I Strings 2025 Conference: Insights, Criticisms, and Key Highlights

[mitchell porter]

The annual big string theory conference took place last week. I thought I would make a thread about it, partly because the most prominent post about it anywhere, is probably Peter Woit dismissing it as worthless.

I skimmed the live video (links here to slides, videos, and posters). What did I personally find to be of interest?

Lara Anderson gave a talk about topology change in heterotic vacua, i.e. transitions between different Calabi-Yau manifolds in the extra dimensions, which turn out to be mediated by the appearance of five-branes. For me there was something classical about this topic, reminiscent of the glory days of the early 1990s when all the different string theories were being unified via duality under the aegis of M-theory. In this case one is trying to go further, and understand how all the different vacua of the string landscape are connected. Brian Greene had some interest in this topic in the late 1990s, and then thanks to Susskind, the landscape was important in the 2000s, but more because that was how the anthropic principle entered the mainstream of string theory.

Apparently there is a revival of the "IKKT model", a matrix model from Japan. The best known matrix model is BFSS, which is a model of M-theory. IKKT, on the other hand, is supposed to be a model for the emergence of the time dimension. I haven't studied it at all.

There was a session on the topic of von Neumann algebras in quantum gravity; Witten spoke in this session. Hong Liu addressed a topic of AdS/CFT - why is physics local in the emergent dimensions - by introducing a "subalgebra-subregion duality", between subalgebras of the CFT on the boundary, and subregions of the emergent bulk spacetime.

During the panel discussions on the final day, Renata Kallosh said that there would soon be more data on the tensor-to-scalar ratio in the CMB, and seemed to think that a particular string model would favor Higgs inflation or Starobinsky inflation (which was the original form of inflation suggested in Russian physics literature).

 

[Ammaniya]

I didn’t watch the whole thing, but Lara Anderson's talk about topology changes sounds intriguing. It feels like there’s so much to unpack with those different Calabi-Yau manifolds. I’m a bit lost on the matrix models, though. IKKT? BFSS? Honestly, I’m not sure where to start with that. Witten’s session on von Neumann algebras seemed like a deep dive too... did I miss something there? And Kallosh's comment about new data is interesting, but I'm not sure how it ties together with everything else.

 

[mad mathematician]

Well, as far as I know M and F "theory" aren't really as of yet theories so to call.
As warren Siegel said here:
http://media.scgp.stonybrook.edu/video/2018/20181022_2_qtp.mp4

Are M, F theories even theories?
Everybody are waiting for the next Einstein to resolve the 90's and 00's conjectures...
It sometimes feel as if all this field is a bit cranky...:-)

 

[weirdoguy]

Well, as far as I know M and F "theory" aren't really as of yet theories so to call.

String "theories" in general are not theories. That is a know "problem", at least to those who focus on the words. And I do, because as a teacher I think wording is important. Laypeople can dimiss real theories (like QFT), beacuse it's "just a theory".

 

[mad mathematician]

String "theories" in general are not theories. That is a know "problem", at least to those who focus on the words. And I do, because as a teacher I think wording is important. Laypeople can dimiss real theories (like QFT), beacuse it's "just a theory".

QFT has the same inconsistencies as RQM that's for sure.
And the methods used there are dubious at best, as Feynman once said that he believes the methods employed in QED will one day be proven inconsistent mathematically.

 

[weirdoguy]

QFT has the same inconsistencies as RQM that's for sure.

For sure it has not. The very reason people don't want to teach RQM but QFT instead is that it does not have the inconsistencies RQM has. You clearly haven't studied none of them.

And the methods used there are dubious at best

So? QED is experimentally verified to an overwhelmly wide extent. And, again, physics is an experimental science. What is your point, in general? That everything we know is not true? Because that what it looks like, regarding all your postings here.

as Feynman once said that he believes the methods employed in QED will one day be proven inconsistent mathematically.

I see that you have read a lot about physics, but did you read the physics itself? Feynman died quite a long time ago, so he's not very knowledgable about modern developements of mathematics of QFT. E.g. perturbative QFT has been formalized, to some extent. We had some Insights here on PF about that. It's not pretty, nor it's useful for physicists, but it is there. What is not known, is that non-perturbative QFT in 4D exists. But physicists don't care. Experiments agree with theory on and on. And that is why we call it theory.

 

[mad mathematician]

For sure it has not. The very reason people don't want to teach RQM but QFT instead is that it does not have the inconsistencies RQM has. You clearly haven't studied none of them.



So? QED is experimentally verified to an overwhelmly wide extent. And, again, physics is an experimental science. What is your point, in general? That everything we know is not true? Because that what it looks like, regarding all your postings here.



I see that you have read a lot about physics, but did you read the physics itself? Feynman died quite a long time ago, so he's not very knowledgable about modern developements of mathematics of QFT. E.g. perturbative QFT has been formalized, to some extent. We had some Insights here on PF about that. It's not pretty, nor it's useful for physicists, but it is there. What is not known, is that non-perturbative QFT in 4D exists. But physicists don't care. Experiments agree with theory on and on. And that is why we call it theory.

pQFT has been formalized? really?
What are the axioms/postulates of pQFT?

