Gia Dvali

Zierler:

OK, this is David Zierler, Oral Historian for the American Institute of Physics. It is May 28, 2021. I am delighted to be here with Professor Giorgi Dvali. Gia, it's great to see you. Thank you for joining me.

Dvali:

Great to see you, David. Thank you.

Zierler:

First things first, professionally, everybody knows you by Gia. When did you get that nickname?

Dvali:

From my childhood. Actually, that's my real name. Georgi is my official name.

Zierler:

Is Gia a common nickname for Giorgi?

Dvali:

It is, yeah. I think it's pretty common. Not every Giorgi is nicknamed as Gia, but in my case, yeah. But it was used in the way that it sort of became a nickname. So it's like a real name. I sign my papers as Gia.

Zierler:

It's Gia for everybody. On a more official note, tell me your current titles and institutional affiliations. And you'll notice that I pluralize that because I know you have more than one.

Dvali:

Yeah, actually I have two. So I'm a Professor at the Ludwig Maximilians University, LMU, in Munich. And I'm also a Director at the Max Planck Institute for Physics, also in Munich.

Zierler:

But no longer with NYU. Your affiliation with NYU has ended?

Dvali:

Yeah, it ended in 2019.

Zierler:

Just for a snapshot in time, what are you working on these days? What's interesting to you?

Dvali:

There are several things that I'm interested in, and they are all interrelated. But basically, what I'm trying to understand is the physics of black holes, but from a slightly different perspective. So I'm trying to understand this phenomena. Black holes have these seemingly mysterious properties, like this extraordinary capacity to store information. It's the most compact way of storing information. For example, if we take a black hole of size 10 to the -27 centimeters, basically tiny, that has the information capacity of a human brain. So it's an extremely compact way of storing quantum information. Energetically, it's not terribly efficient. The human brain is not doing that bad. And the longevity of the human brain, the capacity to maintain memory, is better than of a black hole (of the similar memory volume).

But the interesting thing is the compactness. And because of this, other properties follow that are very interesting. And so, my approach is a little bit different in the sense that I'm trying to ask questions that, surprisingly, we never asked before. In a sense, they seem to be obvious questions. For example, we say that black holes are very special from this point of view. So they have this large capacity of information storage, and correspondingly, they have all these properties. For example, they evaporate by Hawking radiation, but they release information very slowly, extremely slowly. They maintain information until very late stages of operation. And other properties. So the question I tried to ask some time ago was, what happens with other seemingly completely unrelated objects when we push them towards having the maximum memory capacity? So you take a non-gravitational system, for example, a nuclear forces system with QCD-type interactions. And take a baryon, for example, proton-neutron are baryons.

So if I take a baryon-like object in these theories, obviously, of course, immediately, we say that baryons are very different from black holes. But that's not the right question because when we compare a baryon to a black hole, it's like comparing apples with oranges. So the correct question is, what happens with the baryon when you try to push it toward the maximum capacity of information storage? And what is the maximum capacity of information storage that a baryon can have? And it turned out that, first, this maximum capacity of information storage is controlled by certain bounds, which come from so-called unitarity of the theory. So unitary's something extremely important in physics and quantum field theory. Unitary evolution means everything. A theory that violates unitarity doesn't make any sense.

So all the theories that we're trying to work with, by default, must be unitary theories. So in order for the theory not to violate unitarity, there's a maximal capacity of memory storage, information storage. And this is expressed with certain entropy bounds I suggested. Basically, the entropy bounds are given by the inverse of the interaction strength. The weaker the interacting system, the higher the memory capacity can be. Black holes satisfy this, they saturate this bound. And so, I invented this term, saturon. I call the systems that saturate this bound saturons. And so, if I make a baryon into a saturon, or if I make a magnetic monopole into a saturon, or any other object, how different is it from a black hole? And strikingly, what came out was, they start to behave exactly like black holes in all the respects of information storage and processing. So if they're unstable, they emit information extremely slowly.

So they maintain information internally for a very long time. And their lifetime is very similar to the lifetime (the half decay time) of a black hole. All their properties. In the proper limit, they exhibit a sort of information horizon and this kind of stuff. And of course, I'm very excited about this. There are several consequences of this. One thing this indicates is, the way black holes and other saturated systems store information is based on one and the same universal phenomena. And this phenomena have nothing to do necessarily with quantum gravity. So it's a universal phenomenon of saturation. And in physics, we try to understand different systems of nature. And so, when we have a phenomenon of nature, let's say a black hole, and we try to understand it, there are two ways to go.

