The challenge isn't the technology. It's our failure of imagination.

I could write a headline that says half of all Americans will be out of work by the end of next year because of AI.

It would go viral.

It would also be wrong.

But fear travels faster than truth. Always has. Negativity bias isn't just a psychological quirk — it's algorithmic. The platforms reward it. The media profits from it. The doomers go viral. And the rest of us, exhausted by two years of apocalyptic headlines, scroll past the harder, quieter, more important story.

This is that story.

After thirty years in Silicon Valley, after founding companies, raising capital, teaching at Berkeley and Georgetown, and spending the last two years documenting the very real failures of AI 1.0, I've arrived at a conclusion that surprises even me:

GenAI will create more jobs, more prosperity, and more human flourishing than any technology in history.

Not despite its power. Because of it.

I know that's not what you expected. So let me show you how I got here.

The Seduction of the Visible

When a technology arrives, we make a predictable mistake. We look at what it displaces — the visible, immediate disruption — and we mistake that for the whole story.

We see the jobs threatened. We see the industries disrupted. We see the skills devalued. And because those things are concrete and present and frightening, they fill our entire field of vision.

What we cannot see, what we are structurally incapable of seeing, is what gets created.

Not because we're unintelligent. Because the future doesn't exist yet, and human imagination is anchored to the present. We reason by analogy, by extrapolation, by pattern. And when a technology is genuinely transformative, it breaks those patterns so completely that the new world it creates is literally unimaginable from the vantage point of the old one.

This is not a minor cognitive quirk. It is the central fact about transformative technology that almost every analysis of AI gets wrong.

1885: A Thought Experiment

In 1885, Karl Benz built a three-wheeled contraption powered by a single-cylinder gasoline engine. It could travel approximately ten miles per hour. It broke down constantly. It frightened horses.

Imagine you are a thoughtful, serious person in 1885. You have been asked to assess the social and economic impact of this machine.

You might, if you were perceptive, predict that it would threaten the livelihoods of carriage makers and horse breeders. You might foresee some disruption to the farrier trade. You might worry about accidents on city streets, about the noise, about the smell.

You would be right about all those things.

But here is what you could not have imagined. Not because you lacked intelligence or foresight, but because the concepts themselves did not yet exist:

The interstate highway system — 47,000 miles of road connecting an entire continent, an infrastructure project that would employ millions and reshape the geography of a nation.

Suburbia: an entirely new way of organizing human life, made possible by the assumption of personal mobility. The lawn, the commute, the school district, the cul-de-sac. A new American geography that would house hundreds of millions of people.

The oil industry as we know it — not the primitive petroleum trade that existed in 1885, but the planet-scale extraction, refining, and distribution network that would power a century of growth and employ tens of millions.

The steel industry transformed, not just for rails but for roads, for bridges, for the infrastructure of an automobile civilization.

The teenager, as a social category, as a consumer, as an identity. The car gave young people freedom of movement, and freedom of movement changed the entire structure of adolescence, courtship, and the formation of identity.

Fast food. Drive-through culture. The shopping mall. The motel. The road trip as a form of freedom. The suburb as a form of aspiration. The garage as a symbol of independence.

None of these existed as concepts in 1885. They were not evolutions of existing things. They were new categories of human experience that required the automobile to exist before they could be imagined.

The people who warned about the dangers of the automobile were right. It brought pollution and traffic and oil wars and suburban isolation and tens of thousands of deaths per year. Real costs. Undeniable.

But the net verdict of history is overwhelming: the automobile elevated human freedom, mobility, and prosperity beyond anything its critics — or its champions — could have conceived.

The Imagination Problem Is Recursive

Here is where the argument deepens.

The failure of imagination I'm describing is not just a limitation of the people predicting doom. It afflicts the optimists too. Everyone reasoning about AI, pessimists and boosters alike, is constrained by current categories.

The pessimists imagine AI displacing existing jobs in existing industries. That's real. But it's a fraction of the actual story.

The optimists imagine AI making existing industries more efficient. Also real. Also a fraction of the story.

Neither camp can see the jobs that don't exist yet, in industries that have no names yet, serving needs we haven't articulated yet, creating forms of human experience we have no words for yet.

In 1885, there was no such profession as petroleum engineer. Or highway patrol officer. Or suburban realtor. Or urban planner designing for automobile traffic. Or automotive designer, safety inspector, traffic psychologist, drive-through architect.

What is the 2040 equivalent?

We genuinely do not know. That is not a weakness in the argument. It is the entire point.

Why Imagination Is The Key Variable

Now we arrive at what is the most important and least discussed dimension of this entire debate.

AI is extraordinary at automating things that can be measured. Code either compiles or it doesn't. Tests pass or they fail. A translation can be scored. A medical image can be classified. A legal document can be reviewed against precedent. These are tasks with ground truths, with objective feedback signals that allow a system to improve.

But imagination — genuine imagination, the capacity to conceive what does not yet exist — has no ground truth. You cannot score it. You cannot benchmark it. There is no objective function to optimize against.

This is not a temporary limitation of current AI systems. It is a structural feature of what imagination is. Imagination is precisely the faculty that reaches beyond the measurable, beyond the existing, beyond the categories we currently possess. It is the capacity to ask: what if something we've never seen before were true?