 

[mitchell porter]

pQFT has been formalized? really?

@weirdoguy is probably referring to https://ncatlab.org/nlab/show/causal+perturbation+theory

 

[ohwilleke]

QFT has the same inconsistencies as RQM that's for sure.
And the methods used there are dubious at best, as Feynman once said that he believes the methods employed in QED will one day be proven inconsistent mathematically.

Feynman was wrong. The problems he believed might be inconsistent mathematically were, after decades of being used by physicists on a regular basis, ultimately rigorously proven to be valid mathematically.

I don't have a citation readily at hand, but read a less popularized account of this breakthrough that referenced his earlier skepticism (maybe on Science Direct) about a decade ago.

 

[timmdeeg]

After having read Brian Greene's The Elegant Universe I have been a bit enthusiastic, but unfortunately ...

 

[javisot]

The problem with creating theories that cannot be experimentally confirmed, is that they cannot be experimentally confirmed.

 

[mad mathematician]

Well when its biggest advocates wins the Fields medal (a prize in maths), and had to encourage someone to invent a prize (almost specially tailored for him and his advocates); you understand what's the problem here.

 

[pines-demon]

After having read Brian Greene's The Elegant Universe I have been a bit enthusiastic, but unfortunately ...

View attachment 367798

Yet there's still Strings '26 coming up

 

[mitchell porter]

but unfortunately ...

Do you remember what prompt you used to generate that image?

Gizmodo recently ran an article asking "whatever happened to string theory?" My comment:

"In general, the biggest thing that happened is that nothing but the Higgs boson showed up at the LHC. This all but falsified what had been the main paradigm for the future of particle physics (weak-scale supersymmetry), while providing no clear guidance for what should replace it. We were just left with the standard model (which we knew already), and no information about why quantum corrections aren't making the Higgs heavy (something that was supposed to be explained by new particles, like those of supersymmetry).

"The relevance for string theory is that string theorists tended to assume the paradigm of weak-scale supersymmetry, when searching for the real world among the googolplex possible worlds of string theory. Like the particle theorists, they are somewhat adrift without it. There are other ideas out there, but nothing has seized the imagination of the field, and there's no clear empirical phenomenon that can't be explained by just finetuning the standard model and adding some arbitrary fields to explain neutrino masses and dark matter."

 

[Fra]

I have been busy with non-physics stuff lately so i am lagging follow up. But I just stumbled on that funny tombstone.

I must say that - even if I find the foundational ansatz of string theory (ie replace point particles with strings and see what happens) is not conceptually appealing as it's explanatory value is still living in the old paradigm - the problem with string theory is not that it is hard to falsify; this is a good thing for a model; the problem is convergence rate of tuning it. Flexibility is good, but if it's too flexible it may diverge or and up require unreasonable fine tuning.

As someone that is into more agent based model paradigm, I think the problem is not that the confusion of the map and the territory as that tomb says. Because one can argue that from an agent perspective, the map is MORE real, than the territory. The territory is the black box, and the map is our best understanding. This is not a flaw, its i think an accuracte perspective. But the interactions is extremely comlpex! This is not a system dynamics, this a crazy interaction of maps, where the territory may well be emergent from "interacting maps".

In ABM paradigm vs system dynamics, what one really has is "interacting maps", if by maps one considers what is locally encoded in hte physical strucutre of the parts. Here maps are real, the territory is what is forever evolving. Trying to turn that picture around, is problematic.

The biggest problem this i have with string theory, is that the emergence of the territory without finetuning the maps in a fictice super-map-sppace, is missing and the conceptual grips of insights is also fundamentally lacking. Which is not unexpected given the ansatz, and the following "mathematical only" elaborations.

Conceptually I thinking internal and space mixing summetries must happen, in order to connect things. But wether that is in the form of SUSY or some other variant isnt' clear of course. Because the local MAP must encode both internal and external information. From this pricture all the many dualities that exists in string theory can be conceptaully understood as. recoding of the maps. It does not mean reworking territory. All these principal things are good. Except thy are cast in their explicit forms in the string lingo. But the generalisations beyond string lingo are I think good stuff.

/Fredrik

 

[timmdeeg]

Do you remember what prompt you used to generate that image?

Gizmodo recently ran an article asking "whatever happened to string theory?" My comment:

I received the photo from a friend who had it from somewhere using facetime.

Thanks for these comments, very interesting how they differ from each other.

 

[PeterDonis]

ABM paradigm vs system dynamics

Can you give references for these approaches?

 

[PeterDonis]

one can argue that from an agent perspective, the map is MORE real, than the territory.

This seems obviously false to me. Again, is this a viewpoint that appears in the literature? Or is it just your personal opinion? If it's the latter, you're pushing the boundaries of what is allowable here at PF. Even in this subforum, discussions are supposed to be based on the literature.

 

[Fra]

Can you give references for these approaches?

