Freeman Dyson and Boris Tsirelson

Today, as the world braces for the possibility of losing millions of lives to the new coronavirus—to the hunger for pangolin meat, of all things (combined with the evisceration of competent public health agencies like the CDC)—we also mourn the loss of two incredibly special lives, those of Freeman Dyson (age 96) and Boris Tsirelson (age 69).

Freeman Dyson was sufficiently legendary, both within and beyond the worlds of math and physics, that there’s very little I can add to what’s been said. It seemed like he was immortal, although I’d heard from mutual friends that his health was failing over the past year. When I spent a year as a postdoc at the Institute for Advanced Study, in 2004-5, I often sat across from Dyson in the common room, while he drank tea and read the news. That I never once struck up a conversation with him is a regret that I’ll now carry with me forever.

My only exchange with Dyson came when he gave a lecture at UC Berkeley, about how life might persist infinitely far into the future, even after the last stars had burnt out, by feeding off steadily dimishing negentropy flows in the nearly-thermal radiation. During the Q&A, I challenged Dyson that his proposal seemed to assume an analog model of computation. But, I asked, once we took on board the quantum-gravity insights of Jacob Bekenstein and others, suggesting that nature behaves like a (quantum) digital computer at the Planck scale, with at most ~1043 operations per second and ~1069 qubits per square meter and so forth, wasn’t this sort of proposal ruled out? “I’m not going to argue with you,” was Dyson’s response. Yes, he’d assumed an analog computational model; if computation was digital then that surely changed the picture.

Sometimes—and not just with his climate skepticism, but also (e.g.) with his idea that general relativity and quantum mechanics didn’t need to be reconciled, that it was totally fine for the deepest layer of reality to be a patchwork of inconsistent theories—Dyson’s views struck me as not merely contrarian but as a high-level form of trolling. Even so, Dyson’s book Disturbing the Universe had had a major impact on me as a teenager, for the sparkling prose as much as for the ideas.

With Dyson’s passing, the scientific world has lost one of its last direct links to a heroic era, of Einstein and Oppenheimer and von Neumann and a young Richard Feynman, when theoretical physics stood at the helm of civilization like never before or since. Dyson, who apparently remained not only lucid but mathematically powerful (!) well into his last year, clearly remembered when the Golden Age of science fiction looked like simply sober forecasting; when the smartest young people, rather than denouncing each other on Twitter, dreamed of scouting the solar system in thermonuclear-explosion-powered spacecraft and seriously worked to make that happen.

Boris Tsirelson (homepage, Wikipedia), who emigrated from the Soviet Union and then worked at Tel Aviv University (where my wife Dana attended his math lectures), wasn’t nearly as well known as Dyson to the wider world, but was equally beloved within the quantum computing and information community. Tsirelson’s bound, which he proved in the 1980s, showed that even quantum mechanics could only violate the Bell inequality by so much and by no more, could only let Alice and Bob win the CHSH game with probability cos2(π/8). This seminal result anticipated many of the questions that would only be asked decades later with the rise of quantum information. Tsirelson’s investigations of quantum nonlocality also led him to pose the famous Tsirelson’s problem: loosely speaking, can all sets of quantum correlations that can arise from an infinite amount of entanglement, be arbitrarily well approximated using finite amounts of entanglement? The spectacular answer—no—was only announced one month ago, as a corollary of the MIP*=RE breakthrough, something that Tsirelson happily lived to see although I don’t know what his reaction was (update: I’m told that he indeed learned of it in his final weeks, and was happy about it). Sadly, for some reason, I never met Tsirelson in person, although I did have lively email exchanges with him 10-15 years ago about his problem and other topics. This amusing interview with Tsirelson gives some sense for his personality (hat tip to Gil Kalai, who knew Tsirelson well).

Please share any memories of Dyson or Tsirelson in the comments section.

