Humans can intuit quantum physics.

One evening this January, audience members packed into a lecture hall in MIT’s physics building. Undergraduates, members of the public, faculty members, and other scholars came to watch a film premiere and a panel discussion. NOVA had produced the film, “Einstein’s Quantum Riddle,” which stars entanglement. Entanglement is a relationship between quantum systems such as electrons. Measuring two entangled electrons yields two outcomes, analogous to the numbers that face upward after you roll two dice. The quantum measurements’ outcomes can exhibit correlations stronger than any measurements of any classical, or nonquantum, systems can. Which die faces point upward can share only so much correlation, even if the dice hit each other.

einstein's q. riddle

Dice feature in the film’s explanations of entanglement. So does a variation on the shell game, in which one hides a ball under one of three cups, shuffles the cups, and challenges viewers to guess which cup is hiding the ball. The film derives its drama from the Cosmic Bell test. Bell tests are experiments crafted to show that classical physics can’t describe entanglement. Scientists recently enhanced Bell tests using light from quasars—ancient, bright, faraway galaxies. Mix astrophysics with quantum physics, and an edgy, pulsing soundtrack follows.

The Cosmic Bell test grew from a proposal by physicists at MIT and the University of Chicago. The coauthors include David Kaiser, a historian of science and a physicist on MIT’s faculty. Dave co-organized the premiere and the panel discussion that followed. The panel featured Dave; Paola Cappellaro, an MIT quantum experimentalist; Alan Guth, an MIT cosmologist who contributed to the Bell test; Calvin Leung, an MIT PhD student who contributed; Chris Schmidt, the film’s producer; and me. Brindha Muniappan, the Director of Education and Public Programs at the MIT Museum, moderated the discussion.


think that the other panelists were laughing with me.

Brindha asked what challenges I face when explaining quantum physics, such as on this blog. Quantum theory wears the labels “weird,” “counterintuitive,” and “bizarre” in journalism, interviews, blogs, and films. But the thorn in my communicational side reflects quantum “weirdness” less than it reflects humanity’s self-limitation: Many people believe that we can’t grasp quantum physics. They shut down before asking me to explain.

Examples include a friend and Quantum Frontiers follower who asks, year after year, for books about quantum physics. I suggest literature—much by Dave Kaiser—he reads some, and we discuss his impressions. He’s learning, he harbors enough curiosity to have maintained this routine for years, and he has technical experience as a programmer. But he’s demurred, several times, along the lines of “But…I don’t know. I don’t think I’ll ever understand it. Humans can’t understand quantum physics, can we? It’s too weird.” 

Quantum physics defies many expectations sourced from classical physics. Classical physics governs how basketballs arch, how paint dries, how sunlight slants through your window, and other everyday experiences. Yet we can gain intuition about quantum physics. If we couldn’t, how could we solve problems and accomplish research? Physicists often begin solving problems by trying to guess the answer from intuition. We reason our way toward a guess by stripping away complications, constructing toy models, and telling stories. We tell stories about particles hopping from site to site on lattices, particles trapped in wells, and arrows flipping upward and downward. These stories don’t capture all of quantum physics, but they capture the essentials. After grasping the essentials, we translate them into math, check how far our guesses lie from truth, and correct our understanding. Intuition about quantum physics forms the compass that guides problem solving.

Growing able to construct, use, and mathematize such stories requires work. You won’t come to understand quantum theory by watching NOVA films, though films can prime you for study. You can gain a facility with quantum theory through classes, problem sets, testing, research, seminars, and further processing. You might not have the time or inclination to. Even if you have, you might not come to understand why quantum theory describes our universe: Science can’t necessarily answer all “why” questions. But you can grasp what quantum theory implies about our universe.

People grasp physics arguably more exotic than quantum theory, without exciting the disbelief excited by a grasp of quantum theory. Consider the Voyager spacecraft launched in 1977. Voyager has survived solar winds and -452º F weather, imaged planets, and entered interstellar space. Classical physics—the physics of how basketballs arch—describes much of Voyager’s experience. But even if you’ve shot baskets, how much intuition do you have about interstellar space? I know physicists who claim to have more intuition about quantum physics than about much classical. When astrophysicists discuss Voyager and interstellar space, moreover, listeners don’t fret that comprehension lies beyond them. No one need fret when quantum physicists discuss the electrons in us.

Fretting might not occur to future generations: Outreach teams are introducing kids to quantum physics through games and videos. Caltech’s Institute for Quantum Information and Matter has partnered with Google to produce QCraft, a quantum variation on Minecraft, and with the University of Southern California on quantum chess. In 2017, the American Physical Society’s largest annual conference featured a session called “Gamification and other Novel Approaches in Quantum Physics Outreach.” Such outreach exposes kids to quantum terminology and concepts early. Quantum theory becomes a playground to explore, rather than a source of intimidation. Players will grow up primed to think about quantum-mechanics courses not “Will my grade-point average survive this semester?” but “Ah, so this is the math under the hood of entanglement.”

qcraft 2

Sociology restricts people to thinking quantum physics weird. But quantum theory defies classical expectations less than it could. Measurement outcomes could share correlations stronger than the correlations sourced by entanglement. How strong could the correlations grow? How else could physics depart farther from classical physics than quantum physics does? Imagine the worlds governed by all possible types of physics, called “generalized probabilistic theories” (GPTs). GPTs form a landscape in which quantum theory constitutes an island, on which classical physics constitutes a hill. Compared with the landscape’s outskirts, our quantum world looks tame.

GPTs fall under the research category of quantum foundations. Quantum foundations concerns why the math that describes quantum systems describes quantum systems, reformulations of quantum theory, how quantum theory differs from classical mechanics, how quantum theory could deviate but doesn’t, and what happens during measurements of quantum systems. Though questions about quantum foundations remain, they don’t block us from intuiting about quantum theory. A stable owner can sense when a horse has colic despite lacking a veterinary degree.