One way is to come up with a microscopic theory of that phenomenon. For example, imagine you're observing a bucket of water, and you see that water depletes, evaporates. One possibility is that you can simply come up with a theory. You can say, "I know why water evaporates. Because it's probably made out of molecules, and molecules interact with each other, and once in a while, a molecule with high enough energy jumps out. This is the way it works." Or we can try to see if a similar effect happens in other systems. So you say, "I don't know what water is, but let me observe other liquids and substances." And then, we see that all of them evaporate, and we can understand that there's some underlying physical effect defining them.

And so, this is my approach to black holes. So on one hand, this observation about saturation and saturons is an indication that there's a universal underlying physics that works in all saturated systems. On the other hand, I'm trying to develop, together with my collaborators, a theory of a black hole as a composite object. So this is the idea. Since all the other saturated systems that are non-gravitational are composites, it's natural that black holes are also composites of gravitons. This theory, actually, we started quite some time ago. So I'm thinking about it all the time. So this idea that a black hole is a composite object is several years old. So that's a parallel thinking about black holes.

So these two approaches go together. And of course, these approaches are not terribly standard, especially the first one. But they're a lot of fun for many reasons. For example, because the underlying reason is saturation, we can produce saturated systems in a laboratory. That's not unimaginable. For example, in laboratory systems, we've Bose-Einstein condensates of cold atoms which can reach saturation and this kind of stuff. We can also try to produce saturated neural networks. Actually, I also wrote a couple of papers about trying to understand a black hole as a neural network. And again, if you try to rewrite it as a neural network–"neural network.”

Of course, it's not a biological neural network. It's a quantum network. I call it neural because it behaves like a neural network. But it also develops properties of a black hole. So in this way, for example, an interesting idea would be to take the same mechanisms that black holes use to store and process information and use them in real-life quantum computing. For example, one application I'm thinking about is in this direction. But of course, I'm mostly interested in understanding the theoretic properties of black holes and of other saturated systems.

Zierler:

I wonder if you can explain how one goes about quantifying the information capacity of a black hole. How do you do that?

Dvali:

So we have a system, and this system can be in different states. The simplest way is, you can take a cup, and you and I can agree that if it's upside-down, it means yes, and if it's not, it means no. So this is the simplest way to store information, in the simplest binary code. So information is always stored in the state of the system. The more diverse states a system could be in, the more information capacity it has. This is because you can rearrange states. Or it's like a book. We have a book, and we can write symbols in different ways. And each sequence is some sort of a state of this macroscopic system, a book.

Now, among these different states a system could be in, of course, the valuable ones are the ones that do not cost too much energy. That is, where there's not too much energy cost for going from one state to the other. If you want to rearrange some features in your system, it should not cost too much energy, otherwise it's not terribly useful for storing information. So the interesting ones are the states which are nearly degenerate in energy. So this is the way we count information. A black hole, a system that is macroscopic, has certain little features, quantum features, which we cannot observe classically. So if I'm a classical observer, and I'm observing a solar mass black hole, I think it's just one object. But in reality, this black hole can be in an enormous number of different states, which classically, are invisible. So in order to distinguish among those states, I have to do quantum experiments. The number of those states is enormous.

And so, the logarithm out of the number of those states is so-called micro-state entropy, and that's a good measure for the information storage of a black hole. And so, this was known thanks to the works of Bekenstein and Hawking. First, Bekenstein came up with this concept of maximal entropy capacity for an object. Then, he showed that black holes saturate that capacity. And he showed that there's the so-called Bekenstein-Hawking entropy of black holes. And the amazing thing about this entropy is that it scales like area of an object, measured in units of the fundamental scale, the Planck scale. And so, that's the way we measure the information capacity. We measure the micro-state entropy of an object. So basically, we count in how many different micro-states a system can be, and then we take a logarithm out of that number, and that's a good measure of the information capacity. So that's how we quantify it.

Zierler:

To zoom out from this research, in what ways does measuring the information capacity of black holes lead to larger questions of how the universe works?

Dvali:

So that's another part of my research, which is also related to this discussion about information capacity. And it's connected precisely to the universe. So the universe is very much like a black hole from that point of view. So just like black holes have Bekenstein-Hawking entropy, which scales like the area of a black hole, the universe–by which I mean a standard cosmological homogeneous universe, something that classically obeys the Freedman cosmological equation, carries (a so-called Gibbons-Hawking) entropy.

