AI Can Make Abundance. A 2% Inflation Target Will Offset It

Elon Musk told Peter Diamandis something that sounds casual until you sit with it. At the Abundance Summit, already in trillionaire territory, he said AI and robots will make so much stuff that money will stop being relevant. He said the same thing to The Economist. He expects prices of stuff to fall. He talks about a high income for everyone, paid whether you work or not. Print money. Let the machines work.

I want that world. I have spent years arguing that technology is a resource-liberating machine, and I still think Diamandis is directionally right: abundance is a life of possibility, not a life of luxury. I even wrote about his TED talk back in 2015. The problem is what happens after the machines can make almost everything.

On 12 August 2026 Michael Saylor answered Musk in public. Goods can get cheap. Status does not. Give everyone a house and someone wants one twice as big. We are status-oriented animals. You cannot code extra coastline or photocopy a prime address. Vinod Khosla has been making a related point for years: technical plenty does not automatically become shared living standards. Something in the plumbing has to let the surplus through.

That plumbing is monetary policy. And the people who run it already have a job description that fights the outcome Musk is promising.

Asimov's three patterns

I keep rereading Asimov’s Robot books. In Robots and Empire he already mapped the fork.

Earth in the Caves of Steel era is dense, short-lived, suspicious of robots, and politically brittle. The Spacers are the wealthy pole: long-lived, robot-saturated, low population. Solaria is the extreme. In The Naked Sun it has about 20,000 humans and on the order of 10,000 robots each. Comfort does not erase them. They are still there near the end of the saga, in Foundation and Earth. Luxury did not finish them off. It froze them. A society can stay rich, stay automated, and stay still for a very long time. The Settlers are the middle: people who leave the Cities, limit their dependence on robots, accept shorter lives and more risk, and push out. Asimov's point is simple. Total comfort does not have to kill you. It only has to remove the need to reach.

The reports already agree we are past the limit

Optimism about AI abundance has to walk through a door that policy people have been standing in for fifty years. Both big camps in the sustainability fight agree on the diagnosis. Humanity is in overshoot. They fight about the medicine.

One camp says we can keep getting richer if technology does less damage per dollar of GDP. That is the UN green-growth line, from the Brundtland Report onward. The other camp says you cannot grow material use forever on one planet. The Club of Rome's Limits to Growth said that in 1972. Scientists later drew nine “planetary boundaries.” Six were in the red in 2023. The 2025 update put it at seven, after the oceans crossed too.

In July 2026 Keegan, Cafaro, Rees and co-authors argued in Sustainable Development for a voluntary path toward about 4 billion people by 2200. You can reject the 4 billion target. The price point still stands: fewer people, same beaches and city cores, higher rents, even if robots make the manufactured stuff almost free.

I wrote years ago that economics still lacks a true constant, the way physics has c and G, and that Earth itself is the only candidate that stays fixed. Once you treat the planet as that constant, open-ended material growth is a bookkeeping error. A monetary system that pretends otherwise is hiding the scarcity, not removing it.

What a 2 percent inflation target does when machines get cheap

Central banks are told to keep the cost of living rising slowly, about 2 percent a year. That is what they call price stability. It is also a floor. If robots make a fridge cheaper every year, they treat that as a miss and cut interest rates to push prices back up.

They do not mint most of the money in the economy. Banks do, when they lend. The Bank of England spelled that out in 2014. The central bank sets the price of borrowing. In a crisis it buys bonds. Then credit gets easier or tighter, then loans, then the amount of money. Late, and uneven. When machines suddenly make more stuff, the extra value shows up first in house prices and company valuations, not in a typical paycheck.

US data since the early 1970s is the rhyme. Output per hour kept rising while typical pay did not keep up. The official charts and a widely used reconstruction tell the same story: from 1948 to the early 1970s the two lines moved together; after that they split. The numbers are public.

On 15 August 1971 Nixon stopped exchanging dollars for gold. The last hard limit on dollar credit came off. The Club of Rome published Limits to Growth the next year. East-West systems institutes in Vienna and Moscow were running the same global-limit models. The dates are on the record. From that decade on, the tool that scaled was credit.

Steam, electricity, and computing already forced money systems to adapt. People left farms for factories. In the late 1800s Britain and America had decades of rising output with slowly falling shop prices, and pay still rose in real terms. A BIS study of 140 years and up to 38 countries found cheapening goods are a weak predictor of a slump, except in the Great Depression. Falling house and stock prices are what hurt. That does not stop the mandate. If prices fall, they have missed the 2 percent target.

So if AI collapses the cost of things you can copy, the 2 percent rule says: do not let prices fall. Offset. Not by dropping cash in every account. By cheaper credit, buying assets, and making sure sitting on cash does not get you richer. A 2 percent target will not let a fridge get cheaper every year while your bank balance sits still. What you get instead is familiar: gadgets stay roughly the same price or drift up, the extra value piles into land and status, and daily life looks “fine” because stuff is cheap while space, time, and location get dearer.

If the population eases toward that 4 billion path, the split gets sharper. Fewer buyers of cars. More bidders per hectare of coast. Tax the land, not the building might become the rule that drives the economy in that times.

Not Solaria. Not the Settlers either.

Mars will not save this in the near term. Robots will do the dangerous early work cheaper than humans. Until closed-loop habitats exist, packing millions of people into tins is not an individual economic decision. We will not become Caves of Steel Earth either. The machines are already in production, logistics, and care. And we will not become Solaria with eight billion people still here. Twenty thousand humans on a whole planet is a different species of scarcity.

What you get instead is a hybrid, heavily automated, still crowded Earth with enough to live. Discretionary surplus scarce, sitting in land and positional goods. Musk's “money stops mattering” can be true for a box of calories and basic needs. Saylor is describing that split from the asset side. The 2 percent regime describes it from the institutional side.

This will not go to a ballot. Interest rates and what the central bank buys get set in rooms most people never see. The 1930s taught elites that an open fight over money during a crisis can break a democracy.

I do not know what anyone in those rooms intends. Intent is the part I cannot see. The outcome I can see. A mandate that treats falling goods prices as failure will, by default, capture a large share of AI productivity before it shows up as a cheaper life. The machines can still produce abundance. The question is whether the institutions that currently manage scarcity will let you keep it.