Moreover, quantum-foundations research has advanced over the past few decades. Collaborations and tools have helped: Theorists have been partnering with experimentalists, such as on the Cosmic Bell test and on studies of measurement. Information theory has engendered mathematical tools for quantifying entanglement and other quantum phenomena. Information theory has also firmed up an approach called “operationalism.” Operationalists emphasize preparation procedures, evolutions, and measurements. Focusing on actions and data concretizes arguments and facilitates comparisons with experiments. As quantum-foundations research has advanced, so have quantum information theory, quantum experiments, quantum technologies, and interdisciplinary cross-pollination. Twentieth-century quantum physicists didn’t imagine the community, perspectives, and knowledge that we’ve accrued. So don’t adopt 20th-century pessimism about understanding quantum theory. Einstein grasped much, but today’s scientific community grasps more. Richard Feynman said, “I think I can safely say that nobody understands quantum mechanics.” Feynman helped spur the quantum-information revolution; he died before its adolescence. Besides, Feynman understood plenty about quantum theory. Intuition jumps off the pages of his lecture notes and speeches.


Landscape beyond quantum theory

I’ve swum in oceans and lakes, studied how the moon generates tides, and canoed. But piloting a steamboat along the Mississippi would baffle me. I could learn, given time, instruction, and practice; so can you learn quantum theory. Don’t let “weirdness,” “bizarreness,” or “counterintuitiveness” intimidate you. Humans can intuit quantum physics.

My QIP 2019 After-Dinner Speech

Scientists who work on theoretical aspects of quantum computation and information look forward each year to the Conference on Quantum Information Processing (QIP), an annual event since 1998. This year’s meeting, QIP 2019, was hosted this past week by the University of Colorado at Boulder. I attended and had a great time, as I always do.

But this year, in addition to catching up with old friends and talking with colleagues about the latest research advances, I also accepted a humbling assignment: I was the after-dinner speaker at the conference banquet. Here is (approximately) what I said.

QIP 2019 After-Dinner Speech
16 January 2019

Thanks, it’s a great honor to be here, and especially to be introduced by Graeme Smith, my former student. I’m very proud of your success, Graeme. Back in the day, who would have believed it?

And I’m especially glad to join you for these holiday festivities. You do know this is a holiday, don’t you? Yes, as we do every January, we are once again celebrating Gottesman’s birthday! Happy Birthday, Daniel!

Look, I’m kidding of course. Yes, it really is Daniel’s birthday — and I’m sure he appreciates 500 people celebrating in his honor — but I know you’re really here for QIP. We’ve been holding this annual celebration of Quantum Information Processing since 1998 — this is the 22nd QIP. If you are interested in the history of this conference, it’s very helpful that the QIP website includes links to the sites for all previous QIPs. I hope that continues; it conveys a sense of history. For each of those past meetings, you can see what people were talking about, who was there, what they looked like in the conference photo, etc.

Some of you were there the very first time – I was not. But among the attendees at the first QIP, in Arhus in 1998, where a number of brilliant up-and-coming young scientists who have since then become luminaries of our field. Including: Dorit Aharonov, Wim van Dam, Peter Hoyer (who was an organizer), Michele Mosca, John Smolin, Barbara Terhal, and John Watrous. Also somewhat more senior people were there, like Harry Buhrman and Richard Cleve. And pioneers so eminent that we refer to them by their first names alone:  Umesh … Gilles … Charlie. It’s nice to know those people are still around, but it validates the health of our field that so many new faces are here, that so many young people are still drawn to QIP, 21 years after it all began. Over 300 students and postdocs are here this year, among nearly 500 attendees.

QIP has changed since the early days. It was smaller and more informal then; the culture was more like a theoretical physics conference, where the organizing committee brainstorms and conjures up a list of invited speakers. The system changed in 2006, when for the first time there were submissions and a program committee. That more formal system opened up opportunities to speak to a broader community, and the quality of the accepted talks has stayed very high — only 18% of 349 submissions were accepted this year.

In fact it has become a badge of honor to speak here — people put it on their CVs: “I gave a QIP contributed talk, or plenary talk, or invited talk.” But what do you think is the highest honor that QIP can bestow? Well, it’s obvious isn’t it? It’s the after-dinner speech! That’s the talk to rule them all. So Graeme told me, when he invited me to do this. And I checked, Gottesman put it on his website, and everyone knows Daniel is a very serious guy. So it must be important. Look, we’re having a banquet in honor of his birthday, and he can hardly crack a smile!

I hear the snickers. I know what you’re thinking. “John, wake up. Don’t you see what Graeme was trying to tell you: You’re too washed up to get a talk accepted to QIP! This is the only way to get you on the program now!” But no, you’re wrong. Graeme told me this is a great honor. And I trust Graeme. He’s an honest man. What? Why are you laughing? It’s true.

I asked Graeme, what should I talk about? He said, “Well, you might try to be funny.” I said, “What do you mean funny? You mean funny Ha Ha? Or do you mean funny the way cheese smells when it’s been in the fridge for too long?” He said, “No I mean really, really funny. You know, like Scott.”

So there it was, the gauntlet had been thrown. Some of you are too young to remember this, but the most notorious QIP after-dinner speech of them all was Scott Aaronson’s in Paris in 2006. Were you there? He used props, and he skewered his more senior colleagues with razor sharp impressions. And remember, this was 2006, so everybody was Scott’s more senior colleague. He was 12 at the time, if memory serves.

He killed. Even I appreciated some of the jokes; for example, as a physicist I could understand this one: Scott said, “I don’t care about the fine structure constant, it’s just a constant.” Ba ding!  So Scott set the standard back then, and though many have aspired to clear the bar since then, few have come close.

But remember, this was Graeme I was talking to. And I guess many of you know that I’ve had a lot of students through the years, and I’m proud of all of them. But my memory isn’t what it once was; I need to use mnemonic tricks to keep track of them now. So I have a rating system;  I rate them according to how funny they are. And Graeme is practically off the chart, that’s how funny he is. But his is what I call stealth humor. You can’t always tell that he’s being funny, but you assume it.