So we live in an expanding universe. And we are surrounded by a cosmological horizon. And so, the information capacity of the universe is, again, given by the area of that horizon measured in Planck units.

So it's very similar to a black hole. That's not surprising because if you estimate the mass inside our horizon, then we compute the gravitational radius of that mass, the gravitational radius is equal to the Hubble radius. So essentially, the universe inside the Hubble has the maximal mass it can have. So it's like a black hole. Therefore, this fact that the universe has maximal information capacity allows us to understand the universe itself as a saturated system. So the universe is also a saturated system. By the way, one of the striking things is that all the saturated systems have area entropy. That is, their entropy is equal to their surface area. So again, black holes are not special in this respect. The entropy of a saturated baryon state or saturated other states also scales like area.

So this is a universal property of the maximal entropy, that it has to scale like area. And so, it turns out that this is because of unitarity. Unitarity forces maximal entropy to scale like area. And this is related with another part of my research about de Sitter space. So the universe is expanding. We live in an expanding universe. And the way it works, the universe has an energy budget. So we have some energy sources in the universe. They can come in the form of matter, meaning planets, galaxies, dark matter, or radiation, and gravity responds to these energy sources. And so, gravity responds by creating an appropriate space-time geometry. And there is this old story about a very special source called the cosmological constant. This was introduced by Einstein, then retracted, actually. For a good reason, it turns out.

But it's an extremely interesting source. So this is a source that looks like a vacuum energy. So this is a constant term that was introduced by Einstein in Einstein-Hilbert action. And if you ask about the physical meaning of this source, it's that it's what vacuum energy would look like if it existed. What does that mean? For example, normal sources, for instance, normal matter around us, planets, stars, galaxies, us, do not have negative pressure. We have positive energy density and positive pressure. This source (the positive cosmological term) has negative pressure. Because of this negative pressure, it can produce a sort of anti-gravity. So it can lead to the accelerated expansion of the universe. A universe of that type is called a de Sitter universe. de Sitter started these solutions. Such a universe is very interesting.

If we had lambda (a standard notation for cosmological constant) in the universe only –imagine a universe with cosmological constant and gravity only, no matter - such a universe would be pretty boring, but also very interesting. Because any observer in this universe would be surrounded by an eternal horizon. Eternal horizon means that acceleration goes at a constant rate. So the Hubble parameter (which measures the expansion rate of the universe) would be constant in time. And so, the universe would expand eternally. Now, in such a universe, there would also be the so-called Gibbons-Hawking radiation. They showed that in this type of universe, an observer would see thermal radiation basically coming out of nowhere. Because you didn't put in any radiation. There was only a dead cosmological constant. Nevertheless, there is this eternal thermal creation of particles. They get created and diluted immediately. And this is a stationary eternal process.

Now, there is evidence that we are coming from something like that in the past. There's a pretty strong evidence that our universe underwent this type of stage of expansion. This idea of inflation was pioneered by Alan Guth from MIT. And so, basically, this idea says that, temporarily, we lived in this type of de Sitter space. But that was not exact de Sitter space. The energy source controlling this space was not really constant lambda, just a temporary energy source. Lambda-like, but temporary. So of course, this is a current paradigm. And recently, 20 years ago or so, there was a new development observationally.

Before that, the standard picture was that there was this temporary energy density in the universe. That decayed, and now, we're left with, essentially, very small lambda. Either zero or very small. And this created a problem. The problem goes under the name of the cosmological constant puzzle. And this is probably one of the most outstanding problems in cosmology when you view cosmology in terms of particle physics. Once you marry cosmology and fundamental physics, it's one of the most celebrated problems. What's the essence of the problem? In very general terms, we're trying to understand nature in physics. When we're trying to understand certain phenomena, the universe, black holes, superconductors, for example, we come up with a theory. Any theory has a domain of validity. Beyond this domain, we cannot say that the theory's wrong. It's simply not applicable.

For example, Newton's theory is fantastic if you want to describe planetary orbits of relatively slow-moving objects. But if you want to describe relativistic sources, like photons moving, Newtonian theory is useless. So you have to go to Einstein's theories. So these theories are embedded in one another. So when we go from one domain of knowledge to the other, we try to complete the theory. The same is true, for example, when we try to describe nature at large distance, and then we try to go to smaller distances, and we decrease the distance, and so on. For example, if I want to describe air in this room at large distances, I can use theory of sound waves. So there are sound waves in the room. If I want to describe the same system at distances comparable to inter-molecular size, obviously, the sound wave theory breaks down, and I have to go to a molecular description.

