Mini Posts Collection 3 - August 2024

Compromise Coin, Will AI Detach Wealth from Ability, Power flows Up then Down, Lessons from Revolt of the Public, and Empirical Consequentialism

This is my third in a series of “mini-post” posts. I use these to throw out multiple ideas that I’m not sure I have enough to say about to justify a full-length post (or justify sending you an email for just that thing). Some of these might eventually become full length posts.

CompromiseCoin

The most compelling argument I’ve seen for cryptocurrency is current forms of electronic money make it too easy for an authoritarian government to monitor what people do. That might be okay if we still had paper money to use instead, but the convenience of electronic money means that paper money is disappearing. The hope is that cryptocurrency could give us payments that are as convenient as Apple Wallet, and as anonymous as a pocketful of dollar bills.

However most cryptocurrencies are terrible at this use case. They are too slow, too expensive, too volatile, too vulnerable to theft, and often not even particularly anonymous. What Cryptocurrency is great at is scams, speculation, and cybercrime.

So could we create an online payment system that was good at what paper money used to be good at, and bad at crime? To be good at replacing paper money, you’d want it to be reliably anonymous, cheap, and fast. To make it bad at scams, you’d want it to not support large long distance transactions.

I’m not a cryptocurrency expert, but my hunch is that you could do pretty well by doing something along the following lines::

• Register with a bank, and verify your identity and physical location.

• Purchase a limited amount of “e-cash” (e.g. $1000/month) tagged with a physical region, where the amount you are allowed to purchase depends on how close you live to that region. Call it the e-ATM.

• Use slow distributed protocols to route these coins through tumblers, turning them into anonymous coins in anonymous wallets - where these new coins and wallets are still tagged with your original location.

• At payment time, quickly transfer a coin to another person by sending a request to the bank that issued the original coin. The bank doesn’t know who owns the sending or receiving wallet, but can verify that the coin hasn’t already been spent and update their centralized ledger of which anonymous wallet it is in.

• Route the coins you receive through another round of tumblers.

• Withdraw coins at your validated personal bank, subject to location specific withdrawal limits.

In essence, you’ve got an anonymous cryptocurrency, but the ingress and egress rules combined with location segreation make it hard to move large amounts of money long distances. To make large long distance transactions, you’d have to arrange for lots of people to agree to perform “black market” exchanges of different local coins - and anything that requires coordinating large numbers of people gets hard.

Will AI Detach Wealth from Ability?

We’ve got used to the idea that people become wealthy if they have skills that others value. They might be great at writing software, great at fixing cars, great at managing people, or just someone who works really hard. This isn’t entirely true today, but it is at least somewhat true.

But life wasn’t always this way. It’s not clear that medieval monarchs and aristocrats were any more talented than the serfs they oppressed. What gave you power in the past was mostly the extent to which you were able to coordinate the use of force to dominate other groups.

So what happens if AI makes lots of human skills obsolete? What if AI is better than a human at writing software, fixing cars, managing people, and most of the other things that people do? In a world where most people lack the ability to do work that others value, I can imagine we’d return to a world where wealth is largely detached from ability, and the people with wealth are those best able to coordinate the use of force to dominate other groups.

The great thing about AI taking over human work is that most humans would rather do less work. However the ability to do work that others value gives people power, and most people would rather not give up power. 

Power Flows Up then Down

In a well functioning org, power should flow up, then down.

First power flows up, in the form of low level employees having insights about how their org is functioning and how things can be improved.

Then power flows down, in the form of higher level leaders bringing together all the information that is flowing up to them, and using that information to set a direction that everyone can follow.

If power only flows up, then the org will have no direction. If power only flows down, then the org will have the wrong direction. Power should flow up, then down.

Lessons from The Revolt of The Public

I recently read Martin Gurri’s book “The Revolt of the Public” and it makes a lot of interesting points.

Gurri frames society as being a tension between “the Center” and “the Border”. The Center consists of top-down institutions that organize people around a shared vision to create or maintain things. This includes governments, companies, colleges, and civic groups. The Border consists of decentralized groups that challenge the actions of the Center.

Society needs both the Center and the Border. Without the Center, nothing gets done. Without the Border, the Center gets into pathological states where the things it does are bad.

Broadcast TV allowed the center to go rotten. Prior to TV, the Center needed to create bottom-up institutions in order to influence large numbers of people. However TV allowed the Center to broadcast its vision to a mass audience they had no direct contact with. The result was that Elites become self-selecting, rather than needing to earn the admiration of those below them. TV allowed the Center to become incompetent, while also allowing the Center to hide it’s incompetence.

This changed when first Cable TV and then the Internet broke the center’s control over information and gave power to the Border. There was too much information for the Center to control, and, with channels competing to be the most interesting, they naturally amplified the Border’s messages of corruption and failure, while making it hard for people to have a common sense of what other people believed.

Gurri argues that there are three possible paths forwards:

• Collapse - The Border becomes so much stronger than the center that nothing can be created or maintained, and society falls apart.

• Authoritarianism - The Center follows the example of China, Russia, and Iran, and uses state power to shut down free expression on the internet. An incompetent and corrupt Center prevents the Border from ever challenging them.

• Reinvention - Find a way to use the internet to build a healthy society in which the Center and Border are in balance, where everyone feels they have a path to power, people focus on solving problems locally, and people can coordinate around shared truths.

Like all simple frameworks, it’s probably over-simplistic, but it does seem interesting.

Empirical Consequentialism

People often talk about there being two primary moral frameworks:

• Consequentialism -  Something is good if it is expected to have good outcomes.

• Deontology - Something is good if it follows established moral principles.

I think both of these are dumb, and they are both dumb for essentially the same reason. They fail to account for the fact that impressive rational arguments are often utterly wrong.

The problem with consequentialism is that people are terrible at predicting the likely outcomes of their actions. Smart people made convincing rational arguments that the radical change that Stalin, Mao, and Hitler wanted to apply to society would lead to a Utopia. The problem was that those rational arguments were wrong, just as many rational arguments of today will also turn out to be wrong.

The problem with deontology is that often the moral rules we have are dumb. Sometimes they are leftovers from a past situation where it made sense to do different things. Sometimes a rule exists just because it benefited whoever was in power at some point in the past, but it persists long after those people are dead. Things considered to have been moral axioms in the past include the duty of white people to rule over other races, the importance of using violence to discipline kids, and the immorality of homosexuality. And just like consequentialism, every moral principle has a widely accepted rational argument for why it is right, right up until it has a widely accepted rational argument for why it is wrong.

The problem with both consequentialism and deontology is that they are based on “reason”, and broadly speaking, reason doesn’t work very well.

What does work is empiricism. Do whatever has been shown to work the best in the past, in the situation most similar to the one you are in right now. If nothing similar to what you are proposing has been shown to be successful in the past, or if those successes were in situations different to the one you are in now, then be very skeptical. If in doubt, take small actions before big actions, and use those to find evidence that your bigger action will actually have the benefit you want.

In some cases this resolves to something that looks a bit like deontology. Follow the rules that exist, on the assumption that they have led to good outcomes in the past. But always be aware that you may currently have the wrong rules, and try out other approaches at a small scale to discover if something else might work better.