41 Responses to “Freeman Dyson and Boris Tsirelson”

  1. Anthony Says:

    Boris Tsirelson apparently had the time to learn about the result:

    Hat tip @Numerigo on twitter:

  2. David Says:

    I met Dyson once 30 years ago when I was a young graduate student. He was giving a talk at CERN about global warming – back then, he was very worried about it and thought people should do something about it. After the talk, he wanted a tour of my experiment where we were measuring the mass of the antiproton. I started with a very basic sketch of the experiment and he kept encouraging me to get into the nitty-gritty details as he’d already read the papers and understood how it worked. He was more interested in what the systematic errors might be. I tried to answer his questions, though I was a baby graduate student at the time.

  3. Chris Says:

    “With Dyson’s passing, the scientific world has lost one of its last direct links to a heroic era, of Einstein and Oppenheimer and von Neumann and a young Richard Feynman, when theoretical physics stood at the helm of human civilization like never before or since. ”

    I think this means it’s time to enter a new Golden Age.

    People will be talking about how they once met the great Scott Aaronson.


  4. Dana Moshkovitz Says:

    My most vivid memory of Boris Tsirelson was his demonstration of Brownian motion in class. Short and highly energetic, he traced his own motion across the room.
    I’m deeply saddened about his illness and untimely death. He was an excellent teacher.

  5. Mateus Araújo Says:

    I’ve also written a bit about Tsirelson here.

    I do have a memory to share: back when I was an undergrad, I was studying Bell inequalities for the first time, and landed on Tsirelson’s famous paper. Not understanding it well, I wrote to Tsirelson a couple of inane questions about it, who answered them patiently and politely. I also tried reading the Wikipedia article about Tsirelson’s bound to understand what I was going on, but the article was a terrible mess. After some more time studying, I decided I understood the subject well enough to improve on this terrible article, and completely rewrote it. To my surprise, Tsirelson himself kept a watch on the article and thanked me for improving it (Tsirelson was an avid contributor to Wikipedia, but he did not edit the articles about his own work to avoid a conflict of interest).

  6. Big numbers Says:

    Looks like Tsirelson’s last(?) article related to your first(?):

  7. Baruch Garcia Says:

    I am sorry to hear of the passing of Freeman Dyson. I had the pleasure of meeting both him and his lovely wife a few years back while passing through Princeton. He invited me into his office, and we spoke at length about the future of physics and various people he knew throughout the years. He did not care much for Wittgenstein, who lived a few doors down from him at Cambridge, and he knew a side of Godel few others got to see, a caring and welcoming character who invited the young scientist over for tea. His unorthodoxy and kindness combined to make a unique blend. He will be missed.

  8. murmur Says:

    Dyson led the planning for the Project Orion, the nuclear spacecraft. Nuclear propulsion is the only currently available technology that allows humans to explore the solar system. Unfortunately, it never got off the ground due to the anti-nuclear hysteria. Future generations will probably see this point as the time when our 400 year march of scientific progress stalled and eventually came to an end.

  9. James Gallagher Says:

    Dyson was born the year before de Broglie proposed his theory of electron waves, his life really did straddle the entire development of Quantum Theory, and he made a great contribution himself, one of the last survivors of the “Golden Age in Physics”

    It’s still a pleasure to be able read through his own handwritten notes of the famous 1951 Cornell Lecture Course on Quantum Electrodynamics:

    (These were typed up and published as Advanced Quantum Mechanics)

  10. Scott Says:

    murmur #8:

      Unfortunately, [Project Orion] never got off the ground due to the anti-nuclear hysteria. Future generations will probably see this point as the time when our 400 year march of scientific progress stalled and eventually came to an end.

    There are days when I 100% agree with that.

    There are other days when I think that manned exploration of the solar system was never such a great idea. It’s like, Golden Age sci-fi portrayed the other planets and moons as harsh, to be sure, but only somewhat more so than (say) the American Wild West. That turned out to be, literally, one of the understatements of the century. Other planets are so harsh that, for purposes of human habitation, one probably might as well just put artificial space stations in free orbit around the sun. (Under the ocean, or in Antarctica, would of course be paradise compared to either.) I guess Mars has some ice and gravity, but arguably not much more of interest to a potential colonist (as opposed to a planetary scientist).