So then, I have to go the description in terms of atoms, quarks, leptons, and so on. So the way it works is, the more fundamental we try to be, we need more powerful theories. And each theory has a domain of validity. So far, we are always in a domain of validity of some theory, and that's normal. Now, suppose you are in some domain of validity of a certain theory. So you've established a theory that works in that domain. And let's say your theory works starting from a one-mile distance down to a micron. So you have a description from one mile to a micron. Now, in this world of one-mile objects, there are different parameters. And you have to ask, how sensitive are these parameters to the physics beyond my boundary? That is, how sensitive are properties of one-mile objects to the physics at micron scale?

In other words, how sensitive are measured, observed, computed parameters to the boundary of your knowledge? So when we go from large distances to short distances, we have a term for this sensitivity. We call it an ultraviolet sensitivity. And that's one of the most important concepts in physics. But this goes beyond physics. You can apply this to an arbitrary domain of knowledge.

So you have a theory that nicely describes your system. I don't know, sociology, for example, or something. But there always is a boundary of your knowledge and the question is how sensitive your computations are to the unknown, to the boundary of the knowledge. So if you apply this question to the cosmological constant, the cosmological constant, seemingly, turns out to be the most sensitive quantity to the unknown. So in Einstein gravity, for example, we can get away without too explicit computations, such as quantum corrections, up until we hit a certain distance, the so-called Planck distance. The Planck distance is basically an inverse square root from Newton's constant. So this is the shortest length scale of quantum gravity. And that's the length scale beyond which quantum gravity becomes 100% important and by no means can be ignored. And so, when we come up with theories of nature, it's very important to ask, how sensitive are our computations to this very short distance physics, to the cutoff of the theory? And it turns out, seemingly, lambda is the most sensitive quantity. So lambda, the cosmological constant, turned out to be essentially maximally sensitive to the cutoff of the theory. So extremely sensitive.

So now, this created a puzzle. Why is there a puzzle? Nobody has computed lambda. We don't know how to compute it. Because again, we can compute certain low energy contributions to lambda, but even those low energy contributions are computable only within low-energy effective theories, and secondly, these low energy contributions then are matched by contributions from the Planck scale. And that contribution, we don't know how to take into account. So the only thing we can do is to estimate. We can say, "Let me estimate this contribution approximately." And if you estimate this contribution, this contribution comes out to be enormous, of order the Planck scale. The inverse of the Planck length is so-called Planck mass.

By the way, I should make this clear, the Planck length is the most fundamental length scale in gravity, and Planck mass is the most fundamental mass scale from the point of view of particle physics. Why Planck mass is very important. Because Planck mass is the boundary between particles and black holes. So you cannot have an elementary particle which is heavier than the Planck mass because any elementary particle heavier than the Planck mass will become a black hole. And so, essentially, the elementary particle physics world ends at the Planck scale. Planck mass is something like 10 to the 4-5 grams. So it's a huge mass scale for particle physics. So the point is, this lambda is particularly sensitive to the Planck scale. So this is what came out. And this is a puzzle. A so-called cosmological constant puzzle. The mismatch is approximately 120 orders of magnitude or something like that.

So you have a Planck scale quantity, which you are naively estimating, and with that naive estimate, you expect it to be 120 orders of magnitude bigger than what the observational bound tells you. So this is the puzzle.

By the way, here is another reason why lambda is very important. Because the current observations indicate that we are entering the era of accelerated expansion of the universe. So basically, this started approximately 20 years ago or so. So observations indicate that we are entering, again, in de Sitter type epoch. Of course, with a huge Hubble radius. The Hubble radius is approximately 10 to the 28 centimeters (10 billion light years). So de Sitter is very important to understand because we're coming from de Sitter and we're re-entering into the de Sitter.

So they're the sort of two endpoints. And then, in between, we have this cosmological constant puzzle. Why is lambda so small? And now, the point is that this problem of lambda, there's a huge literature on that. I also worked on that a lot. People came up with all possible very interesting ideas. Basically, nothing worked. So we were approaching this problem as a naturalness problem. Somehow, it looked highly unnatural to have the cosmological constant so small. So this is another thing. We're trying to expand our knowledge and go towards the unknown, and we need some guidelines. What are the guidelines? Consistency is one. Sometimes, you formulate a theory, and that theory tells you what kind of consistency requirements it has. Another guideline that we use is naturalness.

Actually, naturalness led to very important breakthroughs, for example, in the case of strong interactions. The question why the pion is lighter than a proton led to an extremely important breakthrough of understanding the pion as a Goldstone boson, etc. So naturalness is an important guideline. This was the story.