I refer the the general modelling categorical approaches - not specific theories. Though there aren't alot of "general theorems" of correspondences between them, perhaps because expcet complexity there is no universally fixed definition of an "agent", except it has some autonomy as opposed to constrained by global rules. The interesting part is the computational properties of the two models. There are still plenty of papers that look at examples in various fields. But here are some papers that gets the perspective.

"The emphasis on modelling the heterogeneity of agents across a population and the emergence of self-organization are two of the distinguishing features of agent-based simulation as compared to other simulation techniques such as discrete event simulation and system dynamics..."

-- Tutorial on agent-based modelling and simulation, CM Macal and MJ North​

​

"It is proven that the asymptotic time and space performance of modular imperative agent-based modeling studies is computationally optimal for a common class of problems. Here ‘optimal’ means that no other technique can solve the same problem computationally using less asymptotic time or space."

-- A theoretical formalism for analyzing agent-based models, Michael J North​

I think it is interesting from conceptually as ABM models is like a "parallell computation", with distributed subprocesses, while a normal PDE/ODE is like a big single computation. If one then associates the computational rules to "laws" one sees that the two modelling categories gives a different picture of what is primary ontologies ~ "more real". And the input and output of the single vs distributed computation is also suggestive.

The very general association to "agent" here could be - any subsystem as in Baranders unistochastics, or perhaps a single "autonomous string". That was my associion to the topic, as it seems the "background to be tuned" might be explained as constructed by other interacting strings. If this would be the case, the selection process should be a physical process, not something theorists need to fine tune. But how does it work? This is what I miss in string theory (and why i mentioned the ABM vs sytem dynamics).

/Fredrik

 

[Fra]

This seems obviously false to me. Again, is this a viewpoint that appears in the literature? Or is it just your personal opinion? If it's the latter, you're pushing the boundaries of what is allowable here at PF. Even in this subforum, discussions are supposed to be based on the literature.

Maybe it could be interpreted so that the map is always right. That wasn't what i meant.

What i mean was just the different modelling perspective, naturally suggests different starting points.
If one considers and ABM, then the agents state, that might reflect its understanding of the territory (ie envirotnmnet), is by definition a primary ontology. The agent could be WRONG, but it learns. but what it is converging towards as it learns, may be in motion, evolving ~ emergent. This is just what i meant. I think this just folllows from the paradigm change.

I don't claim one is better than the other, but you seem to end up with a computational problem that is different, and that is interesting.

/Fredrik

 

[PeterDonis]

Maybe it could be interpreted so that the map is always right.

I wasn't "interpreting" it at all. Again, here is what you said:

one can argue that from an agent perspective, the map is MORE real, than the territory.

I don't see any need to "interpret" this in order to think it's obviously false.

What i mean was just the different modelling perspective, naturally suggests different starting points.
If one considers and ABM, then the agents state, that might reflect its understanding of the territory (ie envirotnmnet), is by definition a primary ontology. The agent could be WRONG, but it learns. but what it is converging towards as it learns, may be in motion, evolving ~ emergent. This is just what i meant. I think this just folllows from the paradigm change.

I don't see how any of this relates to the statement of yours that I quoted above. Even the part that's bolded, which seems to me to be the core claim, doesn't imply that the agent's map is "MORE real" (your emphasis) than the territory it refers to.

 

[Fra]

I don't see how any of this relates to the statement of yours that I quoted above. Even the part that's bolded, which seems to me to be the core claim, doesn't imply that the agent's map is "MORE real" (your emphasis) than the territory it refers to.

Ah I see. I share your objection if if we talk about metaphysical realism! You are right of course. ie what just exists independent of any observational of inferential context.

But I never even use that definition anywhere because i don't know what it means! and even less know how to define less or more of it. For me the metaphysical realism is an empty notion, that has not utility.

For me the only notion of realism that I use is what matters for inference by a physical agent/observer, the primary ontology inside the agents "generative model" — the primary ontological variables that can actually be updated by evidence and that drive prediction and action in an contextual inference. And this is model dependent of course. There is no metaphysical "right/true" or "wrong/false" here.

In multi-agent settings, the "territory" is made up by other agents, it's internal states at microlevel(ie maps) - but at macrolevel the environment is then emergent from interacting maps. It was in this sense I consider the maps to be real. But this is not a metaphysical claim.

But to get on the topic of string theory, the analogy is that something I think the idea is still that spacetime in string theory SHOULD be emergent from many inteacting strings. The fact that it's hand-tuned, is just the perturbative approach because the full understanding is missing. This is where I made the perspective change between system dynamics and ABM.

My main input was that, even given that I am not a fan of string theory, the critique that they mix up map and territory might not be entirely fair IMO, because they might be deeply interacting in a way we apparently do not yet understand, definitely not in terms of an explicit model, and perhaps also only barely conceptually. And all the dualities in string theory I see as reflections of this. And this may be (ignorting string specifics) a deep insight, because nature itself cant tell territory from the map? I think it's just fair to defend some of this no matter that one thinks about string theory as whole. After all, few other programs are as ambitious.

/Fredrik