Not the Same Two Percent

Ideas · Attention · Part three

Not the Same Two Percent

About two in a hundred American adults hold a doctorate. About two in a hundred people, in any era and any culture, give themselves wholly to one idea. The numbers match. The people mostly do not.

By John Rector · October 30, 2026

The coincidence we mistake for an identity

2% ≠ 2%

The titled and the faithful are roughly the same size. They overlap far less than we assume.

In the United States, roughly two adults in a hundred hold a doctoral degree; across developed countries the average is closer to one. The second two percent is counted by no one. It is my estimate of how many people let one idea take nearly all of their bearing, with or without a title, in every century I know anything about.

Any weekday · A red light on Fifth Street

She has driven the same bus route for eleven years. At every red light she is working on the same problem: why the transfer at Fifth Street strands people in the rain, and what would happen if the two routes traded four minutes. She has notebooks of timings. She has a theory. Nobody would call her a researcher, and she would laugh if you did. The idea has had her for a decade.

Every human being alive is in relationship with ideas right now. Not some humans. Not the educated ones. Every one, regardless of age, religion, culture, nation or creed. The child who cannot let go of fairness. The grandmother who has been perfecting one recipe for forty years. The welder, the nurse, the monk, the bus driver at the red light. And the things we usually treat as the ground under ideas, culture and creed and nation, are better understood as their residue: the accumulated record of marks that a great many people made while answering the same conditions. A nation is what a great many marks look like from far away.

So when we talk about the people who “have ideas,” we are looking at a tiny, oddly chosen sample and mistaking it for the species. I am, as far as I can tell, rare in how seriously I hold the other direction: people do not have ideas; ideas have people. If that is right, the question is not who has ideas. Everyone is had. The question is who stays faithful.

Ask an idea

Purely as a metaphor, imagine interviewing an idea. Ask it what share of humanity should be devoted to ideas the way a doctoral candidate is supposed to be. It would vote for one hundred percent. Every one of them would.

I say “purely as a metaphor” on purpose, because I have corrected the literal version in public. For two years I taught that ideas want actualization. They do not. An idea is a named condition, and conditions do not want, wait or vote. The pressure we feel is real, but it is ours: an open term in our own denominator that only a mark on the record can close.

Here is the honest version of the vote, and it turns out to be stronger than the metaphor. In the Reality Equation, every human expectation carries an imaginary component, ideation, alongside prediction. That is what it means to be a host rather than merely an actualizer: you hold a bearing toward what is not yet actual. And a condition that no one ever marks leaves no trace on the record at all. So the hundred percent is already true in the sense that matters. Everyone is a host. Everyone is in relationship. What is nowhere near a hundred percent is fidelity.

Why devotion stays near two percent

Picture each idea you are in relationship with as an arrow pointing in its own direction. Family points one way, faith another, work another, the country another, the recipe another. Add them tip to tail and most of them cancel. What remains, as I wrote in Dominated by a Big Idea, is your bias as an actualizer: the leftover tilt after everything else has had its say.

Figure 01

Two bearings

Argument, not data. Pink arrows are ideas a host is in relationship with; gold is what remains after they are added. Most lives are the left circle, and that is not a failure. The faithful few are the right circle, and they pay for it with everything the other arrows would have claimed.

For most people the remainder is modest, and that is not a defect. It is a wide life: many loves, held at once, none allowed to swallow the rest. For a rare few, nearly every arrow points the same way. One remainder grows so long that it organizes everything: the hours, the friendships, the money, the sleep. That is what wholehearted devotion looks like from the inside of the geometry.

It stays rare because it costs nearly everything attention could otherwise buy, and most people sensibly decline that bill. I see the same small fraction pay it in every field and every era I know anything about. Call it two percent. It does not appear to rise when universities grow, and I do not expect it to rise now.

The credential selects for the wrong trait

Now put the two populations side by side. A doctorate takes years of milestones, committees, exams, revisions and waiting. Whatever else it proves, it proves a high tolerance for process. Real love of an idea tends to run the other way. It has a low tolerance for anything that is not the idea. A filter that rewards tolerance will reliably miss a great many of the people with the least of it.

That is why so many of the people we hold up as the faithful reached the work by another road, or got the title after the work was already done.

Michael Faraday, the bookbinder’s apprentice
Little formal schooling. He read the science books he was binding, took meticulous notes at Humphry Davy’s lectures, sent them to Davy, and was taken on at the Royal Institution in 1813. Electromagnetism had found its man before any institution vouched for him.
The Wright brothers, without diplomas
Neither Wilbur nor Orville finished high school. They ran a bicycle shop in Dayton, built their own wind tunnel there, and used its numbers to fly in 1903, ahead of professionally credentialed rivals.
Freeman Dyson, by choice
One of the great physicists of the century never completed a doctorate, and was proud of it. He called the PhD system an abomination. He collected some two dozen honorary degrees anyway.
Jane Goodall, title afterward
She went to Gombe in 1960 with no degree at all. Cambridge later let her take a doctorate without a bachelor’s, one of only a handful ever allowed to. The title followed the work. It did not license it.