    Furthermore, while progress seems to have stagnated in many forms of technology since the 60s, and while that sucks, the progress in computing and communications has been unbelievably spectacular—as in, it’s fulfilled much of the promise of Star Trek three centuries ahead of schedule. Why has the progress been so lopsided? Well, since computers (by definition) are universal machines, is it possible that in the limit, all technological progress starts to look more and more like progress in computing and communications, that this is a central insight that the Golden Age sci-fi writers missed?

    Related to that, is it possible that the “frontiers” have simply closed? I.e., that once you’ve settled most of the earth (to the point that what was once “the wilderness” has been reduced to minuscule theme parks), visited the moon and the bottom of the ocean (and filled the latter with plastic), set up regular around-the-world commercial air travel, etc. etc.—by that point, the remaining challenges are either nasty byproducts of your own past success (climate change), or else eye-wateringly complicated and hard (curing cancer), or else computer- or communication-related?

    Anyway, so far there’s been no reconciliation between the two worldviews, just days for one and days for the other.

    (Note: if someone wants to respond to the above, but is worried about going too far off-topic, I suppose the relevant question should be, what would Freeman Dyson have considered too far off-topic for his memorial thread? 🙂 )

  11. Scott Says:

    Amusing story: Lily, my 7-year-old, saw me blogging just now and asked me who Freeman Dyson was. So I told her … about atomic-bomb-powered spaceships, and the story of how they got cancelled by the US government in 1963, and also about Dyson spheres, and why they’re so far beyond what human beings can build today, and have also never been seen in space although people have started to look for them. (I figured that the equivalence of Feynman and Schwinger-Tomonaga QED would go a little over her head.)

    Lily listened with great interest but then asked: “so did he ever invent anything that got finished and worked?” I said that that was a good question, but in the realm of practical technologies I wasn’t sure.

  12. Raoul Ohio Says:

    #8, #10, I want to remind everyone of my longstanding even money wager that “No human will travel to the surface of Mars and return alive to the surface of the Earth by 2100”. Please send money now, and I will return double your money when it happens, or in 2100, whichever comes later.

  13. Scott Says:

    A reader pointed me to this extremely interesting Science article, which explains why the transmission vector for the new coronavirus might not have been pangolins after all—it could have been, but it could also have been some other animal sold at the Huanan seafood market.

  14. Aaron G Says:

    Scott #13, with respect to the COVID-19 outbreak, it is worth keeping in mind the following:

    1. This virus most likely emerged from a virus hosted in animal from the live animal markets in China to humans, but viruses constantly mutate and on occasion do so to move from one host animal to another in a variety of circumstances, leading to new outbreaks. This is an inevitable process of evolution.

    2. All information available thus far indicates that the vast majority of those infected who are symptomatic (something like 80-85% of those infected) only suffer mild symptoms. Of those who suffer more severe symptoms (or those who have died), the vast majority were either elderly, had pre-existing illnesses or medical conditions, or had weakened immune systems.

    3. It appears that large numbers of those infected are asymptomatic (although estimates are hard to come by).

    4. The case fatality rate (i.e. percentage of deaths among confirmed cases of infection) falls somewhere between 1-3%, which meand 97-99% of those confirmed infections survive. However, given #3 above, the actual fatality rate from COVID-19 is likely far lower.

  15. Aaron G Says:

    murmur #8 and Scott #10,

    While I agree that progress in computing and communications over the past several decades has been spectacular, the implication that this somehow represents the last “frontier” of science seems misplaced.

    After all, as of 2020, scientists still have only a partial understanding of the workings of the human brain and more broadly of biological neural networks. And there are many areas of biology which continue to be amenable to further exploration that isn’t directly computing related (although advances in computing will serve as an important tool for advances in biological research).

    If anything, could one not make the argument that the 21st century could see the fulfilment of the initial promise of research in complex systems, due in no small part to the increasing ubiquity of data & the corresponding computing power to store & analyze such data?

  16. Noah Says:

    Is it unscientific of me, upon hearing of Dyson’s death, to see 96 and think his abnormally passionate interest in the study of nature boosted his longevity? Irrational as it is, I kinda believe that…

    Given his brilliance, his near centenarian status and that he lived during a time when society could provide him with the resources to do serious science, he’s probably in the running for the “most cumulative brain time spent nerding out over the universe” world record! Who could beat him on that?

