The internet was down at Los Alamos National Laboratory. Two thousand forested feet above Albuquerque, the relatively remote spot was once a hidey-hole for the Manhattan Project. Today, it is tasked with keeping the many resulting nuclear warheads stable. But it has also, and not unrelatedly, become a hub of quantum-computing research—and a bad place for the internet to fritz out.
It was June 17, and quantum thermodynamicist Zoe Holmes was in town during the outage, participating in LANL’s Quantum Computing Summer School, a 10-week boot-camp for students. “Everybody was really frustrated,” she says. But then, instead of Googling and Githubbing and StackExchanging, she and her peers went analog.
“We spent all day by the whiteboard doing calculations,” she said. “I’ve not done that before.” It was… nice.
Since getting plugged in once again, Holmes and 14 other students have been humming away on projects, going to lectures, and generally steeping themselves in the quantum-computing universe. It’s a subject that, for the most part, doesn’t exist in universities. Only a few schools offer courses and even fewer offer specializations.
Since the schools of the present weren’t preparing the quantum workforce of the future, LANL started the summer school last year to do so for them. In 2018, LANL had 90 applicants for 10 spots; this year, the program had 230 applicants for 15 spots.
“It’s sort of a buyers’ market for us,” said Patrick Coles, a quantum physicist and the summer school’s organizer. And indeed, the lab does kind of buy its students, paying them a stipend to spend 10 weeks in the high altitudes of New Mexico, learning about the strange machines of the future.
But why would the Department of Energy care about bringing students like Holmes here to collaborate in the first place? Why has the agency invested hundreds of millions of dollars in quantum computing, just in the past year?
For one, classical computers—while great at scrolling Twitter or even modeling how matter spread out after the Big Bang—suck at simulating quantum systems, ones subject to the physical laws of the super small. But quantum computers can simulate complicated systems that involve lots of individual objects, mimicking atoms or optimizing shipping logistics. If you want to model a chemical reaction, or understand how material behaves in a superconductor, you call D-Wave, not Dell.
“If you have a classical supercomputer that can simulate a molecule with at most 10 atoms, then you would have to double the size of that supercomputer just to simulate 11 atoms,” said Coles. You kind of hit a wall because your computer has to grow exponentially just to add atoms.
If quantum computers can successfully simulate atoms and particles, scientists can understand how the nuclear material inside existing bombs is behaving, what will happen as it sits inert, and what would happen if someone pressed the big red button
Quantum computers, though, only have to grow linearly when you add atoms, giving them more flex for their size. Chemical research of that sort could accelerate pharmaceutical development, whipping up digital chemical reactions you couldn’t digitally replicate before. It could reveal how best to route cargo carriers, or any other system with a ton of moving parts.
And yeah, LANL delves into those topics. But the hot core of LANL’s mission is nuclear security. And if quantum computers can successfully simulate atoms and particles, scientists can understand how the nuclear material inside existing bombs is behaving, what will happen as it sits inert, and what would happen if someone pressed the big red button. Since we’re not supposed to just blow up bombs in the desert anymore to understand them, scientists could simulate nuclear blasts (and the sitting-still bombs) on quantum computers.
While we have a good sense of what they can do, if you ask a quantum computer person how the computers work, you likely won’t get a satisfying answer. That’s partly because it’s complicated. But it’s also because quantum mechanics has no practical bearing on our interactions with the world. Our brains did not evolve while observing Schrodinger’s Cat, and we did not spend our formative years watching entangled electrons. As physicist Richard Feynman famously put it, “I think I can safely say that nobody understands quantum mechanics.”
Nevertheless, here’s an attempt to explain: Classical computers encode information in bits, each either a 0 or a 1. Quantum computers use qubits, entangled quantum systems where each bit represents both 0 and 1 at the same time. That superposition makes certain calculations—like simulating what atoms are up to—faster with less computing power.
While quantum computers are in an embryonic stage of development, scientists have already written algorithms that do certain simple things, like factor numbers. You can look those up and then string them together and plug them into an existing system, like the ones IBM, Rigetti, and D-Wave run in the cloud. These real-world instruments aren’t ready for primetime: They’re more proofs of concept than productive problem-solvers. The biggest universal quantum processor—the traditional sort—comes courtesy of Google and has just 72 qubits.
Right now, for almost every task, a classical computer will beat out its experimental quantum counterpart, so the systems mostly exist so that researchers can figure out how these computers could work in the future. The summer school, says Holmes, focuses not on the distant blue-sky future but what might come for the field in the next 5-10 years.
Also, the idea that other countries will get good at quantum computing spooks the US government: Quantum computers could theoretically break certain kinds of encryption, the security of which relies on the fact that a classical computer could take centuries to brute-force decode it. “This kind of encryption is based on the difficulty of factoring large numbers into primes,” Elizabeth Crosson, a professor in the Center for Quantum Information and Control at the University of New Mexico and a lecturer for the summer school, said. “This is something that a quantum computer can do efficiently but a classical computer can’t.”
A quantum computer can also itself encrypt information, entangling subatomic particles to do so rather than relying on eighth-grade math. That’s why the US wants to build its own functional computers first: because it’s the best, obviously, and because it could decode others’ secrets and more safely store its own. The National Institute of Standards and Technology is currently pouring resources into “post-quantum cryptography,” classical encryption that can work even if a computer has 15 times the qubits of Google’s current superlative system.
For any of that, though, you need human power, not just computing power. And if schools aren’t spitting out such humans, you send them to summer school—a laid-back way, complete with barbecues, to try to save information security.
This is part of what drew Holmes, who started out as a philosophy major and added physics as a double major, to the topic and the LANL program. “I liked it when physics got philosophical,” she said. “And when philosophy was asking questions about the real world.” Quantum mechanics seemed like a place where the two entangled themselves most tightly. Today, she’s doing a PhD focusing on quantum thermodynamics at Imperial College London. She came to the summer school to broaden her subatomic knowledge, and for the collaboration that isn’t always part of theoretical physics.
Even the bigwigs chill out at LANL’s program: Last summer, Coles invited MIT mathematician Peter Shor—the guy who developed the quantum algorithm that does the kind of factoring that could unscramble our secret scrambled bits. “He’s arguably the biggest name in our field,” said Coles. Coles has a picture of him casually eating sourdough bread out of a paper bag. Students and faculty sip beer in the background, taking a break from the future to enjoy the present. Or, perhaps, like a qubit, they’re superimposing those states of being.