I would imagine if Bitcoin continues to expand in 20 years very few users will be running a "satoshi style" wallet and connecting directly to the blockchain.
Agreed! Thought hopefully there will be more net users running the client than today, but it will be a small fraction of overall userbase. Personally, I'll always run the client forever on every computer I own.
But isn't this a problem? Doesn't this remove the decentralized resilience of Bitcoin we often claim is comparable to bittorrent? Doesn't this make it a lot easier to change the rules by someone?
I've asked this question in a couple of other threads and I don't think I received a satisfying answer yet. I'm not trying to spread FUD, I'm obviously very bullish about Bitcoin's future but still it's something that I think we should ask ourselves and find the right answers in order to avoid potential bad future outcomes.
Well, I guess you're asking a few questions:
1) Is a world where many users use ewallets/ebanks more centralized than a world where everyone uses their own client software? Answer: yes.
2) Isn't that increased centralization a problem? Answer: not necessarily.
People need to distinguish between coercive centralization and voluntarily centralization. Centralization of government power, for example, is bad, because people are being coerced into doing what they'd otherwise not do - this is problematic morally, but also can lead to harmful real-world consequences. BUT, this is wholly different than voluntary centralization.
Take the example of Walmart (huge mega-retailer here in the US). Walmart has centralized the distribution of products, and you can find anything you want inside the store. So is this harmful in the same way that coercive government centralization is harmful? I don't think so. If centralization is occurring, but it's occurring voluntarily based on market signals, then that should be taken as a sign that efficiencies and value are being created by that centralization. When the centralization is forced, however, it is unlikely that efficiencies and value are being created (if they were, then the centralization would have occured voluntarily anyway), and thus such coercion should be fought against and resisted.
Bitcoiners are happy to fight against the coercive centralization in the realm of money (vis-a-vis central banks and fiat currency), but sometimes that justified antagonism spills over into also fighting against voluntary centralization, which is problematic.
Centralization can be good or bad, and it typically depends on whether it's occurring under coercion or by choice. If Bitcoiners in the future tend to centralize funds held via ebanks, then so be it. A) there's nothing you can morally do to stop it (other than pleading with people) and B) it likely represents efficiency and value gains which ought not be interfered with anyway.
Does this make Bitcoin less "safe" or less "secure" with voluntary centralization? I'm not sure... it's possible that if safety/security is handled largely by professional private institutions, we'd see a safer and happier blockchain than if it could not be centralized in such a way. Admittedly, some degree of decentralization is lost in this process, but security gains from other dynamics may compensate for this. Also, perfect security is only achievable by spending infinite funds... thus there is always a tradeoff between security and efficiency. The marketplace might discover that the optimal balance means most funds are held by ebanks (but again, you always have a choice to hold your own!)
If all Bitcoins were held at a handful of large US banks, then I'd be worried. But that will never happen, and thus I am not bothered by some marginal or even significant voluntary centralization short of that occurrence.