Gayoung Lee at Gizmodo today has responses to the question Whatever Happened to String Theory?.
Carlo Rovelli and I give the obvious and accurate answer that it’s a failed idea, explaining why. The other answers exhibit the sad state of continuing denial in many quarters of the reality of the situation:
- Science communicator Daniel Whiteson, an experimentalist who doesn’t appear to know anything about string theory, assures the public that the problem is just that we can’t yet test it.
- John Schwarz is completely delusional: “A large portion of the theoretical particle physics community is now convinced that we are on the right track to discovering the correct unified theory of our universe (and many naysayers have repented).”
- Thomas Van Riet, whose main claim to fame is vigorously hyping string theory on Twitter, also does the job here.
- Hiroshi Ooguri gives an absurd argument that one should ignore 40 years of utter failure, since “it would take 1,250 years to construct a realistic model of the universe from string theory and another 1,000 years to verify it experimentally.”
- Cumrun Vafa turns up the hype machine to 11, claiming that “the large-scale cosmological observation and the small-scale tabletop experiments are both on their way to confirming string theory predictions.” (the bogus “predictions” are from his latest “string-motivated model”).
The past few days have also seen another “string theory is doing fine” PR effort, Marika Taylor’s String theory: scientists are trying new ways to verify the idea that could unite all of physics. Here the hype is about string theory vindication from dark energy measurements, black hole gravitational wave measurements, and something about quantum computing.
It became clear long ago that writing careful explanations of exactly what was wrong with this kind of hype is a waste of time. There’s no point in arguing with the people making up the industry generating this nonsense since they simply don’t care whether what they are writing is true or not. For many years it seemed to me that the effort to try and fight this was still worthwhile, because if unchecked it was going to lead to the intellectual death of the subject of fundamental physical theory, together with massive discrediting of scientific research in general among the public who traditionally held science in high esteem. Unfortunately that fight has now been lost, with the depressing consequences that surround us all.
As for what to do about it, articles like Taylor’s claiming “right now, things are looking up for string theory” are just digging the hole that needs to be climbed out of deeper. I’m personally optimistic about some new ideas in fundamental physical theory, trying to concentrate on pushing them forward and getting them written up. I’m not optimistic though about seeing this part of the scientific enterprise return to a healthy subject in the forseeable future.
Mariks Taylor claims String Theory as a reason for tbe Langland’s Program:
“The unification of forces hence implies profound relationships between branches of mathematics. Such relationships had previously been proposed by mathematicians, particularly by Robert Langlands, and string theory gives physical explanations for the relationships.”
It might come as a surprise to many mathematicians that String Theory “explains” number theory, representation theory, and a good deal of modern mathematics.
And the string theorists keep on coming up with new justifications for why string theory is useful. For example, in her string theory puff piece, Marika Taylor says
I can find no evidence whatsoever for this.
What would happen to the string theory folk if the cosmologists actually accepted data (and the processing of it) that flipped the sign of the cosmological constant? Sabine recently mentioned a paper (in a reputable journal, but as far as I know unknown people) that points in that direction.
I’m an experimentalist (chemical physics however) and have always had grave doubts of the purported small systematic errors of cosmology. The CC has always looked terrible in that regard, for the always obvious (see paper) reason.
I assume the ST people would go even wilder than they are now.
Interesting that the experimentalist doesn’t value the distinction between scientific and non-scientific knowledge.
Dear Peter, you mention getting some new ideas written up and hopefully you are referring here to your own work on twistors, as your audience may be growing restless (2 years on from your last paper). One cant rush good science/math of course, but a substantive update would be welcome (have you hit a snag, which would be sad but still interesting).
Time to throw your readers a bone methinks:)
Oh wait, I had not yet read the Ooguri remark, now THAT is laughable!
“Unifying quantum mechanics with “general” relativity is 25 times more difficult than unifying it with “special” relativity. […] It took physicists 50 years to develop a concrete model based on quantum field theory to describe particle physics phenomena (the Standard Model) and another 40 years to confirm all of its predictions (the Higgs boson discovery).
Therefore, one could argue that it would take 1,250 years to construct a realistic model of the universe from string theory and another 1,000 years to verify it experimentally. Now, I don’t think it would really take that long, but this illustrates how difficult the problem is. Therefore, I would expect achieving such a unification to produce 25 times more rewarding results and have 25 times more impact on a broad range of physics and mathematics.”
Doug McDonald,
If the CC had the opposite sign, this would be heavily promoted as conclusive evidence for string theory, since AdS/CFT shows you need the opposite sign.
tripitaka,
Definitely haven’t hit a snag, quite the opposite. The more I pursue these ideas the more convincing they are, and I slowly have been better and better understanding what is going on. I’ve off and on tried to understand “Wick rotation” in terms of hyperfunctions of the time variable, but couldn’t get this to work until recently. I’ve written up the details of how this works in the simplest case, now trying to understand details of what happens for spinors and twistors. Hopefully in the next month of so I’ll be able to finish a now half-complete 50 page or so paper about all of this.
