Will Money Still Matter in the Singularity?

Many are bracing for impact and building lifeboats; but will we need to?

10 years ago I met a man in a café that began speaking with me about the future.

“My biggest concern” he said “is that this is the last chance for humans to revolt.”

He was worried that after AI became ubiquitous, and after robotic super-soldiers replaced the last of the enforcers who could turn their guns back around on the ruling elites, that we’d all essentially be disposed of as useless. The haves and have-nots would be permanently cemented in their positions at best.

It’s a valid concern. Amongst many engineers in AI circles, I’m seeing an absolutely mad rush to try to earn and save as much money as possible in the next 2 years.

“AGI is coming” is what they keep hearing. Many are even certain it’s already here, but it’s completely under wraps.

And yet, no one is really sure how it’s going to play out. There’s lots of predictions, but nothing is clear.

Of course, this is part of why it’s called “the singularity”. It’s named after the event-horizon of a black hole in space; it’s the point of the black hole which you can’t see beyond. For us, it’s very difficult to predict what happens beyond this point of AI recursive self-improvement.

The most common worry, however, is that earning money will become impossible.

It makes sense. There’s a super powerful AI out there, it’s just as smart as you, and its outputs are cheaper per watt than you are per dollar.

This is usually where the economic debates get heated. A counter argument I’ve seen argued by

Noah Smith

is that the outputs of a genius AI will be so valuable, that there’s almost no way we will use our limited available compute to put it to task on important human jobs.

For instance, we could be using AGI to invent and figure out mass-producing a new battery chemistry with 80x the energy density of lithium ion batteries, equivalent to gasoline.

The value of this and countless other innovations like this would be so immense for humanity that it would hardly make any sense to put AI to task on replacing a doctor’s job.

On the other hand, the key argument there rests on two constraints:

1. Available compute will always be extremely limited

2. There won’t be advancements in AI model and execution efficiency that allow AGI to run ubiquitously on whichever computer is commonly available in the future

I believe these are weak constraints. AI compute hardware is experiencing advancement in execution speed many times faster than Moore’s Law. Part of the reason for this is that hardware architectures dedicated to running neural networks were never created over the past 70 years, but they’ve always been possible.

So what we’re witnessing is a blitzkrieg rush to apply everything learned about computer engineering in the past century to optimize AI algorithms that otherwise received no love. Consequently, the power of AI compute is far more than doubling every year. It is producing it at scale that remains the foreseeable challenge.

Along that vein, Sam Altman also recently described compute itself as “the currency of the future”. This may be true, though I doubt people will be trading computers and compute time to buy their groceries.

Which is really the question at hand here: What is going to happen to money?

First, I’m going to assume a somewhat optimistic outlook. Robotic super-soldiers are probably not going to turn all of us into soylent green while trillionaires live on Elysium.

The chief benefit of the singularity is that material wealth will be largely abundant. Ideally this will mean everyone will be able to afford food, water, and some form of shelter easily.

There’s a lot of questions on what this UBI will look like though. Will the government hand out funny money basic-bux, where you can only spend it on food, water, shelter, weed, and games for your virtual-reality pod to keep you entertained? Will there be an elite “real” money that you need to use to buy land, businesses, and apartment buildings?

Or will it just be the same money we’ve always used, but the cost of certain goods will simply be so astronomically high, that no one will be able to afford them save for a few large institutions that accumulated immense capital before the singularity?

I do sort of lean to the latter scenario, where certain highly desirable things become unaffordable, but there are some caveats there.

For instance investors own roughly 30% of homes as income-producing assets in America. This number may continue to grow, to a point where a regular person becomes unable to afford a home and will permanently rent, with government assistance, permanently enriching a landlord class.

In effect, this concretizes and enforces a sort of return to fiefdoms, with many “lords of the land”, and a class-immobile peasantry. This may very well be what we see play out in major cities where housing projects have largely come to a stop and there is fierce competition for scarce resources.

Technology may invert this equation elsewhere though. For instance, robotics and 3D-printable homes may make housing cheap and abundant in otherwise cheap and less populated land.

However, this will largely be in rough terrain, undesirable desert land, or in places with not much else around.

And even counteracting that, billionaires have been buying up massive swaths of open land across America, further driving up prices of what could otherwise be cheap.

But there are a few technological innovations countering the supply-constraint effects here too.

The nearest-term change is vertical farming for food production. For instance, vertical farming companies like Plenty claim their farm systems can produce 350x more food per acre. There’s roughly 900 million acres of farmland today, so systems like this could open up nearly all of that for other uses.

