To serve as a ledger for any major system, you need an architecture that has very high throughput. So if you ever wanted to run, say, the NYSE's ledger, you are

To serve as a ledger for any major system, you need an architecture that has very high throughput. So if you ever wanted to run, say, the NYSE's ledger, you aren't going anywhere if you cannot process many thousands of messages a second, accurately and without duplication. Bitcoin's architecture doesn't really let you get anywhere like that. Adding more computing power won't really help, because you are still maintaining one single ledger, so you have the same synchronization problems as you'd have if you were trying to run a distributed queue with global ordering. There's no way that adding more load will let that sucker scale linearly: Eventually, even if you had infinite computing power, so that it only took you a couple of milliseconds for the next hash to be found, you still have to ferry all the data around to all the nodes of the network, or they can't work: IO becomes a major problem pretty quickly. Big systems like this, but without the crypto requirement, are always IO bound. So to get any serious kind of throughput, you'd end up wanting to have most miners co-located. So a distributed system that anyone can join, as long as they have rackspace in one of 5 buildings, owned by companies that own most of the miners. Doesn't sound all that empowering to me.