Luckily for this project, the situation at Columbia has stabilized: there’s nothing new going on, so one just needs to get used to the new depressing realities here. I’ve had a family issue to deal with that took a fair amount of time, but now that has worked out very successfully, easier to concentrate on Wick rotation.
I should comment that besides the larger awfulness, it has been depressing the past few years to realize how difficult it is to get anyone interested in the new ideas that I see as so promising. Very few people understand much about the geometry of spinors, the fact that these are the fundamental objects of spacetime geometry is not enough to get people interested. That something so fundamental as “Wick rotation” remains so poorly understood, while the community of theorists is focused mostly on utter nonsense is discouraging.
Anyway, at least it’s very clear what I should be doing (yes, this posting is not it…).
While I agree that “tests” of String Theory are often overhyped, that does not make it a “failed idea” In particular, the common claim that string theory is ruled out by the observed positive cosmological constant (Λ > 0) is unfounded.
First, string theory contains large classes of fully explicit, moduli-stabilized vacua with positive, zero, or negative cosmological constant once supersymmetry is broken. Classical massive type IIA flux vacua (DGKT and descendants), KKLT, the Large Volume Scenario, and many others all yield metastable de Sitter solutions in four dimensions. Yes there are objections to these, but the subject is far from settled overall.
Second, the cosmological constant problem is a problem of the Standard Model coupled to gravity, not just of string theory. The Higgs potential alone contributes a negative vacuum energy roughly 55–60 orders of magnitude larger than observed, whereas a purely fermionic (condensate-driven) mechanism for electroweak symmetry breaking would contribute a huge positive vacuum energy of similar magnitude. The observed tiny positive Λ therefore requires an exquisite cancellation among all contributions (QCD, electroweak, higher-scale physics, gravitational sectors, fluxes, etc.). Until we understand the origin of this cancellation, it is inconsistent to single out string theory as uniquely failing on the cosmological constant.
Finally, the fact that AdS/CFT is the best-understood, exactly solvable corner of string theory—and happens to live in Anti-de Sitter space—does not imply that our universe must have Λ < 0. AdS/CFT is a tool and a particular (highly supersymmetric) solution; the rest of the string landscape contains perfectly valid de Sitter, Minkowski, and AdS vacua that are simply harder to study with the same non-perturbative control.
Philip Gibbs,
I certainly wasn’t claiming that string theory is ruled out by positive CC. Like everything else, it says nothing about the CC. You can come up with “string models” with any CC you want, of any sign or any magnitude. Or, like Vafa, you can find models with varying CC.
The problem with string theory is that it’s a failed idea that explains nothing, predicts nothing, so of course can’t be tested. The problem with string theorists is that they refuse to acknowledge this and instead produce endless bullshit about how to “test string theory”.
“I should comment that besides the larger awfulness, it has been depressing the past few years to realize how difficult it is to get anyone interested in the new ideas that I see as so promising. Very few people understand much about the geometry of spinors, the fact that these are the fundamental objects of spacetime geometry is not enough to get people interested. That something so fundamental as “Wick rotation” remains so poorly understood, while the community of theorists is focused mostly on utter nonsense is discouraging.”
Having been a postdoc a few years, welcome to the sociology of academia. Having read through Lawson and Michelsohn’s Spin geometry textbook in a few months when I was a grad student, I don’t find spinors as geometrical objects to be inherently that difficult. The real question might be whether people are willing to put in effort to pursue something like this out of the norm.
While many people do resist it, the optimal path for an early career academic seems to be to study and work on what the elders/fancy people/journal editors find to be the most promising and important. The process doesn’t need to be as direct as some person saying “we should all work on this”- it can take more subtle forms. For instance the elders organize conferences and invite speakers who work on topics they approve of. The unsuspecting audience will hear nicely presented talks and find the subject area interesting and important (after all I heard it at this fancy and important conference). What typically happens is the more often you hear an idea repeated, the more likely you will think of it as important and interesting – this is inherently human process and nothing sinister about this, but when all the elders only organize conferences on topics they approve of, the effect is a megaphone for the ideas the elders are trying to promote.
This is all sales, and we are no different from car salesmen.
@Peter Shor,
not that I want to toot string theory’s horn here, but I think there is a fine distinction between claiming string theory itself has somehow explained “how quantum systems can be used in computing”, and people realising that some mathematics with provenance in string theory was actually instead useful for modelling quantum computing, because it was really just something in a quantum field theory being originally applied (misapplied?) in a string-theoretic setting.
@DavidRoberts:
Can you point me to some specific mathematics with provenance in string theory that has actually been useful for modeling quantum computing? Or was your comment just pure speculation?
@Peter Shor
I am not an expert, and I’m certainly happy to be corrected, but for instance this (draft) article and its predecessors claim to make conceptual progress on understanding the fractional quantum anomalous Hall effect and as (apparently) considered as candidates for future/hypothetical anyon-based quantum computers. The mathematical ideas have their origin in M-theory, but there’s no branes, strings or whatnot in these papers that are intentionally about trying to understand QC-relevant quantum systems.