    And wrt climate change, maybe his legacy there is not yet set in stone… Perhaps someone could appropriate his idea for a plant capable of living on a comet in space and genetically engineer a plant instead optimized to convert C02 into an as massive as possible herbaceous tumor. Toss a few starter cultures of Dyson Seaweed into the Pacific and in a year or two: C02’s gone and the giant kelp tumor has coalesced with the pacific garbage island creating a new continent that we name in his honor.

  17. fred Says:

    Scott #10

    “is it possible that in the limit, all technological progress starts to look more and more like progress in computing and communications, that this is a central insight that the Golden Age sci-fi writers missed?”

    The new frontiers are not physical.
    It’s a matter of time for any intelligent life to realize that the physical world is arbitrarily limited and pretty much a dead end.
    Minds are driven by curiosity and social interactions, and eventually they develop technologies that allow them to live “within” by increasing their communication bandwidth and build/explore/enjoy together an infinite variety of virtual worlds.
    Those virtual worlds will cover everything: from basic natural realities (nostalgia for surviving with a fragile body in a harsh environment) to the most abstract realms (pure mathematics).

  18. Gerard Says:

    Scott #10

    “Well, since computers (by definition) are universal machines, is it possible that in the limit, all technological progress starts to look more and more like progress in computing and communications”

    First comment:

    Computers are universal computing machines, not universal machines. If the only machine you have is a computer and your goal is to build a skyscraper, you would be kind of stuck, I think.

    Second comment:

    I’ve reread this sentence a few times and I’m at a loss as to what it is intended to mean. It’s not that I can’t imagine several possible meanings but that I’m not sure which of those you were aiming at (or something else perhaps). I’m hoping you could provide some clarification/elaboration (maybe even an example).

  19. Scott Says:

    Gerard #18: One obvious example would be that, when I was a kid, we owned separate objects called “calculator,” “camera,” “video camera,” “photo album,” “flashlight,” “notebook,” “calendar,” “alarm clock,” “phone,” “answering machine,” “beeper,” “video game console,” “music player” … whereas now we carry around a single object in our pockets that simulates any of those and countless others as needed. Yes, of course, to get the benefit you need the appropriate input/output mechanisms: computers can simulate the internal logic of any machine, but not necessarily its effects on or input from the external world. The point is that we seem to be clearly remaking the world in such a way that more and more can be done just by changing the internal logic—i.e., by writing software.

  20. Gerard Says:

    Scott #19

    “The point is that we seem to be clearly remaking the world in such a way that more and more can be done just by changing the internal logic—i.e., by writing software.”

    Yes, I (and I suspect most people here) would agree that we are on that path, though I also think we’re still quite far from the endpoint.

    The endpoint is perhaps a true universal machine or fabricator, that is one capable of constructing any other machine, including itself.

    One might think humans fit that definition, but I don’t consider them to be machines. I would define a machine as a physical system which is capable of performing one or more tasks reliably (ie. deterministically). That implies that though a machine is composed of elements that are subject to randomness, it is constructed in such a way as to eliminate or at least greatly suppress the effects of that randomness on it’s functionality.

    By the way I don’t think that a machine, as I defined it, can possess either subjective experience or free-will in anything like the way we do.

  21. Gerard Says:


    There appears to be a bug in your comment submission code. I tried to submit a new comment but I get the error “Duplicate comment detected”.

  22. N2mine Says:

    I’d say the development of computing over the past couple of decades is presents quite a good case that progress has stalled out. Apart from video, “smartphones” really aren’t being used for much that an NTT keitai denwa circa 2000 couldn’t do despite having ~1000x the computing power. The very fact that they’re called “smartphones” and not “pocket computers” or even “pocket supercomputers” speaks to an appalling poverty of imagination.

    I think the best analogy is with imperial China, it had a near-miss with industrialization under the Northern Song, but after the Song downfall subsequent ages never managed to break out of the straightjacket of a medieval political economy despite continued technological advancement.

    Or closer to home, The Ascent of Man pointed out that while 18th century continental Europe was in many respects ahead of the British isles in mechanical ingenuity, it mostly got squandered on constructing useless trinkets designed to edify the rich and privileged while Britain forged onward with creating a whole new industrial way of life.

