Star Wars vs. Star Trek: Which Future Are We Building?

Elon Musk and Dario Amodei imagined as Darth Vader and Spock. AI-generated illustration.

I would rather live in Star Trek.

There are qualifications. Starfleet has an alarming habit of sending its senior management into situations that would normally require a fairly substantial risk assessment. I would probably stay on the ship.

But the underlying proposition appeals. Your access to food, shelter, education and healthcare does not depend on making yourself economically useful to somebody who owns more than you do. People still work, often very hard. They have responsibilities and ambitions. They can also choose a life organised around curiosity, service or the pleasure of getting good at something.

Star Wars has astonishing technology too. Intelligent droids. Interstellar travel. Engineering on a scale that makes our current infrastructure ambitions look rather modest.

It also has slavery, poverty and a great many people whose lives are determined by powerful people they have never met.

That is the comparison I keep coming back to when I think about AI. We could make extraordinary technological progress and still end up somewhere I would very much prefer to watch from the sofa.

The picture accompanying this article puts Elon Musk in Darth Vader's armour and Dario Amodei in Spock's uniform. A little unfair to both, probably. But it gives the question a face.

Star Wars vs. Star Trek. Which future are we actually building?

We already produce enough food

Before we get to humanoid robots and replicators, there is an awkward fact about the world we already inhabit.

The Food and Agriculture Organization says that we produce enough food to nourish everyone. Its account for World Food Day on 16 October 2026 puts the number still facing hunger at 645 million. Far more cannot afford a healthy diet.

Food is produced in particular places, needs to survive a journey and has to reach people who may be living through war or the destruction of their local economy. The World Food Programme identifies conflict as the leading driver of hunger in the worst food crises. Ending hunger involves a great deal more than matching a global total to a global population.

Even so, enough food and widespread hunger exist at the same time. We have managed that combination without any help from artificial intelligence.

A family unable to afford a healthy diet cannot eat a statistic about global agricultural output.

This matters because so much of the optimistic case for AI assumes that producing enough will eventually solve the problem of people having too little. We already have reason to question that assumption. Better harvests, storage and transport would help. So would peace, functioning institutions and people having a reliable claim on what is available.

The machines can contribute to several of those things. They cannot decide that everybody is entitled to eat.

Star Trek imagines a society that has made that decision. Replicators are the memorable prop. The social settlement is the plot.

I can see why people believe the promise

I am enthusiastic about what these tools make possible. My own website is a fairly visible consequence of that enthusiasm.

I have worked with Kiro to take ideas through specification, design, implementation, testing and documentation. There are things on that site which I could describe quite clearly but could never have justified commissioning. Some would have remained ideas indefinitely. Now I can participate directly in building them.

That is a considerable change for somebody whose useful contribution was previously much easier to express in words than in code.

It has also given me ample opportunity to discover how convincing a wrong answer can be. I have written about sessions that revisited rejected approaches and diagnoses that the logs disproved. Knowing what the tool is trying to achieve, keeping a record and checking the result remain part of the work.

The gain survives those qualifications. More people can attempt things that would once have required money, specialist availability or months of somebody else's time. A small organisation can try an operational improvement. Somebody with a particular interest can build a tool around it. An experienced team can take on a scope that would previously have been unreasonable.

I find that exciting. I would like far more people to have that experience.

And it is easy to see how the argument expands. If software becomes easier to build, then the organisations running through that software can change more quickly too. Add capable physical machines and the prospect becomes much larger: less dangerous work, less drudgery, more useful things available at lower cost.

There is a recognisable route towards parts of the Star Trek proposition here.

There is also an invoice for the services I use, issued on terms I have very little influence over. My new freedom depends partly on somebody else's infrastructure and their continued willingness to let me use it.

Both experiences are real. I can build more than I could before, and I have acquired another dependency.

Musk has chosen his droids

In his September 2026 interview with China Media Group, Musk described a future of personal robots, vastly increased production and effectively universal high income. He thought there could be at least a billion humanoid robots within ten years, and questioned whether money would continue to matter.

He reached for C-3PO and R2-D2 as the friendly companions everybody might eventually have.

The interviewer then asked how ordinary people would receive their share, and how a few technology giants could be prevented from controlling the resources. Musk's answer was that robots would become so plentiful that they would exceed anything people could want. The exchange is also available in transcript.

He described a technological solution to an ownership question.

I can follow the production argument. If machines can perform a growing range of useful tasks at falling cost, many things should become easier to obtain. Competition could force prices down. People might own machines that give them capabilities previously confined to large organisations.

But there is a considerable assumption between a robot being able to build me a house and my being entitled to the land, materials, energy and machine time required to have one.

Somebody owns those things. They may have other plans for them.

The Star Wars reference is particularly useful here. The droids work. They are capable, helpful and, in several cases, rather better company than the people around them. Their presence has done very little to settle the political arrangements of the galaxy.

We could have excellent robots and still spend most of our lives trying to earn permission to use what they produce.

The journey is somebody's working life

There is a further difficulty with the promise that work will eventually become optional. Most people still need it to pay for this month.

In *Jevons and Baumol Walk into a Pub*, I explored how output can grow while the number of people required to produce it falls. One of the earliest consequences may be a vacancy that never opens, rather than a redundancy that makes the news.

Anthropic's study of its own workforce, published on 16 October 2026, offers a recent example. Junior employees' share of technical hires fell from 27 per cent to 9 per cent. Technical hiring continued to grow in absolute numbers, but managers described reduced demand for lower-band roles as Claude took on more coding and bug finding. A rapidly expanding AI company is an unusual workplace, and this does not establish what will happen everywhere.

