I don't dispute that this is mostly theoretical. But I wouldn't say it's naive, because we know that when people buy into systems they follow the rules as best they can. The problem occurs when others take advantage of the system to game it with plausible deniability, within the confines of said system. That's the entire point of "the emperor has no clothes" allegory. When people buy in to a broken system, broken things happen.

That's why you get papers "proving" someone's guilt, but no action taken because the handling of such situations lies outside of the common understanding of the system. I'm sure most people think to themselves "I'm sure if the papers were legitimate then that person would be in jail, there must be something else going on". When you codify the system entirely, and people buy into it, then someone submits papers recognized by the system to be authentic that prove someone is guilty of a crime, it's much more likely there will be enough pressure to take action in that case, because then people are thinking more along the lines of "those legitimate papers proved they're guilty, why are they still not in jail?"

Obviously one major problem is how to build a system which handles all of the edge-cases that crop up? One possible answer is to use a DAO which can modify itself through whatever means (usually democratic).

> I'm sure most people think to themselves "I'm sure if the papers were legitimate then that person would be in jail, there must be something else going on". When you codify the system entirely, and people buy into it, then someone submits papers recognized by the system to be authentic that prove someone is guilty of a crime, it's much more likely there will be enough pressure to take action in that case, because then people are thinking more along the lines of "those legitimate papers proved they're guilty, why are they still not in jail?"

Or people think to themselves "well, the papers are authentic but the meaning is open to interpretation" or "well, it's not such a bad thing" or "well, but the others are doing worse", or a myriad of things. All of those are real arguments I've seen made.

It is naive to think that you can codify everything into a system in a way where there's no open interpretation possible. It is naive to think that people will trust the system instead of what they heard the system says, or that people will trust the system over leaders, or the the people responsible for enforcing the system decisions and feeding the system information won't find a way to elude responsibility.

In general, it is naive to think that the problem of corruption, which is a problem about trust (if the person I like says he didn't do it or did it for a good reason, then I trust him above all else), enforcement (the ones responsible for enforcing the rules are the ones breaking them) and tolerance (people don't care that much about corruption if it doesn't harm them) will be fixed by a solution that at most will help a little bit in the enforcement part.

I think of it more in terms of a highly secure and decentralized online videogame. Most people will be happy just to play the game, some people will start guilds and min/max, some people will try to cheat. When someone is found cheating, the system gets patched to prevent others from doing it. The dynamics of a game nudges people into behaving within the incentive structure of the game mechanics - even though technically they could strip naked and /dance for a month straight.

In any well designed online game (please don't use shitty ones against me), cheaters aren't able to hijack the entire system or ruin an economy. Their exploits break the game temporarily and most of the time it doesn't affect that many people. In bad cases, the state gets rolled back, the exploit patched, and the game continues. Sometimes things are just unavoidable, and everyone counts their losses.

I think blockchain has the potential to make fairer and more resilient systems, and with new incentives as well. We can use game theory to try wildly different societal structures, and quickly switch to another one if one doesn't work well.

On some level, trust always has to be present. If we make the rules trustless, enforcement becomes the bottleneck (until Robocop?), but it's still potentially better than what we have now where we simply have to trust millions of other people not to fuck with anything too much.