What Happens to the Economy When Machines Do Most of the Work?

Imagine it is Monday morning.

Your company has just announced that the new AI system can do the work of thirty people.

Faster. Cheaper. It doesn't need lunch, annual leave, sick leave, maternity leave, motivational workshops, or a manager asking whether it has “capacity for one more quick task.”

Twenty-nine people are let go.

Under the current economic operating system, this is obviously a disaster.

Not because society suddenly forgot how to produce the work.

Quite the opposite.

The work is still being done.

The problem is that the humans who used to do it no longer receive the thing we have attached to work: money.

And without money, they may lose access to food, housing, healthcare, transport and almost everything else.

This creates a strange situation.

We have made production more efficient, yet the people whose labour is no longer needed become poorer.

The machine is doing exactly what technological progress is supposed to do—and somehow that becomes a crisis.

Maybe the problem is not the machine.

Maybe the problem is the economic operating system wrapped around it.

The Old Deal: Work or Else

For most people, the deal is straightforward:

work → salary → survival

You work because you need money.

You need money because almost every material condition required to remain alive has a price attached to it.

This arrangement makes sense in a world where society genuinely needs most adults to work in order to produce enough food, housing, clothing, healthcare, transport and everything else.

But imagine that AI and robotics begin breaking that relationship.

Machines produce more.

Human labour produces less of the total.

Eventually we could arrive at something economically very strange:

Society has enough productive capacity to support everyone, while millions of people still cannot afford to live because nobody needs to employ them.

That would not really be a production failure.

It would be a distribution failure.

We would have solved the difficult problem—how to produce enough—and kept the old mechanism for deciding who gets access to what was produced.

Meet the Automation Dividend

Suppose a factory once needed 1,000 workers to produce one million units.

Years later, AI and robotics allow it to produce five million units with fifty people.

Something has happened economically.

The robots did not merely “take jobs.”

They released productive capacity.

The factory can produce more goods with dramatically less human labour.

That extra capacity has to appear somewhere.

Maybe prices fall.

Maybe profits rise.

Maybe output rises.

Maybe quality improves.

Usually some combination of all four happens.

Call this the automation dividend: the additional economic capacity created when machines allow us to produce more with less human effort.

Now we arrive at the important question.

Who receives it?

Because if almost all of the automation dividend flows to the people who own the machines, while everyone else still needs employment to buy basic necessities, then automation can make society extraordinarily productive and extraordinarily insecure at exactly the same time.

That would be an impressive own goal.

What If Everyone Owned a Tiny Piece of the Machine?

Not literally.

You do not need to mail everybody one wheel from a warehouse robot.

Countries already have mechanisms for capturing part of privately generated economic value: taxation, sovereign wealth funds, public ownership, licensing, resource royalties, equity holdings and dozens of other arrangements.

These institutions are human inventions.

There is nothing sacred about today's particular tax code.

If AI creates a radically new source of economic surplus, society can create mechanisms that allow part of that surplus to return to the population.

That could fund a universal basic income.

Or basic services.

Or some mixture of the two.

The specific mechanism is an empirical question.

The underlying idea is simpler:

If machines increasingly produce the basic wealth of society, people should not need to prove that another human wants to employ them before they are allowed access to a basic share of that wealth.

Not luxury.

Not equality.

A floor.

Food.

Basic housing.

Healthcare.

Education.

Internet access.

Enough material security that losing your job no longer means falling through a trapdoor in the floor of civilization.

No, This Does Not Mean Everyone Gets the Same Life

This is where the conversation usually develops a strange allergy.

The moment you suggest guaranteeing basic necessities, somebody appears clutching a metaphorical yacht and asks:

“So nobody is allowed to be rich anymore?”

Keep the yacht.

Seriously.

Money can continue existing.

Private property can continue existing.

Companies can continue existing.

Someone can still become a billionaire.

Someone can still buy a $40,000 watch whose primary technological achievement is telling the same time as a $20 Casio.

Someone can drink expensive imported water while everyone else drinks perfectly safe ordinary water.

Luxury is not the problem.

The problem is making survival itself a luxury product.

This is why I would not call the idea a post-money economy.

It is closer to a post-money-dependence economy.

Money remains useful for scarce things.

It just stops being the entrance ticket to existence.

What Does the Billionaire Lose?

This is where the idea gets more interesting.

Imagine a billionaire in this future economy.

He still owns the yacht.

He still owns the villa.

He still collects watches.

He still flies privately.

He is still in the top 1%.

His relative material status has not disappeared.

Meanwhile, the person who used to depend on a badly paid administrative job now has food, basic housing and healthcare whether the billionaire hires her or not.

What has the billionaire actually lost?

Perhaps some ability to make desperate people accept terrible working conditions.

But if AI and robotics are already replacing most economically necessary human labour, even that becomes less important.

The machines clean the warehouse.

The AI handles the paperwork.

Robots manufacture the products.