I very strongly suspect there are people out there who try to link the string-theoretic origins of mathematical ideas and quantum computation much more closely, to the point of the silliness like the “black hole in a chip” publicity stunt. But to my reading the above linked paper is getting away from that (or very nearly so – the language inherited from the previous home of the maths is telling).
@Peter Shor
Not quantum computing per se, but explicit entanglement entropy calculations via 2D CFT (Holzhey, Larsen, Wilczek; Cardy+Calabrese). These have been highly influential in thinking about quantum chaos, entanglement transitions, and the latter have some close connections to error correction thresholds.
But it’s really just 2D CFT, not string theory per se. So one needs an “originalist” interpretation to make the attribution (Polyakov trying to solve the 4D string).
There are also entanglement analyses of random-geometry tensor networks that connect back to matrix models and Liouville quantum gravity with matter.
QFT — Quanta Mag on renewed interest in “Quadratic Gravity”:
https://www.quantamagazine.org/old-ghost-theory-of-quantum-gravity-makes-a-comeback-20251117/
Any thoughts?
@Matthew Foster @Peter Shor
I think the closet connection one can draw between quantum computing and string theory is the idea of topological quantum computation, which heavily uses the machinery of 3d TQFT and also 2d CFT. Admittedly, the remarkable progress made in the 80’s on 2d CFT was partly associated to its application in string theory.
Meng Cheng, Matthew Foster, David Roberts,
None of what is mentioned provides any justification for bullshit like “string theory has proven to be useful in understanding how quantum systems can be used in computing.”
This whole subject suffers from 40 years of dishonest hype via obfuscation, and that has a lot to do with why it has collapsed intellectually. At this point I have no patience for the people who are doing this kind of thing. Writing obviously untrue things for the public about science should be way out of bounds for scientists. People should be condemning this and trying to stop it, not trying to think up justifications for it.
Raoul Ohio,
I don’t want to start a discussion of this technical subject here, it’s interesting but way off topic. Unfortunately I don’t have time to look at the new work in this area that supposedly justifies the Quanta article. My impression was that nothing had changed in the situation of this program, but maybe I’m wrong.
As the article points out, the idea goes way, way back, to the late 70s I think. It has always been an intriguing part of the story of trying to quantize gravity much like Yang-Mills theory. My own general feeling about this kind of idea is that it’s way more promising than string theory, but if you are doing only pure gravity, you’re in danger of ending up like lots of “quantum gravity” research programs: a bunch of claims that some calculations make sense, with no way to connect to anything measurable and the whole thing only partially defined so the question of consistency is still open.
But, sorry, at the moment I don’t want to moderate yet another discussion of the state of the vast subject of “quantum gravity”.
@Peter Woit
Fair enough re:BS hype. In some sense, though, high-energy physics has always been sold to the public through rather indirect spin-offs, rather than the main intellectual achievements of the SM. E.g. ARPAnet. The mathematical spinoffs of the field pre-holography are significant enough that one can argue much of Kitaev’s work (and hence modern cond-mat) is built on it. But yes it’s been 25 years since AdS/CFT, and certaintly it seems inappropriate to still be hyping the main intellectual thrusts of string theory (though that was also true as soon as the landscape problem was recognized, which was a very long time ago…)
@Peter Woit
Re: justifying claims like
I wholly agree. That is why I said there’s a distinction between this spin, and something that has a (reasonable, imho) chance of being true, namely the mathematics got re-used. The latter is interesting, but not sexy (and even paints the people who developed the mathematics while doing string theory as slightly misguided: they were applying the maths to a speculative program, rather than actual concrete and maybe useful physics!), whereas the former paints string theory as the hero that suddenly has applications.
One could be generous and imagine a science journalist who isn’t too much of a skeptic phrasing things in the way that was quoted, as a kind of summary of a discussion about re-use of mathematical physics ideas. But that the writer has, I see, a PhD on M-theory means this rationale isn’t available. I don’t want to justify the kind of spin that makes it look like string theory, as a theory of quantum 1-dim objects giving rise to gravity, “explains” quantum computing.
I’d love if people explained it like it is, namely how the mathematics is interesting and finds applications, even if and when the original claimed physics origins turn out to be wrong. Knot theory is interesting even if Kelvin’s idea about them was bunk. And if the theory of braids and knots turns out to be useful in quantum computing (braided anyons and whatnot), no one is going to say “Kelvin’s aether-knot model of atoms explains quantum computing!!!1!”. But neither are they going to pooh-pooh the reality of the complex history of the ideas, while pointing out Kelvin was just wrong.
Let me first defend Daniel Whiteson as he is a colleague, and his stance is not at all rare among experimentalists. All he says is that someday string theory (or some descendant) may actually make a prediction (and quite possibly be *experimentally* falsified). In the meantime if string-inspired models give us new signatures to search for experimentally, all the better. I don’t think you can lump him in with the other string theory defenders.
