The Great Filter for Ideas

AI removed almost every barrier to creation. Where are the creations?

In Arthur C. Clarke’s Childhood’s End, the Overlords arrive without warning. They are more advanced, more powerful, and more generous than anything humanity has encountered. They end war, eliminate poverty, and cure disease. Every external barrier to human flourishing, every obstacle we had spent millennia pointing to as the reason we couldn’t reach our potential, is systematically removed.

And then something unexpected happens. Humanity loses its momentum. Art production slows. Ambition gradually disappears. People are comfortable, healthy, and free, but they have stopped reaching for anything beyond comfort. The golden age arrives and the species discovers, too late, that the obstacles were never the source of the drive. The drive was something else entirely, something the Overlords could not provide and their gifts could not replace.

Clarke wrote that in 1953. Seventy years later, we didn’t need the Overlords from outer space. We built them ourselves. We just call it AI.

---

I grew up reading Clarke, Asimov, Gibson. For most of my childhood, the idea that occupied me most was the oracle, the machine you could ask anything and receive the truth. I was so taken with the concept that my AIM screen name as a kid was th3oracl3, years before The Matrix put one on screen.

My brother still reminds me of this. He likes to say I was ahead of the curve, that I saw where technology was going before the people around us were thinking about it. I think the truth is simpler: I was a kid who believed that if you could ask the right questions, you could understand anything, and I desperately wanted the machine that would let me.

I spent years imagining what it would feel like to sit in front of a system that could answer whatever question was large enough to deserve asking, and in every version of that fantasy, the arrival of the oracle was the turning point. Humanity leapt forward. The species matured. We became what we were supposed to become.

A few years ago, something resembling the oracle arrived. We can ask it anything.

People are asking it how many calories are in a serving of broccoli.

---

Democratization has been the central promise of the technology industry for as long as the industry has existed. Lower the barriers and people will build, create, participate. The argument has repeated at every turn, from the personal computer to the internet to open source to the smartphone: the reason most people aren’t building is that building is too hard, too expensive, too exclusive. Remove the friction and watch what happens…

For a long time, that story was convincing. You could point to real barriers: access to capital, access to information, access to technical skill, access to the networks of people who knew how to get things done. I believed it. Over the course of my life, I’ve met hundreds of people who carried ideas they genuinely cared about, the book they wanted to write, the app they wanted to build, the business they were certain would change their lives. And in nearly every case, nothing came of it. I assumed the barriers were to blame. It was the only explanation that made sense, because the alternative was too uncomfortable to consider.

Then AI arrived and dismantled almost every single barrier.

You no longer need capital to prototype a product. You no longer need a technical cofounder, or a design team, or a strategist, or a year of runway. The most powerful creative and analytical tool ever built is available to anyone with an internet connection, and the most capable version of it costs less than a gym membership.

Every person who told me they had a novel inside them can now sit with a collaborator of extraordinary patience and skill, for as long as they want, and write it. Every person who sketched an app idea on a napkin can build a working prototype in a weekend. The tool is here. The excuses have dissolved. The conditions for a creative explosion, by every measure we once used, have been met.

And almost nobody is shipping with it.

---

What If the Tools Were Never the Problem?

A few weeks ago, a tweet from a developer named Jim Raptis caught my attention. He was responding to Paul Graham, who had written that in the AI age, taste would become the great differentiator, that when anyone can make anything, what matters is what you choose to make. Raptis replied simply: not a single person he knew in real life actually wanted to ship an app. These were his friends. Many of them worked in tech.

I have watched the same thing play out in my own circles. I know people, technical people, creative people, people with means and education and every structural advantage, who have talked about the same ideas for years. AI did not change them. It did not even accelerate them. The distance between who they are and who they said they would become, if only they had the tools, has not closed by a single inch.

For decades, we had a simple narrative: the impulse to create is universal, and the barriers are external. Remove the barriers and the floodgates open. AI removed them. The floodgates stayed shut. And once you’ve watched that happen, once you’ve sat with it long enough, you arrive at a question that the simple narrative has no room for.

What if the impulse to create is something closer to intrinsic, present in some people, absent in others, and unchanged by even the most powerful new technology? I am not talking about motivation, which comes and goes with circumstance. I mean something more permanent. A restlessness. A compulsion to make something exist that did not exist before. The kind of pull that keeps a person in front of a blank screen long after comfort has told them to stop.

If that compulsion is genuinely internal, then every tool ever created, regardless of how powerful, will leave the fundamental distribution unchanged. People who carry it will use the oracle to build. People who don’t will use it to answer questions about broccoli, faster and more efficiently than they could before.

---

In 1950, the physicist Enrico Fermi sat at lunch with colleagues at Los Alamos and asked a question that has haunted anyone with their eye to the stars ever since. The universe is staggeringly vast and nearly fourteen billion years old. There are hundreds of billions of galaxies, each containing hundreds of billions of stars, many of them orbited by planets in conditions hospitable to life. By sheer probability, intelligent civilizations should be common. So where is everyone?