  23. Mateus Araújo Says:

    Gerard #18: I think you would really enjoy chapter 5 of David Deutsch’s book The Fabric of Reality.

  24. Scott Says:

    Gerard #21: Sorry, but don’t worry about it! When that happens, it means that the comment indeed went through.

    (I confess that the Star Trek computer probably did not have these bizarre bugs. 😀 )

  25. mjgeddes Says:

    Dyson seems to have had a similar vision to David Deutsch – the idea that the possibility of progress is in some sense unbounded – ‘Infinite In All Directions’ /’The Beginning of Infinity’.

    I have an idea of how this might work. Is it scientific? No. Wild speculation? Yes ! 😀

    In the modeling of knowledge, I see 3 levels of abstraction, and I think this is recursive…no matter how far in or out you zoom, you can always perform this 3-level splitting of your ontological models.

    So I think that 2 of the ontological ‘views’ are *objective* (a high-level representation and a low-level representation of a given domain), but the 3rd view is *subjective* (it’s the language that performs the mapping between the high- and low-level views).

    So the idea is that perhaps the creation of the universe isn’t complete… it’s actually on-going! There are objective components (what *is*), but there’s always a subjective component (what *could be*). So I’m suggesting that there’s always this creative component to reality.

    Let me give you an example of this by performing the 3-level splitting for the domain ‘Mathematics’.

    Here are what I think the 2 *objective* ‘views’ of math are:

    (1) Algebra – LOW-LEVEL VIEW
    (2) Analysis – HIGH-LEVEL VIEW

    And here’s the 3rd *subjective* view of math:

    (3) Logic – MAPPING LANGUAGE

    Do you see the idea? Algebra and Analysis are objectively real components of reality, but Logic is not, it’s a human creation. The idea is that logic is the *mapping* between the objective components of math- logic generates the dualities that map concepts from algebra to concepts from analysis. But this process is always a creative act…a work in progress that will never end.

    Now this 3-level ontological split, I think, applies to all knowledge domains as I mentioned, you can zoom in and out and find the same components.

    So let’s zoom right out to the highest level of abstraction – knowledge domain ‘Reality’ (or ‘Universe’) and attempt the same 3-level splitting. Here’s the result:

    2 objective components…

    (1) Matter (Physics) – LOW-LEVEL VIEW
    (2) Abstraction (Math) – HIGH-LEVEL VIEW

    And 1 subjective mapping language… drum-roll please…

    (3) Thought (Mind) – MAPPING LANGUAGE !

    Could minds *be* the creative element to reality that continues the process of creation itself?

    Could minds *be* the mapping that links matter and abstraction (physics and math) together?

  26. Norm Margolus Says:

    Scott #10

    “Furthermore, while progress seems to have stagnated in many forms of technology since the 60s, and while that sucks, the progress in computing and communications has been unbelievably spectacular—as in, it’s fulfilled much of the promise of Star Trek three centuries ahead of schedule. Why has the progress been so lopsided?”

    Most technology needs to be macroscopic for us to use it. A microscopic car isn’t useful. Microscopic bits are. Our technologies were using such a small fraction of the computational capacity of matter that we were able to double that fraction every couple of years for about eight decades. It’s not crazy to consider all physical dynamics as computation.

  27. fred Says:

    Gerard #20

    “The endpoint is perhaps a true universal machine or fabricator, that is one capable of constructing any other machine, including itself.”

    I recently got into 3D printing via a cheap 250$ device (Ender 3), and while this technology is only baby steps in the direction of a universal fabricator, I’ve been blown away by it.
    On the physical engineering side, I think that the two major breakthroughs in the last couple decades have been strong magnets and battery improvements, allowing for super compact and powerful electrical motors: it’s allowed for things like drones, agile robots (Boston Dynamics stuff), and 3D printers… but all those things are mainly software driving very precise step motors.

  28. Pangolin3 Says:

    Note there is no evisceration of the CDC, administration proposed funding cut but it was ignored by a bipartisan congress, congress actually raised CDC’s budget. So any incompetence on CDC’s part is not caused by funding issues.


