Conceptually it is very simple to understand.

The point that Moxie was making at the time was that, pragmatically, most users would end up not doing this. Just as most people share their thoughts on Facebook or Twitter and not on their own blogs, and if they do have their own blogs they don't run their own server, and if they do it's a virtual machine on Amazon's hardware.

So something can be theoretically capable of enabling enormous decentralisation but actually not have that effect at all.

People look at decentralization as an ideology rather then a threat model. The security you might implement to protect $100 vs $100k vs $100 million is different. Running your blog on AWS is running your own server in a decentralized sense; but if your threat model includes Amazon shutting you down or something like the US Govt/public compelling Amazon to shut you down then it might not be good enough. Which is why Target, Walmart, Parler (assuming that's still true) don't run on AWS. Any threat model is about mitigation and acceptance of risk. Easiest way to secure a database/system is to isolate it from the internet and let nobody read/write from it; probably wouldn't be a very useful system for most usecases but something you might do for the nuclear weapons arsenal.

Centralization vs Decentralization is a similar concept to closed source vs open source. Just because the code is open source doesn't mean that people are going to download the source code and compile it themselves. The option to fork the codebase is also there. Neither of those options are possible for a closed source project. And to your point, the fact that a project is open source might not "actually have any effect at all". Doesn't mean the concept is inherently bad.

> The security you might implement to protect $100 vs $100k vs $100 million is different.

This almost makes sense, except that you have to take into the relative value of that amount, in particular relative to one's total net worth. If I want to protect $100, and that is all the money I have, I'll worry about security a lot. If I have $1,000,000, protecting $100 is not a priority.

The general assumption is that people put all of their money in one place, which is not true for anyone I know -- poor, middle class, or rich. Everyone has cash, bank accounts, various physical valuables, stock, etc.

It could be absolute and/or relative depending on the mechanisms. Relying on FDIC insurance to protect $1 million in a bank account is probably not a good idea.

If you cannot trust a single third party to hold your money, it's no better than cash under the mattress.

We have guard rails in banking for a reason.

Hm, I am not quite sure I understand. Are you saying that people should not have the ability to withdraw their USD into cash either? Cryptographic currencies are exactly the same thing imo- they are not meant to replace trusted third parties in the traditional sense, they are just the digital equivalent of cash. A cryptocurrency wallet is far closer to a physical object than being a 'promise' from networks of trusted authorities. Apples and oranges.

banking is for fiat currencies whose value halves ever dozens of years. you dont have to worry about bank robbers. inflation does the same thing.

> whose value halves ever dozens of years

This seems demonstrably better than halving over 6 months...

Or six hours.

Looking at you not-stable coin.

If this is true, crypto is dead in the water and always has been. It's not hard to understand that people are simply not going to go through that trouble in order to own something.

People go thru a lot more trouble to ensure the value of their savings does not go to zero.

They absolutely do not. They put their savings in an account held by a centralized 3rd party (bank) with the backing of government-guaranteed insurance (FDIC).

Maybe some people store cash or gold in a secret place but that is the exception.

How is this any better than the current banking system?