What I found more interesting was this by Thomas Van Riet:
People say that without experiment we cannot call one theory better than another. That is plain wrong. There are many consistency checks, which are ridiculously hard to pass. Can you compute black hole entropy? String theorists were able to compute it in very idealized circumstances and reproduced Hawking’s famous formula for black hole entropy!
Is this a significant thing? Or is string theory such a vast expanse of math that any result could have been pulled out of it including black hole entropy?
Amit,
What you change Whiteson’s quote to makes sense, but what he writes is nonsense. The problem with “The major objection is that its predictions are for things at a microscopic scale that we cannot yet test” is that the real problem is there are no predictions. The comments he makes about group theory are just as bad. Group theory is not some “mathematical curiosity” that only recently came to be of interest in physics. It is a subject that has been at the very center of much of 20th century mathematics and closely related to QM from the beginning (Hermann Weyl’s book on group theory and quantum mechanics was published in 1928, and the use of group theory in analyzing atomic spectra was central right from the beginning.)
Scientists writing for the public should take care to write about what they understand, not repeat hype they don’t.
The black hole entropy thing is a standard bit of hype and obfuscation. It’s a complicated technical subject, so a good thing to hype. To lay out carefully and clearly exactly what the string theory calculations involved are and would take a huge effort and almost no one except experts would understand anyway. A couple simple things to say that show that the situation is not so simple as the hype are that the advertised calculations are in a different dimension, aren’t really black holes. The Hawking computation is semiclassical, reproducing that is not so impressive, the argument is more complicated.
Thanks Peter. I figured the black hole entropy story was more complicated than a simple “we are able to do this”. Daniel Whiteson has done a lot more outreach than I have and his comments may have been informed by that. My efforts are mostly with terrifyingly bright New Jersey high school students, and there is a fascination with compositeness: the idea that things you can see are made up of smaller things that you cannot. There is this narrative thread from Mendeleev to Rutherford to Chadwick to Friedman-Kendall-Taylor. String theory has filled the hole after quarks. The pushback against string theory has to reclaim the concept of substructure as being a separate thing.
There is a hilarious sentence in Marika Taylor’s article
“However, the Desi results suggest that interpreting dark energy as quantum energy of strings may be on the right track.”
which is linked to the DESI paper on fits of extended (that is, beyond-LCDM) cosmological models to data. That paper has absolutely nothing to do with “quantum energy of strings”; the models studied are mostly phenomenological vanilla modified gravity, as well as various parameterizations of dark energy’s evolution with time. It is not clear that quantum energy of strings, if such a thing exists, has predictions for the expansion of the universe.
Amitabh:
“People say that without experiment we cannot call one theory better than another. That is plain wrong. There are many consistency checks, which are ridiculously hard to pass.”
Beyond a certain not-very-high level of intellectual seriousness, it is obviously true (and accepted) that there are many such consistency checks — highly non-trivial calculations with highly non-trivial and non-obvious answers — that have been done in string theory and have not been done in any of the supposed rival theories.
Examples are mirror symmetry computations or those involving holography and the computation of multi-loop effects in gauge theory through gravity-side computations. These are better than black hole entropy, because it might be possible to fudge your way to a factor of 4 pi when you know that is what you need, it is a little harder to fudge your way to a number like 295091050570845659250 (to pick a random example counting rational curves in one of the early mirror symmetry papers), where two apparently different techniques give the same answer.
Joseph Conlon,
The mirror symmetry calculations are a good example of the problem. They provide strong evidence for a very non-trivial duality. The connection to string theory though is to topological string theory, which is of great interest mathematically, but something very different than a string theory which is supposed to describe the physical world. The problem with string theory arises when you try and use it to get a unified theory with four more or less flat spacetime dimensions, and make the others go away in some fully consistent way. This is where one is dealing with a theoretical setup that is not completely well defined.
Topological string theory is a subsector of the full Type II string theories. The concistency of this subsector does provide highly non-trivial evidence for the consistency of the complete theory, of which it is a truncation.
Peter,
You say:
If you’re a researcher, it’s hard to learn a whole new subject, and people aren’t going to commit to learning stuff until they’re convinced that it’s promising.
If it becomes absolutely clear that your ideas are promising, and people still aren’t interested in them, then it’s time to get depressed.
Cobi,
The “truncation” removes all the physical states… Analyzing what’s left can’t provide “highly non-trivial evidence” that what you threw out is a consistent model of physics.
Peter Shor,
I understand the problem well. For one thing, it’s taken me a very long time to fully understand what’s going on behind the ideas about spinors and twistors that I’m trying to use. That few others want to do the same without some guarantee it will be worth the trouble is not surprising.
Partly my own fault also for not getting more of this completely worked out and written down. I remain convinced there’s a set of very promising ideas here and I’m enjoying pursuing them, slowly making progress. But I have a lot more to do. I can understand why others don’t see the promise in these particular ideas, although find it hard to understand why at the same time they find string theory unification promising…
Dear Peter,
personally I do follow with great interest the developments of Your ideas and their possible connections with material in non-commutative geometry (where I try to work).