So I said, “Graeme, What’s the secret? Teach me how to be funny.” I meant it sincerely, and he responded sympathetically. Graeme said, “Well, if you want to be funny, you have to believe you are funny. So when I want to be funny, I think of someone who is funny, and I pretend to be that person.” I said, “Aha, so you go out there and pretend to be Graeme Smith?” And Graeme said, “No, that wouldn’t work for me. I close my eyes and pretend I’m … John Smolin!” I said, “Graeme, you mean you want me to be indistinguishable from John Smolin to an audience of computationally bounded quantum adversaries?” He nodded. “But Graeme, I don’t know any plausible cryptographic assumptions under which that’s possible!”

Fortunately, I had another idea. “I write poems,” I said. “What if I recite a poem? This would set a great precedent. From now on, everyone would know: the QIP after-dinner speech will be a poetry slam!”

Graeme replied “Well, that sounds [long pause] really [pause] boring. But how about a limerick? People love limericks.” I objected, “Graeme, I don’t do limericks. I’m not good at limericks.” But he wouldn’t back down. “Try a limerick,” Graeme said. “People like limericks. They’re so [pause] short.”

But I don’t do limericks. You see:

I was invited to speak here by Graeme.
He knows me well, just as I am.
He was really quite nice
When he gave this advice:
Please don’t do a poetry slam.

Well, like I said, I don’t do limericks.

So now I’m starting to wonder: Why did they invite me to do this anyway? And I think I figured that out. See, Graeme asked me to speak just a few days ago. This must be what happened. Like any smoothly functioning organizing committee, they lined up an after-dinner speaker months in advance, as is the usual practice.

But then, just a few days before the conference began, they began to worry. “We better comb through the speaker’s Twitter feed. Maybe, years ago, our speaker said something offensive, something disqualifying.” And guess what? They found something, something really bad. It turned out that the designated after-dinner speaker had once made a deeply offensive remark about something called “quantum supremacy” … No, wait … that can’t be it.

Can’t you picture the panicky meeting of the organizers? QIP is about to start, and there’s no after-dinner speaker! So people started throwing out suggestions, starting with the usual suspects.

“How about Schroedinger’s Rat?”
“No, he’s booked.”
“Are you telling me Schroedinger’s Rat has another gig that same night?”
“No, no, I mean they booked him.  A high-profile journal filed a complaint and he’s in the slammer.”
“Well, how about RogueQIPConference?”
“No, same problem.”

“I’ve got it,” someone says: “How about the hottest quantum Twitter account out there? Yes, I’m talking about Quantum Computing Memes for QMA-Complete Teens!”

Are you all following that account? You should be. That’s where I go for all the latest fast-breaking quantum news. And that’s where you can get advice about what a quantumist should wear on Halloween. Your costume should combine Sexy with your greatest fear.  Right, I mean Sexy P = BQP.

Hey does that worry you? That maybe P = BQP? Does it keep you up at night? It’s possible, isn’t it? But it doesn’t worry me much. If it turns out that P = BQP, I’m just going to make up another word. How about NISP? Noisy Intermediate-Scale Polynomial.

I guess they weren’t able to smoke out whoever is behind Quantum Computing Memes for QMA-Complete Teens. So here I am.

Aside from Limericks, Graeme had another suggestion. He said, “You can reminisce. Tell us what QIP was like in the old days.” “The old days?” I said. “Yes, you know. You could be one of those stooped-over white-haired old men who tells interminable stories that nobody cares about.” I hesitated. “Yeah, I think I could do that.”

Okay, if that’s what you want, I’ll tell a story about my first QIP; that was QIP 2000, which was actually in Montreal in December 1999. It was back in the BPC era — Before Program Committee — and I was an invited speaker (I talked about decoding the toric code). Attending with me was Michael Nielsen, then a Caltech postdoc. Michael’s good friend Ike Chuang was also in the hotel, and they were in adjacent rooms. Both had brought laptops (not a given in 1999), and they wanted to share files. Well, hotels did not routinely offer Internet access back then, and certainly not wireless. But Ike had brought along a spool of Ethernet cable. So Ike and Mike both opened their windows, even though it was freezing cold. And Ike leaned out his window and made repeated attempts to toss the cable though Michael’s window before he finally succeeded, and they connected their computers.

I demanded to know, why the urgent need for a connection? And that was the day I found what most of the rest of the quantum world already knew: Mike and Ike were writing a book! By then they were in the final stages of writing, after some four years of effort (they sent the final draft of the book off to Cambridge University Press the following June).

So, QIP really has changed. The Mike and Ike book is out now. And it’s no longer necessary to open your window on a frigid Montreal evening to share a file with your collaborator.

Boy, it was cold that week in Montreal. [How cold was it?] Well, we went to lunch one day during the conference, and were walking single file down a narrow sidewalk toward the restaurant, when Harry Buhrman, who was right behind me, said: “John, there’s an icicle on your backpack!” You see, I hadn’t screwed the cap all the way shut on my water bottle, water was leaking out of the bottle, soaking through the backback, and immediately freezing on contact with the air; hence the icicle. And ever since then I’ve always been sure to screw my bottle cap shut tight. But over the years since then, lots of other things have spilled in my backpack just the same, and I’d love to tell you about that, but …

Well, my stories may be too lacking in drama to carry the evening ….  Look, I don’t care what Graeme says, I’m gonna recite some poems!

I can’t remember how this got started, but some years ago I started writing a poem whenever I needed to introduce a speaker at the Caltech physics colloquium. I don’t do this so much anymore. Partly because I realized that my poetry might reveal my disturbing innermost thoughts, which are best kept private.

Actually, one of my colleagues, after hearing one of my poems, suggested throwing the poem into a black hole. And when we tried it … boom …. it bounced right back, but in a highly scrambled form! And ever since then I’ve had that excuse. If someone says “That’s not such a great poem,” I can shoot back, “Yeah, but it was better before it got scrambled.”

