Dr. Rolf Ekelund in his office at the California Institute of Technology in Pasadena on Wednesday, with the ledger in which he has recorded every problem in physics he considers to have been solvable by one person at a desk. The final entry is dated 1930. Credit: Marguerite Sallander/The New York Time5
Dr. Rolf Ekelund in his office at the California Institute of Technology in Pasadena on Wednesday, with the ledger in which he has recorded every problem in physics he considers to have been solvable by one person at a desk. The final entry is dated 1930. Credit: Marguerite Sallander/The New York Time5

PASADENA, Calif. — At 6:40 on Wednesday morning, Dr. Rolf Ekelund sat down at the desk in his office at the California Institute of Technology, sharpened a pencil with a pocketknife, and solved the hydrogen atom.

It took him fifty minutes. He wrote the energy levels at the bottom of the page, checked them against nothing, signed and dated the sheet, and placed it in a drawer with the 211 others. He has solved the hydrogen atom, by his count, 212 times since 2019, a practice he describes as “keeping the hands in.” Then he did what he has done every morning for seven years, which is to open a green clothbound ledger to its last written page and look for something else to solve. There was nothing there. There has been nothing there since 1930.

Dr. Ekelund, 61, a professor of theoretical physics, is what researchers are now calling an artisanal physicist, and he may be the purest example of the type. He does not use a computer. He does not use a collaborator. He will not use a result he has not obtained himself, and he will not accept a problem that cannot be solved by one person, at a desk, with a pencil, in what he calls “a human amount of time.” These constraints, which he adopted in 2019, have had one consequence that he did not anticipate and now cannot escape. Every problem that satisfies them has already been solved.

“I was born sixty years too late,” Dr. Ekelund said, in an interview in the office, from which the computer was removed at his request in 2019 and where the Ethernet jack is covered with a square of masking tape. “In 1925 a man could sit down at a desk after breakfast and know something about the universe by lunch that nobody had known before. The hydrogen atom. The harmonic oscillator. The particle in a box. These were lying on the ground. By 1930 the ground was clear. Everything since has needed a machine, a committee or a hole in the ground.”

The ledger is his evidence. Begun in 2020 and compiled by hand from the primary literature, which he reads in print, it lists every result in the history of physics that Dr. Ekelund judges to have been obtained by a single person working alone with pencil and paper, together with the date and, where he could establish it, the time it took. Kepler’s laws occupy the first page and, by his reckoning, several years. The Bohr atom took a few months in 1913. The wave equation took a holiday in the Alps over Christmas of 1925. The Dirac equation, in 1928, he records as having taken “an afternoon, by looking at it.” The entries thicken through the 1920s, a decade he refers to as “the harvest,” and then they stop. The final entry is dated 1930 and records a calculation performed by a nineteen-year-old on a steamship between Bombay and Venice, establishing the maximum mass of a white dwarf star. “Between meals,” Dr. Ekelund said. “That was the last one. After that you needed a cyclotron.”

Working physicists have a different name for the problems in the ledger. They are the exactly solvable problems: the handful of physical systems whose equations can be written down and then, in the manner of a recipe, followed through to the end. There are not many of them. They are taught to undergraduates, and they are taught to undergraduates because, like the mother sauces of a classical kitchen, every harder dish is built on top of them. Dr. Ekelund does not dispute any of this. He regards it as the problem.

He did not begin his career this way. For two decades he was a computational physicist who studied the behavior of electrons in disordered materials and ran simulations on some of the largest machines then available. The break came in 2018, when a numerical result that had occupied his group for three years was reproduced in an afternoon by a graduate student in Lausanne, using different software, and Dr. Ekelund realized, he said, “that neither of us had understood it. The machine had understood it. We had been in the room.” He published the result, withdrew from the collaboration, had the computer removed and bought a box of pencils. He has not published anything since, because he has not found anything since, because there is nothing at his size left to find.