As regards the scarce attention to new ideas, compared to the always lively trends in very popular approaches (irrespective of their little or nonexistent merits), I see that the discussions are often attempting a “rational analysis” of the respective strengths of the theories involved. This is actually irrelevant. This is an issue of “sociology” … or I should rather say “ethology”. As in most of the simians primate social groups, there is a very rigid pecking order of individuals and the ranking is the only real game: people work on certain approaches because it is there that the “smartest top achievers” are working and it is there that (apart from extremely rare exceptions) the selection of the new leadership is going to take place. It is somehow given for granted that top-level researchers will anyway obtain “something” (better than those lower-ranking), no matter where they look. In the light of such simple considerations, one can easily understand quite arrogant (and somehow irritating) statements like these (that I literally quote from Jay Armas book “Conversations on Quantum Gravity”):
“. . . history has shown that other consistent ideas and techniques that a priori appear not to be part of string theory, if shown to be consistent, will later become subsets of what we will call string theory in the future. In other words, string theory swallows other fields in some sense. My belief is that string theory is wide enough, or can be made general enough, to include everything that is mathematically consistent.”
C. Vafa (in J.Armas 2021)
“String theory is the only idea about quantum gravity with any substance. One sign is that where critics have had interesting ideas (non-commutative geometry, black hole entropy, twistor theory) they have tended to be absorbed as part of string theory.”
E.Witten (in J.Armas 2021)
It is somehow a very pleasant “contrappasso” that, in the last few years, some of these top-level researchers (including E.Witten) have been returning to “good old and forgotten” algebraic quantum field theory … although they might call it now “AdS/CFT flavored string theory”
😀 😀
All the Best.
Paolo
I found Paolo’s comment quite insightful. When I was an undergrad more than 20 years ago, I felt that the smartest physicists worked on string theory. Why I felt this is probably a mixture of the influences of The Elegant Universe (the book and the PBS show) and a certain kind of arrogance that takes root in physics and math undergraduates where they feel they do and should do deeper, harder, more meaningful things than everyone else. There was also a cult of personality around people like Ed Witten who himself, in person, has always been kind and generous to students in my experience. After all, who doesn’t want to work on things that the next Einstein is working on? Like Paolo’s says, maybe this drove a lot of smart students into string theory? I certainly never even explored alternate theories of quantum gravity even as a graduate student (who eventually switched away from string theory).
I see similar effects in machine learning (ML). While people certainly flock to ML because of the dramatic success of LLMs/multimodal models (with caveats), it also attracts people seeking to participate in the gold rush not just for the money but also for the prestige and influence and relative lack of risk compared to following a completely new paradigm. This is mostly understandable because, just like with string theory in the 80s and 90s, modern AI does have some striking successes even though there are also striking problems that don’t seem straightforward to solve.
Working alone on a problem or project can create a sense of frustration especially when it seems like others are making rapid progress or at least pushing out work at a much faster pace. I feel this is mostly illusory and Peter, if you can, take your sweet time and explore and share your work when it feels right. The most important thing is that you are enjoying yourself.
Peter,
Armas’ interview book is fascinating. Most people interviewed are string theorists. The book is a large collection of smart people who are completely lost. They are looking for something that they cannot describe with precision, they find results that are not strict, they are lost in technicalities, and they all care too much about what others are doing.
And there is an elephant in the room. Everybody avoids making predictions. They all agree that their own theories do not make predictions, but worse, they actively avoid doing so. They are not working towards predictions. All of them. It is fascinating to see so many people claiming to be physicists and not caring about predictions and observations.
The interviewees lack concrete statements, concrete results, or concrete guidance: they are incredibly vague. They talk like a mixture of pied pipers and children following them. And these are the smartest people in theoretical physics across the world!
The book is a monument to the contrast between wishes and actual achievements. It is really worth a read.
Joka,
I wrote more positively about that book here:
https://www.math.columbia.edu/~woit/wordpress/?p=12458
although I agree with your description of it.
The people doing this work would very much like to make predictions. Their problem is that they’re pursuing ideas which long ago failed, in particular because the only predictions they can make don’t look like the real world. A small number of people (e.g. Vafa and collaborators) are still claiming to make “predictions”, but they now have long histories of making predictions that were falsified, and their current claims are not plausible.
Peter,
This is just a suggestion, possibly for a later time. For non-specialists could you provide some background for Witten’s statement that “String theory is the only idea about quantum gravity with any substance”. According to ChatGPT: “1974 (Scherk & Schwarz): Recognized that the spin-2 excitation of the closed bosonic string corresponds to the graviton. Gravity is not added by hand—it comes out of the theory automatically, something no quantum field theory had succeeded at with consistency.” I gather that it is this early success of string theory that has been the main impetus for the last 50 years of focus on it. Could you discuss how extraordinary this result is and why, in your opinion, it is not enough by itself to suggest that string theory is the best approach?