But anyway, here’s one I wrote to honor Ben Schumacher, the pioneer of quantum information theory who named the qubit, and whose compression theorem you all know well.

He rocks.
I remember
He showed me how to fit
A qubit
In a small box.

I wonder how it feels
To be compressed.
And then to pass
A fidelity test.
Or does it feel
At all, and if it does
Would I squeal
Or be just as I was?

If not undone
I’d become as I’d begun
And write a memorandum
On being random.
Had it felt like a belt
Of rum?

And might it be predicted
That I’d become addicted,
Longing for my session
Of compression?

I’d crawl
To Ben again.
And call,
“Put down your pen!
Don’t stall!
Make me small!”


Yeah that’s the response I usually get when I recite this poem — embarrassed silence, followed by a few nervous titters.

So, as you can see, as in Ben Schumacher’s case, I use poetry to acknowledge our debt to the guiding intellects of our discipline. It doesn’t always work, though. I once tried to write a poem about someone I admire very much, Daniel Gottesman, and it started like this:

When the weather’s hottest, then
I call for Daniel Gottesman.
My apples are less spotted when
Daniel eats the rottenest ten …

It just wasn’t working, so I stopped there. Someday, I’ll go back and finish it. But it’s tough to rhyme “Gottesman.”

More apropos of QIP, some of you may recall that about 12 years ago, one of the hot topics was quantum speedups for formula evaluation, a subject ignited by a brilliant paper by Eddie Farhi, Jeffrey Goldstone, and Sam Gutmann. They showed there’s a polynomial speedup if we use a quantum computer to, say, determine whether a two-player game has a winning strategy. That breakthrough inspired me to write an homage to Eddie, which went:

We’re very sorry, Eddie Farhi
Your algorithm’s quantum.
Can’t run it on those mean machines
Until we’ve actually got ‘em.

You’re not alone, so go on home,
Tell Jeffrey and tell Sam:
Come up with something classical
Or else it’s just a scam.

Unless … you think it’s on the brink
A quantum-cal device.
That solves a game and brings you fame.
Damn! That would be nice!

Now, one thing that Graeme explained to me is that the white-haired-old-man talk has a mandatory feature: It must go on too long. Maybe I have met that criterion by now. Except …

There’s one thing Graeme neglected to say. He never told me that I must not sing at QIP.

You see, there’s a problem: Tragically, though I like to sing, I don’t sing very well at all. And unfortunately, I am totally unaware of this fact. So I sometimes I sing in public, despite strongly worded advice not to do so.

When I was about to leave home on my way to QIP, my wife Roberta asked me, “When are you going to prepare your after-dinner talk?” I said, “Well, I guess I’ll work on it on the plane.” She said, “LA to Denver, that’s not a long enough flight.” I said, “I know!”

What I didn’t say, is that I was thinking of singing a song. If I had, Roberta would have tried to stop me from boarding the plane.

So I guess it’s up to you, what do you think? Should we stop here while I’m (sort of) ahead, or should we take the plunge. Song or no song? How many say song?

All right, that’s good enough for me! This is a song that I usually perform in front of a full orchestra, and I hoped the Denver Symphony Orchestra would be here to back me up. But it turns out they don’t exist anymore. So I’ll just have to do my best.

If you are a fan of Rodgers and Hammerstein, you’ll recognize the tune as a butchered version of Some Enchanted Evening, But the lyrics have changed. This song is called One Entangled Evening.

One entangled evening
We will see a qubit
And another qubit
Across a crowded lab.

And somehow we’ll know
We’ll know even then
This qubit’s entangled
Aligned with its friend.

One entangled evening
We’ll cool down a circuit
See if we can work it
At twenty milli-K.

A circuit that cold
Is worth more than gold
For qubits within it.
Will do as they’re told.

Quantum’s inviting, just as Feynman knew.
The future’s exciting, if we see it through

One entangled evening
Anyons will be braiding
And thereby evading
The noise that haunts the lab.

Then our quantum goods
Will work as they should
Solving the problems
No old gadget could!

Once we have dreamt it, we can make it so.
Once we have dreamt it, we can make it so!

The song lyrics are meant to be uplifting, and I admit they’re corny. No one can promise you that, in the words of another song, “the dreams that you dare to dream really do come true.” That’s not always the case.

At this time in the field of quantum information processing, there are very big dreams, and many of us worry about unrealistic expectations concerning the time scale for quantum computing to have a transformative impact on society. Progress will be incremental. New technology does not change the world all at once; it’s a gradual process.

But I do feel that from the perspective of the broad sweep of history, we (the QIP community and the broader quantum community) are very privileged to be working in this field at a pivotal time in the history of science and technology on earth. We should deeply cherish that good fortune, and the opportunities it affords. I’m confident that great discoveries lie ahead for us.

It’s been a great privilege for me to be a part of a thriving quantum community for more than 20 years. By now, QIP has become one of our venerable traditions, and I hope it continues to flourish for many years ahead. Now it’s up to all of you to make our quantum dreams come true. We are on a great intellectual adventure. Let’s savor it and enjoy it to the hilt!

Thanks for putting up with me tonight.

[And here’s proof that I really did sing.]

Chasing Ed Jaynes’s ghost

You can’t escape him, working where information theory meets statistical mechanics.

Information theory concerns how efficiently we can encode information, compute, evade eavesdroppers, and communicate. Statistical mechanics is the physics of  many particles. We can’t track every particle in a material, such as a sheet of glass. Instead, we reason about how the conglomerate likely behaves. Since we can’t know how all the particles behave, uncertainty blunts our predictions. Uncertainty underlies also information theory: You can think that your brother wished you a happy birthday on the phone. But noise corroded the signal; he might have wished you a madcap Earth Day. 