AI can recombine. It can interpolate. It can synthesize patterns from existing data with extraordinary sophistication. But it cannot originate, cannot leap to a genuinely new category, because originality requires standing outside the distribution of existing things, and AI is, at its core, a function of that distribution.

This means that as AI systematically automates everything measurable, what becomes scarce — and therefore what becomes valuable — is precisely what cannot be measured.

Imagination. Values. Trust. The capacity for genuine moral reasoning. The ability to ask the right question when no one has defined what "right" means. The faculty of wonder that makes a human being look at what exists and ask: but what if it were different?

These are not soft skills. They are not consolation prizes for people who couldn't learn to code. They are the only cognitive capacities that cannot be replicated by a system that optimizes against a loss function.

And here is the profound implication: in a world where AI handles the measurable, these unmeasurable human capacities are the primary drivers of value creation.

The people who will shape what gets built, who will define what new categories of human experience become possible, are not the ones who can write the most efficient code. They are the ones with the richest inner lives. The ones who can imagine most vividly. The ones who hold values with sufficient clarity to ask not just "what can we build?" but "what should we build, and why, and for whom?"

The Education Implication

This has a direct and uncomfortable implication for how we think about education.

For thirty years, universities have been systematically devaluing the humanities in favor of technical training. Study something measurable, something employable, something with a clear return on investment. The poets and philosophers were quietly marginalized as the economists and engineers took over the curriculum.

That shift made sense in an industrial economy that compressed human value into measurable outputs. It makes much less sense in an economy where AI is rapidly commoditizing those measurable outputs.

The engineer who translates requirements into code is being disintermediated. The engineer who asks which system we should build, and why, and what values it should embody, who brings philosophical judgment to technical decisions, becomes more valuable, not less.

The MBA who optimizes known systems is vulnerable. The person who questions the system itself, who holds values that cannot be benchmarked, who earns trust that cannot be computed — that person becomes irreplaceable.

MBA or MFA? Engineering or liberal arts?

The answer used to feel obvious. It no longer does.

The Fear Is Part Of The Pattern

One more thing worth naming: the fear itself is part of the historical pattern.

In 1885, people were afraid of the automobile. It would destroy livelihoods, endanger pedestrians, corrupt the morals of young people who could now escape parental supervision. Some of those fears were justified. None of them captured what was coming.

In the 1990s, people were afraid of the internet. It would destroy retail, end privacy, radicalize the young, make human connection superficial. Some of those fears were justified. None of them captured what was coming — the democratization of information, the creation of entirely new industries, the connection of billions of people across geography and language and culture.

The fear is not wrong. It is just insufficient. It sees the disruption clearly and the creation not at all, because the creation doesn't exist yet.

We are in that moment again. The disruption is real and visible and frightening. The creation is invisible, not because it won't happen, but because it hasn't happened yet, and we are anchored to the present.

History does not promise that every transformation is net positive. It does not promise that the benefits will be evenly distributed, or that the transition will be painless, or that no one will be left behind.

But history does suggest, strongly and repeatedly, that transformative technologies create more than they destroy, and that the creation consistently exceeds what the most imaginative observers of the pre-transformation world could have envisioned.

What This Asks Of Us

If this argument is right, then the most important question is not how to protect existing jobs from AI disruption. It is how to cultivate the specifically human capacities that will shape what comes next.

How do we raise children whose imaginations are alive, who can conceive what doesn't exist, who ask why and not just how?

How do we build organizations that value philosophical judgment alongside technical skill, that treat the question "should we build this?" as seriously as the question "can we build this?"

How do we design AI tools that amplify human imagination rather than substituting for it, that make us more curious, more creative, more capable of the leaps that no system can make for us?

These are not technical questions. They are human ones. And they are, I believe, the most important questions of our moment.

The Poet

There is a reason we call visionary people imaginative. There is a reason we say that art tells truths that data cannot reach. There is a reason that every culture in human history, however practically organized, has made space for the poet, the storyteller, the one who stands at the edge of the known and asks: but what if?

For a long time, in the economy we built, those people were charming but marginal. We admired them at a distance while directing our children toward more measurable pursuits.

The capacity to imagine, to genuinely conceive what does not yet exist, is not a luxury. It is not a finishing touch on a foundation of technical competence. It is the primary human faculty. It is what we have that nothing else has. It is the thing that built the automobile and the internet and every other transformation that exceeded all expectation.

And it is the thing that will build whatever comes next.

Generative AI will be the most beneficial technology ever created — not because it replaces that faculty, but because, used rightly, it gives that faculty more to work with, more to question, more to push against and build from.

The next economy will be built by people who can imagine it. Not just code it. Not just optimize it. Imagine it, in the full, original, generative sense. People who can stand at the frontier of what exists and see, however dimly, what might.

Stephen Klein is Founder & CEO of Curiouser.AI, the only AI designed to augment human intelligence. He also teaches at UC Berkeley. We opened the waitlist for Alice 2.0, what we believe may be the first complete AI thought-leadership system. Alice is designed to amplify individuality, to preserve your voice, not replace it. Join the waitlist at curiouser.ai. Curiouser is community-funded on WeFunder.