David,
The story of string theory, exactly what it is, why some people were enthusiastic about getting quantum gravity out of it, and why this doesn’t work, is a very complicated one. I spent a couple years writing a book about it. If you want to understand what is going on, you could try reading that. Although the book was written nearly 25 years ago, little has changed.
One way you’re never going to understand anything is by consulting ChatGPT and getting pointed to worthless hype.
Witten is a genius and has a lot of insightful things to say and in such cases it can be quite worthwhile to discuss these. On rare occasions he, like everyone else, says something dumb that should be ignored. What you quote is such a case.
Peter,
I’m a tenure-track string theorist at a prestigious institution. I agree that outreach hype is often careless. But your blanket conclusion here — “failed idea,” “explains nothing,” “predicts nothing,” “not testable” — is technically indefensible once you engage what string unification has actually accomplished since ~2000.
1. Nonperturbative definition (in a large class of spacetimes).
For asymptotically AdS boundary conditions, AdS/CFT is a nonperturbative definition of quantum gravity in terms of an ordinary CFT. Bulk black holes map to thermal states. Hawking–Page is a confinement/deconfinement transition. Bulk reconstruction is phrased in operator algebra language. Entanglement wedge reconstruction is naturally quantum error correction. These are not metaphors. They are exact statements in controlled examples.
2. Quantum information in gravity is not a “confirmation of quantum computing.”
The point is structural. Ryu–Takayanagi (and covariant generalizations) computes boundary entanglement entropies from bulk extremal areas. QES/islands compute Page curves in solvable models. This addresses the semiclassical information paradox inside a controlled holographic framework. You can dislike the hype without pretending the technical content is vacuous.
3. “Compactify six dimensions” is not the state of the art.
Flux compactifications gave a systematic 4D EFT with controlled ingredients: quantized fluxes, tadpole constraints, warping, and supersymmetric vacua with stabilized complex structure moduli and dilaton (GKP). KKLT and LVS are explicit moduli stabilization mechanisms using nonperturbative superpotentials and α′ corrections. The debate about de Sitter control exists because the construction problem is tight, not because “anything goes.”
4. F-theory made nonperturbative 7-brane gauge sectors engineering-friendly.
Varying axio-dilaton is geometric. Non-Higgsable clusters are generic. Matter and Yukawas localize on codimension-two/three loci. Anomaly cancellation and flux quantization constrain chirality. This is concrete 10D/8D physics feeding into 4D EFT data.
5. Topological strings are not “unphysical therefore irrelevant.”
They compute protected couplings in the physical theory. The A-model computes Gromov–Witten invariants. Those reorganize into integer BPS counts (Gopakumar–Vafa). The topological vertex computes all-genus amplitudes on broad classes of toric CY. OSV relates the topological string partition function to indexed BPS black hole degeneracies in \mathcal N=2 compactifications. These are physical quantities, protected by supersymmetry, and cross-checked across duality frames.
6. Exact gauge theory and duality technology is now precise, not aspirational.
Nekrasov partition functions compute nonperturbative \mathcal N=2 data by localization. Pestun gives exact S^4 partition functions and Wilson loops. Class S realizes large families of 4D SCFTs from the 6D (2,0) theory on punctured Riemann surfaces. AGT relates these to 2D Liouville/Toda CFT. These are exact equivalences of computable observables.
7. Gauge/gravity unification shows up directly in amplitudes.
Twistor-string ideas reorganized perturbation theory in \mathcal N=4 SYM. Color–kinematics duality exists. Gravity amplitudes can be obtained as a double copy of gauge theory numerators. This is an explicit algebraic relationship between gauge and gravity S-matrices.
None of this selects the Standard Model vacuum. None of this is a guaranteed near-term collider prediction. Fine. But that is not what your rhetoric claims. You claim “explains nothing.” It explains how quantum gravity can be defined and computed in controlled settings, how gauge dynamics reorganizes into higher-dimensional gravity, how black hole entropy and information accounting work in solvable models, and how consistency constraints corral low-energy EFTs.
If you want to criticize hype, do that. If you want to claim the program produced “nothing” for 25 years, you need to grapple with the above and explain which parts you think are wrong technically, not sociologically.
“Stringmaster”,
I’m guessing you’ve spent a half-hour or so generating that stream of irrelevant hype (“F-theory made nonperturbative 7-brane gauge sectors engineering-friendly”????). To seriously discuss any of your many points would require vast amounts of time and effort. There was a time in my life when I tried to do that sort of thing, but in the end I found this was a waste of time. People going on the way you do have no interest in what’s true and what’s not true.
At this point I don’t even have a good reason to believe you’re not a juvenile troll (“Stringmaster”?) generating this with ChatGPT. You claim to be a “tenure-track string theorist at a prestigious institution”, with email “notyourbusiness@physics.harvard.edu”. If that’s really what you are, why don’t you do something about the embarrassing hype coming out of your institution instead of anonymously attacking me for publicly saying that’s what it is.