Edwin Thompson Jaynes united the fields, in two 1957 papers entitled “Information theory and statistical mechanics.” I’ve cited the papers in at least two of mine. Those 1957 papers, and Jaynes’s philosophy, permeate pockets of quantum information theory, statistical mechanics, and biophysics. Say you know a little about some system, Jaynes wrote, like a gas’s average energy. Say you want to describe the gas’s state mathematically. Which state can you most reasonably ascribe to the gas? The state that, upon satisfying the average-energy constraint, reflects our ignorance of the rest of the gas’s properties. Information theorists quantify ignorance with a function called entropy, so we ascribe to the gas a large-entropy state. Jaynes’s Principle of Maximum Entropy has spread from statistical mechanics to image processing and computer science and beyond. You can’t evade Ed Jaynes.

I decided to turn the tables on him this December. I was visiting to Washington University in St. Louis, where Jaynes worked until six years before his 1998 death. Haunted by Jaynes, I’d hunt down his ghost.


I began with my host, Kater Murch. Kater’s lab performs experiments with superconducting qubits. These quantum circuits sustain currents that can flow forever, without dissipating. I questioned Kater over hummus, the evening after I presented a seminar about quantum uncertainty and equilibration. Kater had arrived at WashU a decade-and-a-half after Jaynes’s passing but had kept his ears open.

Ed Jaynes, Kater said, consulted for a startup, decades ago. The company lacked the funds to pay him, so it offered him stock. That company was Varian, and Jaynes wound up with a pretty penny. He bought a mansion, across the street from campus, where he hosted the physics faculty and grad students every Friday. He’d play a grand piano, and guests would accompany him on instruments they’d bring. The department doubled as his family. 

The library kept a binder of Jaynes’s papers, which Kater had skimmed the previous year. What clarity shined through those papers! With a touch of pride, Kater added that he inhabited Jaynes’s former office. Or the office next door. He wasn’t certain.

I passed the hummus to a grad student of Kater’s. Do you hear stories about Jaynes around the department? I asked. I’d heard plenty about Feynman, as a PhD student at Caltech.

Not many, he answered. Just in conversations like this.

Later that evening, I exchanged emails with Kater. A contemporary of Jaynes’s had attended my seminar, he mentioned. Pity that I’d missed meeting the contemporary.

The following afternoon, I climbed to the physics library on the third floor of Crow Hall. Portraits of suited men greeted me. At the circulation desk, I asked for the binders of Jaynes’s papers.

Who? asked the student behind the granola bars advertised as “Free study snacks—help yourself!” 

E.T. Jaynes, I repeated. He worked here as a faculty member.

She turned to her computer. Can you spell that?

I obeyed while typing the name into the computer for patrons. The catalogue proffered several entries, one of which resembled my target. I wrote down the call number, then glanced at the notes over which the student was bending: “The harmonic oscillator.” An undergrad studying physics, I surmised. Maybe she’ll encounter Jaynes in a couple of years. 

I hiked upstairs, located the statistical-mechanics section, and ran a finger along the shelf. Hurt and Hermann, Itzykson and Drouffe, …Kadanoff and Baym. No Jaynes? I double-checked. No Jaynes. 

Library books

Upon descending the stairs, I queried the student at the circulation desk. She checked the catalogue entry, then ahhhed. You’d have go to the main campus library for this, she said. Do you want directions? I declined, thanked her, and prepared to return to Kater’s lab. Calculations awaited me there; I’d have no time for the main library. 

As I reached the physics library’s door, a placard caught my eye. It appeared to list the men whose portraits lined the walls. Arthur Compton…I only glanced at the placard, but I didn’t notice any “Jaynes.”

Arthur Compton greeted me also from an engraving en route to Kater’s lab. Down the hall lay a narrow staircase on whose installation, according to Kater, Jaynes had insisted. Physicists would have, in the stairs’ absence, had to trek down the hall to access the third floor. Of course I wouldn’t photograph the staircase for a blog post. I might belong to the millenial generation, but I aim and click only with purpose. What, though, could I report in a blog post? 

That night, I googled “e.t. jaynes.” His Wikipedia page contained only introductory and “Notes” sections. A WashU website offered a biography and unpublished works. But another tidbit I’d heard in the department yielded no Google hits, at first glance. I forbore a second glance, navigated to my inbox, and emailed Kater about plans for the next day.

I’d almost given up on Jaynes when Kater responded. After agreeing to my suggestion, he reported feedback about my seminar: A fellow faculty member “thought that Ed Jaynes (his contemporary) would have been very pleased.” 

The email landed in my “Nice messages” folder within two shakes. 

Leaning back, I reevaluated my data about Jaynes. I’d unearthed little, and little surprise: According to the WashU website, Jaynes “would undoubtedly be uncomfortable with all of the attention being lavished on him now that he is dead.” I appreciate privacy and modesty. Nor does Jaynes need portraits or engravings. His legacy lives in ideas, in people. Faculty from across his department attended a seminar about equilibration and about how much we can know about quantum systems. Kater might or might not inhabit Jaynes’s office. But Kater wears a strip cut from Jaynes’s mantle: Kater’s lab probes the intersection of information theory and statistical mechanics. They’ve built a Maxwell demon, a device that uses information as a sort of fuel to perform thermodynamic work. 

I’ve blogged about legacies that last. Assyrian reliefs carved in alabaster survive for millennia, as do ideas. Jaynes’s ideas thrive; they live even in me.

Did I find Ed Jaynes’s ghost at WashU? I think I honored it, by pursuing calculations instead of pursuing his ghost further. I can’t say whether I found his ghost. But I gained enough information.


With thanks to Kater and to the Washington University Department of Physics for their hospitality.

Theoretical physics has not gone to the dogs.

I was surprised to learn, last week, that my profession has gone to the dogs. I’d introduced myself to a nonscientist as a theoretical physicist.

“I think,” he said, “that theoretical physics has lost its way in symmetry and beauty and math. It’s too far from experiments to be science.”

The accusation triggered an identity crisis. I lost my faith in my work, bit my nails to the quick, and enrolled in workshops about machine learning and Chinese.

Or I might have, if all theoretical physicists pursued quantum gravity.

