So when you think about tracking on-chain transactions, what you're really talking about is tracking government-level, giant transactions that are largely divorced from the kind of detailed forensics that people would want to do. Which... there are better ways to audit giant government-level agencies than a blockchain.

Similarly, leveraging the blockchain to handle things like proof-of-identity doesn't work if the transaction speed for ordinary people can't keep up. If you want to use Bitcoin for proof of identity transactions, you need ordinary people interacting with the network, not a central exchange. Because again, if you have a few central exchanges that everyone is interacting with, then it's almost strictly better to just let them handle proof of identity and proof of ownership in shared centralized databases.

The only reason Bitcoin could matter is if it actually did scale enough that it was feasible for ordinary people to use it for micro-transactions and as a performant, robust API. But the problem is that Bitcoin as a technology is poorly suited for that use case, and none of the people hyping are interesting in solving those problems, evolving the technology, or moving to other coins that would be better suited.

In a world with centralized exchanges and secondary layers on top of bitcoin that consolidate and batch transactions, all of the problems you're describing end up being easier and better to solve by just having the centralized exchanges coordinate with each other, the same way that banks already do. And to the extent that banks don't coordinate with each other right now, it's unrealistic to assume that the blockchain is going to suddenly change their incentives or force them to do so.