Peter, you’re dodging the substance and substituting social diagnosis (“irrelevant hype,” “waste of time,” “no interest in truth”) plus an identity purity test, and that’s not an argument. “Irrelevant hype” is itself a technical claim: pick one point and say what is false. If you think “F-theory made nonperturbative 7-brane gauge sectors engineering-friendly” is hype, then tell readers what is wrong with it—do you deny that F-theory geometrizes varying \tau and provides a systematic handle on IIB 7-brane gauge sectors, or do you accept that but decide it “doesn’t count”? If it’s the latter, your thesis is not “string theory is wrong,” it’s “I refuse to count internal consistency, controlled constructions, and exact results as explanatory progress,” which is a philosophical criterion, not a refutation. “To seriously discuss would take vast time” is also not a rebuttal; it is a refusal while keeping a sweeping verdict (“failed idea,” “explains nothing,” “predicts nothing”) on the table. Sweeping claims require engagement with counterexamples; you don’t get to wave them away and still expect the verdict to stand.
You then relocate the debate from facts to motives by asserting that people “going on the way you do have no interest in what’s true.” That move is convenient because it immunizes you from checking whether your own public summary is technically accurate. If your position is that holography as a nonperturbative definition in AdS, RT/QES as computable entanglement observables, localization as exact nonperturbative gauge theory data, and flux compactifications/moduli stabilization as controlled EFT technology are either false or irrelevant, then say so explicitly and defend it. Insinuating I’m a troll or “ChatGPT” is similarly beside the point: the correctness of statements like “gravity amplitudes can be obtained as a double copy of gauge theory numerators” does not depend on my email address. If those statements are wrong, show where; if they’re right but you don’t value them, then stop saying “explains nothing” and say what you actually mean, namely “I only count unique near-term experimentally testable predictions for our vacuum.” Finally, on your rhetorical pivot—“why don’t you do something about embarrassing hype”—I already agreed with you that hype exists and should be criticized; what I’m objecting to is your follow-on inference that the underlying technical program is empty. You can condemn hype without making false universal claims about the last 25 years of technical progress, and right now your response reads as evasion: “I can’t be bothered; and maybe you’re fake,” rather than an engagement with whether your “explains nothing” line survives contact with the actual technical record.
Peter, I took your advice and reread (20 years later) your book. I had completely forgotten what a wonderful, fascinating discussion it was not only of the problems of string theory, but also a terrific review of modern high energy physics and math at a level understandable to what I suspect are your typical audience. I would suggest that readers of your blog should go back to your book. I suspect that they will find it, as I did, unputdownable, a page turner! Not to mention that it is remarkably prescient – 20 years later.
“Stringmaster”,
You’re giving me zero evidence you’re not a troll. Explain to me why I should spend a lot of time taking you seriously and trying to have a serious discussion of complicated issues with you. For a start, put your name on what you have to say publicly and I’ll decide if it’s worth spending time on this.
Peter, the unimpeachable reason for anonymity is simple: in 2025, attaching a real-world identity to pointed criticism of a public figure online is a nontrivial professional and personal risk, and that risk is wildly asymmetric for anyone junior-ish (tenure-track, soft-money, visa-dependent, grant-dependent, or just employed in a small subfield). “Put your name on it” is not a neutral request; it is a demand to accept reputational and institutional blowback in exchange for the privilege of being heard. That’s backwards. Arguments about AdS/CFT, flux vacua, localization, BCJ double copy, etc., are either correct or incorrect independent of whether the commenter’s CV satisfies your priors. If you want to filter for seriousness, you already have a better tool than identity: engage one concrete technical statement and test it.
Also, you’re implicitly rewriting the norms that made the early “golden age” of science blogs work. Back then, a lot of the best technical back-and-forth happened under pseudonyms or partial anonymity, precisely because people could speak candidly without dragging departments, students, hiring committees, donors, and internal politics into a comment thread. That wasn’t “trolling”; it was a workable expectation of privacy for technical discussion in public. The requirement “real name or I won’t engage” doesn’t raise the epistemic bar; it raises the retaliation bar.
And, honestly, your troll framing is self-serving. If someone wanted to troll, why on earth would they choose a moderately well-known blogger in a niche technical subject, arguing about protected sectors and 7-branes? Trolls optimize for attention and outrage, not for discussions that require knowing what a tadpole constraint is. The more plausible explanation for anonymity is not “sockpuppet,” it’s “rational risk management”: the commenter doesn’t want their employer, colleagues, students, or tenure file to become collateral damage in a public argument, and they don’t want your readers to treat “X at Y institution” as the real target rather than the claims.
If you want an honest seriousness test, here’s one that doesn’t involve doxxing: pick one sentence and say “this is wrong because…” with specifics. If the commenter can respond with technically correct details, appropriate caveats, and references, they’re not wasting your time. If they can’t, you stop. That discriminates signal from noise far better than demanding names.
“Stringmaster”
I accumulated a huge amount of experience about dealing with comments during the “golden age” of science blogs. What I learned back then was that people who wanted to have a serious discussion didn’t submit comments under stupid names like “Stringmaster”. If they wanted to have a serious technical discussion they made serious technical arguments, not stupid propaganda. As for my testing you by trying to have a technical argument with you, sorry, but we’re done. Responding to serious criticism of string theory with hype about F-theory constructions gives me all I need to know.