  29. Doug Says:

    I think it’s right to class all of our progress into “information,” but I think the other broad possible class where we still need work would be “energy.” We’ve made only paltry progress in batteries and solar cells in the years that computers have made huge bounds, and a lot of that only just recently. Sustainable, useful fusion and low temperature super conductors are out of our grasp. We strain under our grid and supply systems. Right now we’ve been doing more with more – and our environment has been seriously paying the price of our increased energy needs. I think we’re overdue for a set of revolutions int he “energy” side of the story.

  30. Stella Biderman Says:

    Katherine Johnson, a NASA computer and applied mathematician who was the focus of the movie Hidden Figures, also passed away this week.

  31. Scott Says:

    Stella #30: Yes, thanks for adding that!

  32. Numerigo Says:

    Boris Tsirelson, always the enlightened mind. I think he felt pride in his final decision. It took courage, but was also very rational.

  33. Raoul Ohio Says:

    Noah, #16: Two obvious scientists are Humboldt and Wigner, who both lived into their 90’s and were reportedly sharp to the end.

    I am confident that Knuth will hang in until he finishes vol 7 of “Art of Computer Programming”, at which time he will be well over 200.

  34. Jelmer Renema Says:

    I must confess that my first thought when reading this was ‘ah yes, that Heroic Age of science, when the brightest minds, instead of denouncing each other on Twitter, denounced each other in AEC security hearings’. I agree with your eulogies of Tsirelson and Dyson, but blanket idolization of the past never ends well.

  35. Stella Biderman Says:

    My girlfriend asked me who Dyson was, after seeing me text a friend about his death. I explained that he was a physicist who was well known for his futurology and had worked on the Manhattan Project.

    I’m wrong, he didn’t work on the Manhattan Project and he got his BA in 1945. However he was at IAS and I seem to have a shortcut in my head that says “famous physicists at Chicago or IAS who pass away worked on the Manhattan project.” I tried to find living scientists who had worked on the Manhattan Project and was unable to turn any up. Does anyone know if any of them are still alive?

    Raul #33: Knuth and Berry Mazur are 82. They are only octogenarians who comes to mind as being active in mathematics and computer science.

  36. OC Says:

    In the late ’70s I had the interesting experience of taking a “reading course” with Dyson, as an undergrad at the university down the road from the IAS. I had become very interested in issues related to reducing nuclear weaponry, which, IIRC, Dyson was engaged in promoting (the reduction, that is) in some fashion that now escapes me.

    A friend, who was much bolder than I, suggested that we see if he would be our “tutor” (British style) for the semester on some related topic, and he agreed to meet with us once a week. My main recollection from the experience was of finding conversation with him both fascinating and discomfiting. Frequently what seemed like an ordinary conversational pause, where we were expecting him to say something, would turn into 10, then 20, then 30 seconds of silence. Finally when we were ready to explode and one of us was about to say whatever popped into his head just to keep the conversation going, Dyson would start to speak, producing a paragraph or two of perfectly formed ideas. You could have transcribed his speech with no editing at all into a clear and cogent text.

    A few years later, when I was a grad student, he came to Berkeley for a month or so, and I mustered my courage and invited him to join a group of us for dinner at a local Chinese restaurant, which he did. Every now and then something would amuse him, and he would laugh voicelessly, the only sign being a big smile and the shaking of his shoulders. He was very willing to engage with our bunch of 1st & 2nd year students. Really a humble and kind man.

  37. Scott Says:

    Stella #35: Roy Glauber may have been the last, or close to it.

  38. Raoul Ohio Says:

    NYT on Dyson:

  39. Raoul Ohio Says:

    Stella Biderman #35:

    I’ll report on another one in seven and a half years

  40. Raoul Ohio Says:

    Knuth was the student manager of the basketball team at Case while in grad school. While in high school, I played in basketball pickup games with guys on Case’s team, so it is possible that I met Knuth when I was a kid.

  41. Trees not Cubes! Memories of Boris Tsirelson | Combinatorics and more Says:

    […] Wikipidea user page is now devoted to his memory; Here is a great interview with Boris; A very nice memorial post on Freeman Dyson and Boris Tsirelson on the Shtetl […]