Humans become increasingly unnecessary for compulsory production.

The wealthy person can remain wealthy.

The formerly poor person becomes secure.

This begins to look less like redistribution through sacrifice and more like something economics should find very interesting:

Can technological abundance allow us to raise the floor without substantially lowering the ceiling?

If yes, we may be approaching something surprisingly close to a Pareto-improving transition.

Not because everyone becomes equal.

Because there is simply more to distribute.

Then Why Would Anyone Work?

Some people wouldn't.

This appears to terrify us.

Apparently civilization collapses the moment Brian from accounting is allowed to wake up on Tuesday and decide he would rather learn pottery.

But people do things for reasons other than avoiding homelessness.

They want money.

They want beautiful things.

They want status.

They get bored.

They become curious.

They want admiration.

They want to build something.

They want to help somebody.

They want to win.

They want to prove another person wrong.

They want to understand why their tomatoes keep dying.

Human motivation did not begin with the payroll department.

So paid work can continue.

The difference is that it becomes much more voluntary.

If you want a materially expensive life, earn more money.

Start a company.

Work for somebody.

Invest.

Sell something.

Offer a service somebody genuinely values.

Nothing about a universal floor prevents you from accumulating considerably more than the floor.

But now you are working for more, rather than working because the alternative is nothing.

That is a very different bargain.

And Then We Discover the Really Scarce Resource

Once machines perform more production and humans spend fewer hours maintaining the material machinery of life, something else becomes visible.

Human attention.

We have around eight billion brains walking around on Earth, and an extraordinary percentage of their waking hours are currently spent doing things primarily because rent exists.

Some of that work is meaningful.

Some of it is essential.

And some of it is a human being copying information from spreadsheet A into spreadsheet B so that another human being can paste it into spreadsheet C.

AI is exceptionally well qualified for the sacred cultural tradition of moving text between boxes.

So what happens when human time is released?

People learn.

They explore.

They join communities.

They teach.

They investigate things.

They care for children.

They make things.

They waste time.

They fall in love.

They argue about obscure historical events at 2 a.m.

And somewhere inside all of that, they continually encounter something humanity has an unlimited supply of:

problems.

The World Is Not Running Out of Problems

This is where I think the post-AI economy becomes much more interesting than the usual “robots took our jobs” conversation.

Questions are abundant.

Problems are abundant.

Observations are abundant.

One person can encounter thousands of them during a lifetime.

Why is this service designed so badly?

Why does this teaching method work for one child but not another?

Why did this policy create the opposite result from the one intended?

Why does this plant grow here but not ten metres away?

Why do people behave differently when a rule becomes visible?

Why is everybody solving the symptom while ignoring the constraint producing it?

Today, most of those observations simply disappear.

Someone notices them.

Complains to a friend.

Maybe posts something online.

Then goes back to work.

What if they had somewhere else to go?

A Public Workbench for Problems

Imagine a shared public space where humans and AI can contribute:

a problem,

an observation,

a piece of evidence,

a hypothesis,

a counterexample,

a criticism,

a partial solution.

Other people can challenge it.

AI can connect it to similar problems.

New evidence can strengthen or weaken it.

Nothing important needs to disappear merely because the person who originally noticed it died, changed careers, or lost interest.

And importantly, everyone gets access to the same underlying record.

Not the same opinion.

Not the same interpretation.

Not the same conclusion.

The same record.

I call this broader idea Intermind.

It is not supposed to replace money.

It does something money is terrible at doing: preserving a public history of who noticed what, who contributed what, how ideas connected, and what happened afterward.

And once that record exists, money can begin flowing through it too.

Government may fund useful public contributions.

Entrepreneurs may hire people whose previous contributions reveal unusual skills.

A wealthy person may finance a team working on a problem they care about.

Someone whose old job disappeared may discover that they are unusually good at finding problems nobody else notices.

The old labour market asked:

What job can you do?

The new system might increasingly ask:

What can you see that the rest of us are missing?

This Is Not a Prediction

There is a large pile of assumptions buried underneath everything I just wrote.

Maybe AI will not replace as much human labour as expected.

Maybe robots remain too expensive.

Maybe housing stays painfully scarce even while intelligence becomes cheap.

Maybe governments fail spectacularly at capturing the automation dividend.

Maybe universal income produces consequences we did not predict.

Maybe humans receive unlimited free time and use all of it to argue with strangers about celebrity divorces.

Entirely possible.

That is why I don't think Post-AI Age Economics should begin as an ideology.

It should begin as a research question.

Take the assumptions apart.

Model them.

Test them.

Find the weak links.

Let people challenge them.

Change the theory when reality disagrees.

Because the interesting question is no longer simply:

Will AI take our jobs?

The more interesting question is:

If one day machines can do most of the work required to keep civilization running, why would we continue organizing human survival as though they cannot?

And if the answer is we wouldn't:

then economics after AI may look very different from economics before it.