And Andrew Wiles, who did hold the title, shows the same thing from the inside. With the PhD in hand and a professorship at Princeton, he went up to his attic and hid for seven years from the very community that had credentialed him, because, as he put it, years of undivided concentration were impossible with too many spectators. The title did not make him faithful. It just stopped mattering.

The complicated marriages

None of this means the titled are frauds. Almost all of them began as true lovers. Most of them still are. But many have what can only be called a complicated relationship with the idea, and from the outside we can sometimes see them cheating on it.

In this framework the other woman is not a mystery. She is another idea. Tenure is an idea. The grant cycle is an idea. The citation count, the prestige of the journal, the salary, the title itself: these are conditions too, and they have people the same way the first idea did. I wrote about the business version of this in Financialization Has People. The academic version looks like the safe topic chosen over the true one, the finding sliced into four papers, the question shrunk to fit what reviewers will fund. Each choice is defensible. Together they produce exactly what I warned founders about in Do Not Negotiate the Idea: a blend of competing purposes whose relationship to the original idea has become faint.

I hold this with some compassion. The institution makes the second relationship a condition of keeping the first one funded. Many people keep both honorably for a whole career. We only notice when the second one wins. But it is worth naming, because it is the clearest way to see that the title and the love are two different things that sometimes share a house.

In 2026, the lover does not need the title

For most of history the credential was also the price of admission: to the library, the colleague who could check your math, the senior person who would tell you that you were wrong, the press that would print you. That price was real, and it kept many faithful people outside.

Most of it has just collapsed. A patient, tireless interlocutor is available at two in the morning. The literature, the math, the code, the statistics, the drafting: all of it now arrives on request. What remains genuinely exclusive is the expensive physical layer, the wet lab, the telescope, the collider. Outside that layer, a person can be wholly in love with an idea and work at the frontier of it with no title at all. Most of the faithful care little about the title anyway. Now they do not have to pay for it.

Spring 2030 · Fifth Street

The bus driver finally put her notebooks into her AI and asked it to show her what she had found. It built the model with her over three evenings, checked it against a year of the city’s public schedule data, and helped her write it up. The transit authority moved the two routes four minutes apart that fall. She has still never set foot in a seminar. The record did not ask for her credentials. It only takes marks.

Five predictions

The devoted fraction stays near two percent.
AI lowers the cost of acting on an idea. It does not lower the cost of giving an idea everything. Wholehearted devotion was never limited by tools. It was limited by what people were willing to give up.
The overlap between the two two-percents keeps shrinking.
Through the 2030s, a growing share of work that experts recognize as landmark will come from people without doctorates, outside the expensive physical sciences where the instruments still live behind institutional walls.
Titles follow the work.
More institutions will do what Cambridge did for Goodall: grant credentials, fellowships and degrees by published work to people whose marks are already on the record, rather than requiring the title before the work is allowed to count.
The hundred percent becomes visible as participation.
Not devotion, but marks. When making a real mark costs an afternoon instead of a career, far more people close at least one open term in their lives: the recipe written down, the route fixed, the small proof finished. That is the honest answer to whether the two percent should be higher. Fidelity, no. Participation, enormously.
Fidelity becomes legible.
When output is nearly free, volume stops meaning anything. Paper counts and publication lists inflate into noise. The only signal left is a long run of marks all pointing the same way. The complicated marriages will get easier to see, and so will the faithful ones.

Where I am in this

I am in the titled two percent. I have a PhD, and I was never in love with the dissertation or any other part of the process, only with the idea. I have spent years on this site arguing that ideas have people, and I have had to correct myself in public along the way, which I think is what fidelity to an idea actually looks like when the idea is bigger than your last sentence about it. The title has never been the point. It is a record that I tolerated a process once. The love is a different record, and it is still being written.

The ledger

Already true: Every human expectation carries ideation, so every human is a host. Roughly two in a hundred American adults hold a doctorate, and about one in a hundred across developed countries. Many of the most celebrated faithful, Faraday, the Wrights, Dyson, Goodall, did their defining work without the title or before it.
What has to happen: Institutions have to recognize work that arrives without a credential attached, and the tools that collapsed the price of admission have to stay cheap and open. People also have to learn to tell an idea from a wound. Persistence is not proof, and not everything that will not leave you alone deserves your life.
Where I am probably wrong: My two percent is an observation, not a measurement. Nobody has ever counted devotion, and I may be seeing a round number partly because the doctorate number sits right next to it. The credential also does real work I have undersold: training, community, and above all correction. Wiles needed Richard Taylor to help repair the gap in his proof. A “low tolerance for nonsense” can be a flattering name for a low tolerance for being told you are wrong. If the faithful lose the people who check them, love may produce more beautiful errors, not more Faradays.

Go back to the red light on Fifth Street. Eleven years, one problem, notebooks full of timings. She is not in anyone’s count of people with ideas. She is in the other count, the one nobody keeps.

The record does not ask for credentials. It only takes marks.

Background

Not the Same Two Percent