blockchain solutions than with USD wire transfers. I know I would.