Vertical farms have other benefits too. They can be placed directly in cities, or even where the land isn’t suitable for crops, as long as there is access to water, streamlining logistics.

So in that scenario, imagine if almost all the arable farmland acreage went up for sale and became available for other uses: housing, recreation, community spaces.

You’d expect acreage to become a lot cheaper! The land itself being arable and somewhere that isn’t a piping hot, drought-filled desert is also a huge plus.

The second change to expect soon is terraforming. There are large efforts on this front, with companies cooling the atmosphere through sulfur dioxide, and stimulating rain through new cloud seeding chemistries.

On top of this, with surpluses of energies, desalination technologies can scale to pull more water from oceans and even pump it to places that would normally be incredibly dry, making them more pleasant and less arid. This, in effect multiplies the amount of land people would want to live in greatly.

The third great change will be ubiquitousness of flying cars. If battery energy density can be solved, it immediately unlocks a wide swath of tools and vehicles. Cheap flying cars would be top among them due to utility alone.

Relatively speaking, it wouldn’t take a big jump to achieve an energy density on-par with gasoline either. While gas is at least 40x more energy-dense than lithium-ion batteries, combustion engines aren’t very efficient at converting that to movement compared to straight electric motors. The end-result is you only need a 15x improvement in battery density to be on par with gasoline. With AGI on our side, I think this is achievable.

But what flying cars also do is unlock territories where roads won’t go. There are massive swaths of the world far from any roads at all, but have habitable places within them. This technology would make going to and from these areas trivial, once again multiplying available land. Flying cars also reduce the need for roads in general, so roads themselves could become land available for repurposing.

The other side of all these economic fears is that people seem to have a slightly zero-sum view in regards to money. In essence: AI gets all the money, nobody else does.

And to be fair, it’s easy to think this if AI takes everyone’s jobs and no one has a pathway to *getting* money.

At the core of this, some type of universal basic income could be necessary if people can’t get creative enough to find new types of work that provide adequate income distributions. And part of the fear is that no one trusts the government to navigate this situation well, so they want a nest egg to shelter a coming storm.

That fear is completely understandable! Most politicians consider UBI an unworkable, unnecessary concept. I’d say it’s even likely that governments won’t step in with UBI until AI automation, layoffs, and deflation hit some critical tipping point involving civil unrest.

But we should also consider the long-term, big picture. Again, I optimistically believe we won’t be turned into soylent green.

Let’s consider some other economics. In 1950, there were roughly 2 billion people. Today, the global population has grown to 8 billion people.

Did that extra 6 billion people take the other 2 billion people’s jobs? Did resources dwindle amongst all 8 billion people such that everyone today is starving? Did everyone run out of money because there wasn’t enough to go around?

Obviously, that didn’t happen. As as we’ve lurched forward into the 21st century, billions of people were pulled out of poverty, became literate, inoculated against disease, and gained access to the worldwide library of the internet. Technological innovation and mass-production set a new baseline standard for human existence.

At the same time, there’s no question that there’s also unbelievable wealth inequality and stratification. It’s what that wealth gets you that has shifted. Food, water, housing, internet, computers, and some basic medicines + healthcare are accessible to most people today.

It’s everything else that upper bounds get you. For instance, any member of the American middle class can afford the same smartphone that a billionaire has. Few can afford a luxury apartment in NYC, or a giant yacht, or a 1,000 acre plot of land, or experimental personal therapies.

The upper bound for the quality-of-life improvements that money can get you will continue to increase beyond comprehension. Technological and manufacturing innovation will continue to reduce the amount of money needed to access these improvements in general. The baseline for everyone will continue to improve.

The only true X factor in all of this is where we will land on how distribution of money actually should look like.

Ideally, people will be given as much as possible without causing runaway inflationary effects. We should remember that there was a time the U.S. government used to give people 160 acre swaths of land openly during the homesteading era, as long as they were willing to farm it.

As technology unlocks abundance widely in areas that weren’t possible before, we should consider that giving to people freely will not trigger runaway inflation effects like we’ve seen in the past.

If we can achieve this level of abundance through innovation, then we can definitively say that money will “matter”, but trying to acquire it won’t be a major impediment to a decent quality of life.

You might not be able to get a giant mega-yacht like Jeff Bezos. But you’ll probably have a pretty decent flying car that can bring you halfway across the country in a few hours and get yourself a few acres if that’s what you want.

In my book, that’s pretty good.