The constraint has fallen hardest on Wendy Oyelaran, 27, a fourth-year graduate student in Dr. Ekelund’s group and, since 2023, its only member. She does not have a thesis problem. “He says he’s looking for one my size,” she said. “He’s been looking for four years. He’s very thorough. He’s gone through everything up to 1930.” Ms. Oyelaran has, in the meantime, been running numerical simulations of a spin glass on a laptop at night, in her car, in a parking structure on Wilson Avenue, with the dome light off. “I’m running Monte Carlo in a Honda,” she said. “It’s not what I pictured. He doesn’t know. He would know if he had email.”

The constraints Dr. Ekelund observes are what practitioners in other fields call artisanal thinking, a movement, if a practice without an organization can be called one, that emerged in the East Bay and reached public attention this week when Dr. Everett Lindqvist, a surgeon in Berkeley, dismissed a robot midway through a gallbladder operation and completed it by hand. Dr. Ekelund receives the movement’s newsletter, The Unassisted, by mail, in an envelope. He read the account of the operation with attention and without sympathy. “A gallbladder is an easy problem,” he said. “He had one. I would like to know where he got it. Surgeons have easy problems the way people in 1925 had easy problems. They have them every day. They are taking them out of people. I have been looking for one for seven years.”

Not everyone at Caltech shares the diagnosis. Down the corridor, Dr. Anders Ruhl, a colleague who has spent eleven years deriving Western philosophy from first principles without reading any of it, said that he encountered easy problems constantly. “Rolf has run out of problems because he has stayed in physics,” Dr. Ruhl said. “That is a choice. Philosophy is almost entirely easy problems. I close one a week. The consciousness thing is giving me some trouble, but that is one problem, and when it is done I will go and look at economics.” He added that he had offered Dr. Ekelund the induction thing in 2021 and that Dr. Ekelund had declined it on the grounds that it was not physics. “It is physics,” Dr. Ruhl said. “It is whether tomorrow will happen. I do not see what else it could be.”

Dr. Helene Marquardt, who holds the Enrico Fermi Chair in Particle Theory at Columbia University, said the ledger was accurate and the conclusion drawn from it was, in a sense, the point. “Well, the short answer is that he’s right,” Dr. Marquardt said. “The long answer is that he has defined an easy problem as one a person can solve alone at a desk, and the reason those problems are gone is that they were solved, and the reason they look easy is that they were solved. A problem is easy after someone has solved it. Before that, it is a problem. In the standard formulation we do not call them the easy problems. We call them the first ones.” She disputed the date. “That’s an interesting claim, 1930,” she said, and named a theorem from 1964 and a result from 1974, each obtained with pencil and paper by a single person. Dr. Ekelund has rulings on both. The first, he said, is mathematics. The second relied on a framework its author did not obtain himself. “He has an appeals process,” Dr. Marquardt said. “He is also the appeals process.”

Dr. Franklin Krause, the director of the Institute for Occupational Cognitive Health in Baltimore, which has tracked unassisted reasoning in operating rooms since 2022, said that physicists working alone had been the institute’s control group from the beginning. “The surgeons came to this in 2024,” Dr. Krause said. “The physicists have been doing it since 1687. We did not need to recruit them.” The institute’s 2026 survey of theoretical physicists found that 14 percent had, in the past year, rederived a textbook result for reasons they described as “not teaching,” and that 3 percent had done so more than fifty times. Dr. Ekelund is in the 3 percent, Dr. Krause said, “and in some respects is the 3 percent.” The institute does not recommend the practice. It recognizes it. It has been warning about it, he said, since before it was founded, “which is the only kind of warning that would have reached him.”

What the exhaustion of the easy problems could mean for ordinary Americans is difficult to say. Most of them have not solved the hydrogen atom once, a point Dr. Ekelund made himself and, it should be said, without envy. “They still have it,” he said. “It is right there.”

He has not stopped looking. He has written to the American Physical Society to ask whether any easy problems remain and received a reply that referred him to a website. He has considered the three-body problem, which cannot be solved, “which is a different thing from hard,” and turbulence, which he described as “a problem for people who own a computer.” And on the last page of the ledger, beneath the entry from the steamship, he has recently written a new line, in pencil, undated. It is a problem he has set for himself: to determine, from first principles and by hand, whether any easy problem remains. He expects it to take several years. “It is not easy,” he said. “That is how I know it is mine.”