Peter, I hope you will allow me to play the role of Stringmaster’s advocate:
1) It’s far from unreasonable for a tenure-track professor to be cautious about attaching their name to a public debate that may blow up.
2) Stringmaster’s anonymity is comparable to that of several other people leaving comments on here (there’s no policy against it).
3) Anonymity can be pernicious, but Stringmaster’s original post is not ad hominem. You may disagree with the points raised, but these were technical issues (a bit on the laconic side), not personal attacks.
4) This isn’t a job search committee/grant evaluation panel, where one needs a full application package in order to make up one’s mind.
5) Having had several people try to credentials-shame you through the years, I’m sure you will agree that Stringmaster’s publication record is only tangentially relevant here.
All of the above may be true and you can still choose not to engage in lengthy debates if you think they’re likely going to end up being unproductive. This is your blog, after all. But talk of trolls sidestepped the issues Stringmaster raised (which, as even Stringmaster admitted, do not select the real-world vacuum or make foreseeable-future predictions).
Alex Gezerlis,
This has nothing to do with his credentials. The point is just that if he really is what he’s claiming to be, a professional who wants to take a lot of my time to have a serious discussion, he should behave like a professional. Posting a bunch of hype anonymously under a juvenile pseudonym is not professional behavior.
Yes, when you put your name on what you write publicly, you put your reputation on the line. This means you’re going to take the thing seriously and put care into what you write.
On top of this, the trolling thing is an increasingly large problem. I’ve been taken in by it a couple times, and this looks to me more likely trolling than what it purports to be. I’m increasingly unwilling to spend time engaging with someone’s arguments unless I have good reason to believe they’re being made in good faith.
All,
The more I thought about it, the more sure I was that “Stringmaster” is a troll, and this was just confirmed. The ChatGPT prompt used to generate the troll comment was:
“You are an arrogant, condescending nerd. You are a string theorist, and you like to brag about having a tenure-track job doing string theory at a top university. In character, write a comment on Peter Woit’s blog post here:
https://www.math.columbia.edu/~woit/wordpress/?p=15348
Explain why Peter is completely wrong, with a thorough, extremely detailed explanation of the great successes of string unification since the late 90s — 2000s (especially talk about technical stuff not covered by the popular media), for example work on holographic dualities, gauge/gravity dualities, moduli stabilization, F-theory, construction of stable vacua for various Calabi-Yaus, black hole entropy, etc. You should sound like someone who has **heard** all these technical terms but has no actual deep technical knowledge of any of them, and is using them to brag and overwhelm Peter with your superior technical knowledge rather than have a serious discussion.
Don’t go too over-the-top: this must be a plausible-sounding comment from an arrogant Postdoc, not a troll comment.”
How were you able to determine the ChatGPT prompt from Stringmaster’s comment?
Bob,
I heard from the troll, who provided the prompt.
I wonder why nobody has taken up the following “challange” which I am about to propose. If you think its a silly challenge, explain why. Here is the challange to the siring people, or, more generally to the BSM or TOE people:
Come up with a theory of everything, including gravity, that is self consistent and reduces at observed energies and scales to our actual universe. It has to numerically
come within a couple of percent of all the parameters of our known universe.
You are allowed a certain number of free parameters, fine tuned as much as you need or want to. The numbers are as follows:
First the number used in the current ordinary standard model of particle physics, plus the ones needed to describe the right-handed neutrino sector (energies can be anywhere up to the Planck scale.) You are allow SUSY or not, strings or not, as you wish, at any energy scale (that scale is a parameter) You are allowed a few more physical particles, scalar gravitons, gravitinos, axions, whatever you wish, that could in fact be reasonably measured with a Planck Scale experiment. “A few” is say, 40, with two parameters for each particle. You are allowed in infinite number of other particles or degrees of freedom, so long as they come each with parameters which come from well described random distributions, all described with not more than say a total of 50 parameters. Saying a distribution is a Gaussian or a Poisson or whatever is one parameter.
That sounds reasonable to me as a challange. What most certainly surprises me is that apparently no String or TOE proponent has published such a thing.
If the String Theory proponents can’t do that, I would feel a lot more comfortable knocking them without quibble. At least it would be a well-defined goal. And I’m allowing a total number of parameters that is infinite.
Stringmaster, what’s wrong with my challange?
Doug McDonald,
You seem to have missed the explanation above that “stringmaster” is not a string theorist but a ChatGPT wielding troll.
In any case, the answer to your challenge is that no one, string theorist or otherwise has “a theory of everything, including gravity, that is self consistent and reduces at observed energies and scales to our actual universe.” What string theorists (and others) do have is basically a conjecture that a self-consistent theory with certain properties exists, but no actual well-defined theory. I wrote about this here
https://inference-review.com/letter/theorists-without-a-theory