This is Fermi’s Paradox, and the most reasonable proposed explanation is something called the Great Filter. The idea is that somewhere between dead matter and a civilization capable of reaching the stars, there is a barrier that almost no species gets past. In the optimistic version, the filter is behind us: the leap from single cells to complex life was so improbable that we have already cleared the hardest part. In the terrifying version, the filter is ahead. Something intrinsic to intelligent life, something about what we are, prevents us from becoming what we could become, whether by our own hand or by the cold indifference of a universe that does not care how far a species has come before an asteroid, a gamma-ray burst, or its own self-destruction wipes it from the record.

For decades, the technology industry has operated under its own optimistic interpretation. The Great Filter for human creativity was external. It was resources, access, friction. Remove those things and the latent genius of humanity would bloom.

AI removed them. The bloom did not come.

---

I should be transparent about my own position in this. I am one of the people for whom AI has been transformative. I started programming at thirteen. I studied computer science. I built companies and products across web, mobile, and VR. Then I spent years running a studio where my role shifted away from hands-on building, and development became a luxury I no longer had time for.

When AI tools became good enough to function as genuine collaborators, something came back. I built an entire website from scratch in five nights and a weekend, the kind of project that would have taken months with a team, completed solo and many times faster than I had ever worked before. I spent five months of spare time building a social platform from nothing. Ideas I had been carrying for years could finally become real, and the feeling of that return was electric.

But here is what I cannot look away from. I was building before AI. I was the kid who spent every available hour in the computer lab, making websites at seventeen, starting a company in college. AI did not create the impulse. It supercharged an impulse that was already there, one that has been present since I was five years old, lying on the floor in front of a glowing green monitor, pressing keys to see what would happen.

The people I know who are doing extraordinary things with AI were already doing things before AI. They carried the restlessness. They had whatever the unnamed thing is that makes someone sit down, face the blank screen, start, and most importantly, not stop until it is done.

The people who were not building before AI have the oracle now. Sadly, they are still not building and launching. The truth is that many of them are creating, tinkering, asking questions, generating outputs. But there is a vast distance between creating and shipping, between asking the oracle for answers and using those answers to put something real into the world. Almost no one seems to have an interest in crossing that distance.

---

This is where the story of democratization curdles into something harder to look at.

If the filter is internal, if it was never about access, then each new wave of technological empowerment does something other than level the playing field. It tilts the field further. The people who were already building receive a jet engine. Everyone else has access to the same jet engine, but they choose to keep it in the garage.

There is an obvious counter to this, and it deserves acknowledgement. Companies everywhere are leveraging AI to build faster, increase margins, and create enormous value. Entire industries are being restructured around the technology. Surely that represents broad adoption?

But look closer. In most of these cases, employees are being directed by their employers to use AI to produce more output at lower cost, increasing profit margins for a select few. That is capitalism doing what capitalism does, allocating powerful tools toward the extraction of value. It is a different question entirely from whether individuals, given the same tools and the freedom to use them however they wish, will choose to create something of their own. The company using AI to cut costs and the person using AI to build a dream are not the same story, and conflating them only reinforces the question: who is this technology actually democratizing for?

Economists describe what happens when growth diverges rather than converges as a K-shaped economy. Wages split. Opportunity splits. The top of the K accelerates upward while the bottom drifts down, and the distance between them only grows.

I believe we are watching a K-shaped divergence in human capability, and AI is the hinge point. What should unsettle us most is the rate. This divergence is not the slow, grinding separation of wages over decades. AI compounds. The technology improves exponentially. The ecosystem of research, investment, tooling, and capability is steepening, and the people who already knew how to build are riding that curve toward the clouds while the rest of the world watches from the basement.

The unlock came. It just did not come for humanity at large. It came for the people who were going to build with or without it. For everyone else, the oracle is a slightly faster search engine.

And the technology industry, whose founding mythology depends on access as the bottleneck, has no language for this. No framework. No honest accounting. If the filter is us, if the thing preventing most people from building is something that lives inside the individual rather than inside the system, then the central promise Silicon Valley has been making for forty years requires serious revision. “We are democratizing creation” becomes a much more complicated sentence when you have to ask: democratizing it for whom?

---

I do not have an answer. I am not sure anyone does yet.

It is possible the filter is more nuanced than a binary. Perhaps what we are calling drive is itself shaped by conditions, by childhood environments, by early exposure to making, by models of agency encountered at the right age. Perhaps the filter is cultural rather than innate, and can be taught, or at least cultivated. I hope so. The alternative is bleak in ways I do not think we have begun to reckon with.

But I find myself less certain than I used to be. The technology industry has spent four decades telling us that if we gave people the right tools, we would unlock something extraordinary in the human species. The tools have never been better. The unlock, for the vast majority, has not come. And that gap between what we promised and what actually happened is something we owe it to ourselves to sit with, uncomfortable as it is.

Clarke saw it seventy years ago. The Overlords gave humanity everything, and humanity discovered that everything was not enough. The gifts were sufficient. What was missing was the thing that makes a species reach, and that was never about the obstacles in the way.

We built the oracle. We gave it to everyone.

And what that revealed was a story about us.