Quantum-gravity physicists attempt to reconcile two physical theories, quantum mechanics and general relativity. Quantum theory manifests on small length scales, such as atoms’ and electrons’. General relativity manifests in massive systems, such as the solar system. A few settings unite smallness with massiveness, such as black holes and the universe’s origin. Understanding these settings requires a unification of quantum theory and general relativity.

Try to unify the theories, and you’ll find yourself writing equations that contain infinities. Such infinities can’t describe physical reality, but they’ve withstood decades of onslaughts. For guidance, many quantum-gravity theorists appeal to mathematical symmetries. Symmetries, they reason, helped 20th-century particle theorists predict experimental outcomes with accuracies better than any achieved with any other scientific theory. Perhaps symmetries can extend particle physics to a theory of quantum gravity.

Some physicists have criticized certain approaches to quantum gravity, certain approaches to high-energy physics more generally, and the high-energy community’s philosophy and sociology. Much criticism has centered on string theory, according to which our space-time has up to 26 dimensions, most too small for you to notice. Critics include Lee Smolin, the author of The Trouble with Physics, Peter Woit, who blogs on Not Even Wrong, and Sabine Hossenfelder, who published Lost in Math this year. This article contains no criticism of their crusade. I see merit in arguments of theirs, as in arguments of string theorists.

Science requires criticism to progress. So thank goodness that Smolin, Woit, Hossenfelder, and others are criticizing string theory. Thank goodness that the criticized respond. Thank goodness that debate rages, like the occasional wildfire needed to maintain a forest’s health.

The debate might appear to impugn the integrity of theoretical physics. But quantum gravity constitutes one pot in the greenhouse of theoretical physics. Theoretical physicists study lasers, star formation, atomic clocks, biological cells, gravitational waves, artificial materials, and more. Theoretical physicists are explaining, guiding, and collaborating on experiments. So many successes have piled up recently, I had trouble picking examples for this article. 

One example—fluctuation relations—I’ve blogged about beforeThese equalities generalize the second law of thermodynamics, which illuminates why time flows in just one direction. Fluctuation relations also provide a route to measuring an energetic quantity applied in pharmacology, biology, and chemistry. Experimentalists have shown, over the past 15 years, that fluctuation relations govern RNA, DNA, electronic systems, and trapped ions (artificial atoms). 

Second, experimentalists are exercising, over quantum systems, control that physicists didn’t dream of decades ago. Harvard physicists can position over 50 atoms however they please, using tweezers formed from light. Google has built a noisy quantum computer of 72 superconducting qubits, circuits through which charge flows without resistance. Also trapped ions, defects in diamonds, photonics, and topological materials are breaking barriers. These experiments advance partially due to motivation from theorists and partially through collaborations with theorists. In turn, experimental data guide theorists’ explanations and our proposals of experiments.

In one example, theorists teamed with experimentalists to probe quantum correlations spread across space and time. In another example, theorists posited a mechanism by which superconducting qubits interact with a hot environment. Other illustrations from the past five years include discrete time crystals, manybody scars, magic-angle materials, and quantum chaos. 

These collaborations even offer hope for steering quantum gravity with experiments. Certain quantum-gravity systems share properties with certain many-particle quantum systems. This similarity, we call “the AdS/CFT duality.” Experimentalists have many-particle quantum systems and are stretching those systems toward the AdS/CFT regime. Experimental results, with the duality, might illuminate where quantum-gravity theorists should and shouldn’t search. Perhaps no such experiments will take place for decades. Perhaps AdS/CFT can’t shed light on our universe. But theorists and experimentalists are partnering to try.

These illustrations demonstrate that theoretical physics, on the whole, remains healthy, grounded, and thriving. This thriving is failing to register with part of the public. Evidence thwacked me in the face last week, as explained at the start of this article. The Wall Street Journal published another example last month: John Horgan wrote that “physics, which should serve as the bedrock of science, is in some respects the most troubled field of” science. The evidence presented consists of one neighborhood in the theoretical fraction of the metropolis of physics: string and multiverse models.

Horgan’s article reflects decades of experience in science journalism, a field I respect. I sympathize, moreover, with those who interface so much with quantum gravity, the subfield appears to eclipse the rest of theoretical physics. Horgan was reviewing books by Stephen Hawking and Martin Rees, who discuss string and multiverse models. Smolin, Woit, Hossenfelder, and others garner much press, which they deserve: They provoke debate and articulate their messages eloquently. Such press can blot out, say, profiles of the theoretical astrophysicists licking their lips over gravitational-wave data.

If any theory bears flaws, those flaws need correcting. But most theoretical physicists don’t pursue quantum gravity, let alone string theory. Any flaws of string theory do not mar all theoretical physics. These points need a megaphone, because misconceptions about theoretical physics endanger society. First, companies need workers who have technical skills and critical reasoning. Both come from training in theoretical physics. Besmirching theoretical physics can divert students from programs that can benefit the economy and nurture thoughtful citizens.1 

Second, some nonscientists are attempting to discredit the scientific community for political gain. Misconceptions about theoretical physics can appear to support these nonscientists’ claims. The ensuing confusion can lead astray voters and parents who face choices about vaccination, global health, national security, and budget allocations.

Last week, I heard that my profession has wandered too far from experiments. Hours earlier, I’d skyped with an experimentalist with whom I’m collaborating. A disconnect separates the reality of theoretical physicists from impressions harbored by part of the public. Let’s clear up the misconceptions. Theoretical physics, as a whole, remains healthy, grounded, and thriving.



1Nurturing thoughtful citizens takes also humanities, social-sciences, language, and arts programs.

A Roman in a Modern Court

Yesterday I spent some time wondering how to explain the modern economy to an ancient Roman brought forward from the first millennium BCE. For now I’ll assume language isn’t a barrier, but not much more. Here’s my rough take:

“There have been five really important things that were discovered since when you left and now.