It does make the transition easier to picture. Experienced people direct more work through agents. The organisation grows. The graduate who would once have joined the team has to look elsewhere.

For the person who owns the company, increasing production with fewer people can be excellent news. For someone whose main asset is their ability to work, the same development can remove both income and the opportunity to acquire experience.

Calling the eventual destination abundance does not pay for the years in between.

If work is going to become less necessary to production, we need to decide how much access to life should continue to depend on having a job. Otherwise we are asking people to celebrate their approaching freedom while making them steadily less able to afford it.

That feels much closer to Star Wars than Star Trek.

The uniform is not a guarantee

Amodei has at least acknowledged that the distribution question needs an answer.

On 18 October 2026, Anthropic announced its Public Benefit Fund. Subject to shareholder approval, it proposes committing 5 per cent of the company's shares to a wholly owned entity. Further allocations could follow at the board's discretion as the company's valuation increases. Donations and distributions to particular populations are among the possible uses. Details of governance are still to come.

That is a commitment worth taking seriously. I would much rather see companies considering how the value they create can support people affected by the change than assuming the matter belongs entirely to somebody else.

On 19 October 2026, Musk pledged a SpaceX share for every child in America in 2027. The delivery arrangements were unspecified when announced. Giving people shares directly raises a useful possibility: they could own part of the productive asset themselves.

Direct ownership has an obvious attraction. People would hold some of the productive assets themselves. What I would want to understand in either proposal is how much security the recipient gains and how much influence they have. The costumes are considerably easier to sort out.

I spend much of my working life examining the distance between institutional commitments and what happens when those commitments are tested. A public benefit fund invites the same attention. Who decides which people benefit? What are those people entitled to expect? Can they challenge a decision? What survives a change of leadership?

A fund might do substantial good while leaving all the consequential decisions with the company. A share can provide a financial stake without giving its holder any practical influence over the service on which they depend. Both may improve somebody's position. The scale and terms matter.

I would be more persuaded by an arrangement whose beneficiaries could explain their rights than by one whose directors could explain their intentions.

Amodei can keep the uniform for the picture. There is still some work to do before I would call it Starfleet.

Who owns the replicator?

The strongest objection to my concern is that I may be underestimating how widely the capability will spread.

Open models, competing suppliers and falling costs could put useful AI within reach of almost everyone. Small firms could gain room to compete with larger ones. People might run capable systems themselves and move between providers without losing their work. That would make a considerable difference to the future I am describing.

My own experience gives me a reason to take this seriously. Capabilities once beyond my budget are already available to me.

But cheap access can coexist with concentrated ownership. A thousand small businesses might each gain a useful new capability while all becoming dependent on the same few suppliers. The businesses remain inventive. Their provider still decides what access costs and which uses it permits.

The physical requirements matter too. Models need chips, power and facilities. A capable model available for download does not make those resources free. Some costs may fall dramatically while the people controlling the remaining bottlenecks become more powerful.

This is why I keep coming back to the replicator. Suppose somebody develops a machine that can make almost anything, and its output costs very little to produce. That would be a magnificent engineering achievement. The person who owns it could lower prices, sell subscriptions, supply public services or reserve its capacity for the most profitable customers.

The machine does not choose between those arrangements. Its owner does, within whatever rules society has established.

In Star Trek, we are invited to imagine that access to essentials has become secure. In our version, I can rather easily imagine the replicator asking me to upgrade my account.

Getting to Star Trek

I do not expect money to disappear over the next decade. I do expect more organisations to change around AI, with useful gains appearing alongside new dependencies and pressure on the people whose work it changes. How far that goes remains uncertain.

What concerns me is how much can be decided along the way through ordinary contracts, hiring choices and investment decisions. Each can make sense on its own. Together they can determine who owns the useful capacity and who has to keep asking for access.

I think getting closer to Star Trek requires us to make access to essentials more secure as production becomes less dependent on human labour. That can include public services and income support, funded in part from the growing productivity we are so pleased about. The purpose would be to give people a reliable basis for living through the change, including those who have no shares to appreciate on their behalf.

Shared ownership deserves attention too. Worker equity or a public investment fund could give more people a continuing claim on the value being created. Each needs rules about who benefits and who makes decisions. Calling a fund public does not settle either question.

Organisations also have choices to make now. If AI reduces the amount of routine work available to juniors, they need another way for people to learn. If a supplier becomes essential to operations, there should be a practical way to change supplier. The communities hosting the infrastructure should have a say in what they accept and what they receive in return.

Those are manageable questions to ask of a particular company, contract or policy. They give the larger comparison somewhere to land.

I would take sustained improvements in ordinary people's security, choices and share of the gains as evidence that we were moving towards Star Trek. Falling prices would count. So would better public services, more widely held productive assets and people having time they could actually afford to use as they wished.

Company valuations and model benchmarks cannot tell us that on their own.

I would like a future in which somebody can spend more time caring for a parent, studying a subject, making music or building something peculiar because the necessary work takes less of everybody's life. AI has made parts of that prospect easier for me to imagine. I can already do things I would once have abandoned as impractical.

That possibility is worth pursuing well beyond the people who happen to own the systems.

The image is playful. The choice underneath it is consequential. Both futures have impressive machines. We have already demonstrated, with food, that producing enough and ensuring people have enough are separate achievements.

We could spend the next decade getting remarkably good at building droids and still leave most people on Tatooine.

I would rather we aimed for Star Trek.

I work with organisations navigating this shift, fractionally, as an adviser, or as a trusted collaborator. See how I work →