First, every living thing has a tiny blueprint inside it. We learned how to rewrite those, and now we can make crops that resist pests, grow healthy, and take minimal effort to cultivate. The same tool also let us make creatures that manufacture medicine, as well as animals different from anything that existed before. Food became cheap because of this.

Second, we learned that hot air and steam expand. This means you can burn oil or coal and use that to push air around, which in turn can push against solid objects. With this we’ve made vehicles that can go the span of the Empire from Rome to Londinium and back in hours rather than weeks. Similar mechanisms can be used to work farms, forge metal, and so on. Manufactured goods became cheap as a result.

Third, we discovered an invisible fluid that lives in metals. It flows unimaginably quickly and with minimal force through even very narrow channels, so by pushing on it in one city it may be made to move almost instantly in another. That lets you work with energy as a kind of commodity, rather than something that hooks up and is generated specifically for each device.

Fourth, we found that this fluid can be pushed around by light, including a kind human eyes can’t see. This lets a device make light in one place and push on the fluid in a different device with no metal in between. Communication became fast, cheap, and easy.

Finally, and this one takes some explaining, our machines can make decisions. Imagine you had a channel for water with a fork. You can insert a blade to control which route the water takes. If you attach that blade to a lever you can change the direction of the flow. If you dip that lever in another channel of water, then what flows in one channel can set which way another channel goes. It turns out that that’s all you need to make simple decisions like “If water is in this channel, flow down that other one.”, which can then be turned into useful statements like “Put water in this channel if you’re attacked. It’ll redirect the other channel and release boiling oil.” With enough of these water switches you can do really complicated things like tracking money, searching for patterns, predicting the weather, and so on. While water is hard to work with, you can make these channels and switches almost perfect for the invisible fluid, and you can make them tiny, vastly smaller than the width of a hair. A device that fits in your hand might have more switches than there are grains in a cubic meter of sand. The number of switches we’ve made so far outnumbers all the grains of sand on Earth, and we’re just getting started.”

“Methinks, I know one kind like you.”

I was expecting to pore over a poem handwritten by one of history’s most influential chemists. Sir Humphry Davy lived in Britain around the turn of the 19th century. He invented a lamp that saved miners’ lives, discovered and isolated chemical elements, coined the term “laughing gas,” and inspired younger researchers through public lectures.


Humphry Davy

Davy wrote not only scientific papers, but also poetry. He befriended contemporaries known today as “Romantic poets,” including Samuel Taylor Coleridge. English literature and the history of science rank among the specialties of the Huntington Library in San Marino, CA. The Huntington collects manuscripts and rare books, and I secured a reader card this July. I aspired to find a poem by Davy.

Bingo: The online catalogue contained an entry entitled “To the glow worm.” I requested the manuscript and settled into the hushed, wood-paneled reading room.

Davy had written scarcely legibly, in black ink, on a page that had creased and torn. I glanced over the lines, then realized that the manuscript folder contained two other pages. The pages had stuck together, so I gently flipped the lot over.

Davy poem

Poem “To the glow worm,” by Humphry Davy

A line at the top of the back page seized the wheel of my attention.

“Methinks, I know one kind like you.”

The line’s intimacy arrested me. I heard a speaker contemplating someone whom he or she had met recently, turning the person over in the speaker’s mind, gaining purchase on the person’s identity. “I know you,” I heard the speaker saying, and I saw the speaker wagging a finger at the person. “I know your type…I think.”

The line’s final six words suggested impulsiveness. How can you know someone you’re still wrapping your head around? I felt inclined to suggest a spoonful of circumspection. But perhaps the speaker was reflecting more than I’d allowed: “Methinks” suggested temperance, an acknowledgement of uncertainty.

I backpedaled to the folder’s cover. “Includes verse and letter by Lady Davy,” it read. Jane Apreece, a wealthy widow, acquired the title Lady Davy upon marrying Sir Humphry. She enjoyed a reputation for social savvy, fashionableness, and sharpness. I’d intruded on her poem, a response to Davy’s. Apreece’s pages begged for a transcription, which I struggled through until the reading room closed 45 minutes later. Dan Lewis, the Huntington’s Dibner Senior Curator of the History of Science and Technology, later improved upon my attempt (parenthesized text ours):

Methinks, I know one kind like you,

Thine(?) to peace, & Nature true;

Kindled by Feeling’s purest flame,

In Storm, or Calm, for ages(?) the same.

Bestowing most its brilliant Light,

Amidst the tranquil shades of Night;

And prompt to solace, raise, & cheer(?),

The heart, subdued by Doubt or Care.

Though not of busy Life afraid

Yet loving best, the pastoral Shade;

Shedding a Ray, more clear & pure,

A Ray, which longer shall endure,

As Friendships light must ever prove

More steadfast than the Flame of Love.

Light recurs throughout the verse: The speaker refers to two flames, to a “Ray,” and to a “brilliant Light // Amidst the tranquil shades of Night.” Comparisons with light suit a scientist, who reveals aspects of nature never witnessed before. (I expect that the speaker directs the apostrophe toward Davy.) Comparisons with light suit Davy not only professionally, but also, to Apreece, personally: Each member of the couple inspired the other to learn. Their poems reflect their intellectual symbiosis: Apreece’s references to light complement the glow worm, which Davy called “lively living lamp of night.”

The final two lines arrested me as the first line did. The speaker contrasts “Friendship[’]s light” with “the Flame of Love.” Finite resources can’t sustain flames, which consume candles, wood, and oxygen. Once its fuel disappears, flame proves less than “steadfast.” Similarly, love can’t survive on passion’s flames. Love should rest on friendship, which sheds the “light” extolled throughout the poem. Light enhances our vision, providing the wisdom needed to sustain love throughout life’s vicissitudes. 

These two lines reveal the temperance hinted at by the “Methinks.” The speaker argues for levelheadedness, for balancing emotion with sustainability. Spoonful of circumspection retracted.


The clock struck 4:45, and readers began returning their manuscripts and books to the circulation desk. I stood up—and pricked myself on a thorn of realization. The catalogue dated the manuscript to “perhaps [ . . . ] 1811 – they [Davy and Apreece] were married in 1812.” The lovers exchanged these poems without knowing that their marriage would sour years later. I’d read about their relationship—as about Davy’s science and poetry—in Richard Holmes’s The Age of Wonder. 

At least the Davys reunited when Sir Humphry’s last illness struck. At least they remained together until he died. At least a reader can step, through the manuscript, into the couple’s patch of happiness. One can hope see more clearly for their—a scientist’s, a societal navigator’s, and two human beings’—light.

Jane letter + p.1 of poem.JPG

Letter and poem by Jane Apreece (p. 1). The top segment constitutes a letter written “by Lady Davy to a ‘Miss Talbot’ (1852, January 2),” according to the catalogue.

Jane poem, p. 2.JPG

Poem by Jane Apreece (p. 2)

If anyone has insights or has corrections to the transcription, please comment. I haven’t transcribed Davy’s poem, which might illuminate Lady Davy’s response.

With thanks to the Huntington Library of San Marino, CA, for the use of its collection. With thanks to Dan Lewis for improving upon my transcription and for prodding, for five years, toward a reader card.

Doctrine of the (measurement) mean

Don’t invite me to dinner the night before an academic year begins.

You’ll find me in an armchair or sitting on my bed, laptop on my lap, journaling. I initiated the tradition the night before beginning college. I take stock of the past year, my present state, and hopes for the coming year.

Much of the exercise fosters what my high-school physics teacher called “an attitude of gratitude”: I reflect on cities I’ve visited, projects firing me up, family events attended, and subfields sampled. Other paragraphs, I want off my chest: Have I pushed this collaborator too hard or that project too little? Miscommunicated or misunderstood? Strayed too far into heuristics or into mathematical formalisms?

If only the “too much” errors, I end up thinking, could cancel the “too little.”

In one quantum-information context, they can.


Imagine that you’ve fabricated the material that will topple steel and graphene; let’s call it a supermetatopoconsulator. How, you wonder, do charge, energy, and particles move through this material? You’ll learn by measuring correlators.

A correlator signals how much, if you poke this piece here, that piece there responds. At least, a two-point correlator does: \langle A(0) B(\tau) \rangle. A(0) represents the poke, which occurs at time t = 0. B(\tau) represents the observable measured there at t = \tau. The \langle . \rangle encapsulates which state \rho the system started in.

Condensed-matter, quantum-optics, and particle experimentalists have measured two-point correlators for years. But consider the three-point correlator \langle A(0) B(\tau) C (\tau' ) \rangle, or a k-point \langle \underbrace{ A(0) \ldots M (\tau^{(k)}) }_k \rangle, for any k \geq 2. Higher-point correlators relate more-complicated relationships amongst events. Four-pointcorrelators associated with multiple times signal quantum chaos and information scrambling. Quantum information scrambles upon spreading across a system through many-body entanglement. Could you measure arbitrary-point, arbitrary-time correlators?

New material

Supermetatopoconsulator (artist’s conception)

Yes, collaborators and I have written, using weak measurements. Weak measurements barely disturb the system being measured. But they extract little information about the measured system. So, to measure a correlator, you’d have to perform many trials. Moreover, your postdocs and students might have little experience with weak measurements. They might not want to learn the techniques required, to recalibrate their detectors, etc. Could you measure these correlators easily?

Yes, if the material consists of qubits,2 according to a paper I published with Justin Dressel, José Raúl González Alsonso, and Mordecai Waegell this summer. You could build such a system from, e.g., superconducting circuits, trapped ions, or quantum dots.

You can measure \langle \underbrace{ A(0) B (\tau') C (\tau'') \ldots M (\tau^{(k)}) }_k \rangle, we show, by measuring A at t = 0, waiting until t = \tau', measuring B, and so on until measuring M at t = \tau^{(k)}. The t-values needn’t increase sequentially: \tau'' could be less than \tau', for instance. You’d have to effectively reverse the flow of time experienced by the qubits. Experimentalists can do so by, for example, flipping magnetic fields upside-down.

Each measurement requires an ancilla, or helper qubit. The ancilla acts as a detector that records the measurement’s outcome. Suppose that A is an observable of qubit #1 of the system of interest. You bring an ancilla to qubit 1, entangle the qubits (force them to interact), and look at the ancilla. (Experts: You perform a controlled rotation on the ancilla, conditioning on the system qubit.)

Each trial yields k measurement outcomes. They form a sequence S, such as (1, 1, 1, -1, -1, \ldots). You should compute a number \alpha, according to a formula we provide, from each measurement outcome and from the measurement’s settings. These numbers form a new sequence S' = \mathbf{(} \alpha_S(1), \alpha_S(1), \ldots \mathbf{)}. Why bother? So that you can force errors to cancel.

Multiply the \alpha’s together, \alpha_S(1) \times \alpha_S(1) \times \ldots, and average the product over the possible sequences S. This average equals the correlator \langle \underbrace{ A(0) \ldots M (\tau^{(k)}) }_k \rangle. Congratulations; you’ve characterized transport in your supermetatopoconsulator.


When measuring, you can couple the ancillas to the system weakly or strongly, disturbing the system a little or a lot. Wouldn’t strong measurements perturb the state \rho whose properties you hope to measure? Wouldn’t the perturbations by measurements one through \ell throw off measurement \ell + 1?

Yes. But the errors introduced by those perturbations cancel in the average. The reason stems from how we construct \alpha’s: Our formula makes some products positive and some negative. The positive and negative terms sum to zero.

Balance 2

The cancellation offers hope for my journal assessment: Errors can come out in the wash. Not of their own accord, not without forethought. But errors can cancel out in the wash—if you soap your \alpha’s with care.


1and six-point, eight-point, etc.

2Rather, each measured observable must square to the identity, e.g., A^2 = 1. Qubit Pauli operators satisfy this requirement.


With apologies to Aristotle.