Now, the way the previous discussion about saturation applies to the universe is–I don't know how much history I should go into, but it's a personal history, also. So there was the story that in '98, a collaborator of mine, Henry Tye from Cornell, and myself tried to ask the question of how to generate a de Sitter-like state in string theory. And so, we were mostly motivated by inflation. We wanted to construct some kind of temporary de Sitter state in string theory.

So we realized that string theory had ingredients for this. In other words, it had ingredients for the energy density required for de Sitter in the form of so-called D-branes. In string theory, they (D-branes) are these extended objects, like surfaces, and they have this interesting property that open strings can end on them. So the strings are attached to them and vibrate. Now, D-branes have this property that is similar to the cosmological constant. So basically, they effectively generate a cosmological constant-like term, because of their tension which provides a negative pressure. So our idea was to create a temporary de Sitter in string theory using these D-branes. So for example, we were thinking of having D-branes and anti-D-branes. And then, these D-branes move toward each other. While they move toward each other, the universe inflates, and then they collide and annihilate, and the universe stops inflating, and the hot Big Bang starts. And later we continued to w work on this and Henry independently was writing some interesting papers about cosmic strings in this scenario.

I was trying to find out whether this system could have been stabilized. In other words, whether there was a possibility to really create some sort of an eternal de Sitter in string theory. And it was impossible. So every attempt would fail, and every time, there was a very good reason for that. And so, it was sort of clear that somehow, string theory–by the way, string theory is a theory of quantum gravity, which we pretty often use as a reference point because it's sufficiently well-understood that it's consistent. So therefore, it makes a good reference point for understanding many things about quantum gravity. And so, it was clear that string theory hates de Sitter. After a few years, I concluded this was impossible because there's a good reason for it. Then, what happened was, people got interested in this type of inflation model building using D-branes. And very often, people would make the assumption, "OK, let's assume that you stabilize it."

This is one of these things that looks plausible because if you look naively, you think that there must be an infinite number of possibilities to produce de Sitter in string theory. But if you look closely, you'll see that every time you try, it fails. You have to make some assumptions. So it's like a duct tape. You need assumptions to patch things up. You say, "OK, this should work." So that was the situation.

And so, I was thinking that there must be some more fundamental reason why it was impossible because it's very hard to argue on a case-by-case basis. Because every time, construction may be extremely smart, very complicated, and it's very hard to find the problem or the mistake.

So I said to myself that there must be a general reason why string theory cannot tolerate a de Sitter-type state. And then, a collaborator of mine, Cesar Gomez, and myself, we stated that de Sitter has to be viewed in the light of consistency. So there is something fundamentally inconsistent about de Sitter from the point of view of quantum gravity, and so it has to be viewed in this light. And we showed that de Sitter has sort of an anomaly. In other words, a de Sitter-like state in quantum gravity exhibits a certain type of anomalous behavior, which gives inconsistency. And the reason is very general. It has to do with saturation. So the reason, again, has to do with the fact that de Sitter is also one of those saturated states. So basically, we concluded that in quantum gravity, and particularly in string theory, you cannot have any de Sitter-like vacuum or anything like that. There cannot be any eternal de Sitter, and there cannot be any eternally inflating Hubble patch.

Inflation must have a strictly bounded number of e-foldings, number of Hubble times that it lasted. And the cosmological constant must be exactly zero. So basically, our point is that quantum gravity predicts the cosmological constant to be exactly zero. Now, I must say that this is completely opposite to a competing view, which was based on so-called anthropic reasoning. Because this reasoning says, "OK, we don't understand why lambda is small." But some time ago, Weinberg showed that with lambda larger than a certain value, pretty close to the observed one, galaxies would not be formed.

And so, this sort of became a basis for this anthropic reasoning. Because the anthropic reasoning says, probably, quantum gravity gives us all possible de Sitter-like vacua, all possible lambdas or cosmological constants, and we just live where we can. Because with a larger lambda, we could not live. And that's completely opposite to what I'm trying to say. I'm saying that quantum gravity predicts strictly zero lambda. In our budget of the universe, there's no lambda. It's exactly zero.

Zierler:

Is there anything going on right now observationally in astrophysics or cosmology that's particularly relevant for this research?

Dvali:

Oh, yeah, absolutely. So before, we could say, "We live in a universe with lambda, and that's the simplest explanation." Because in order to explain the accelerated universe, by using the Standard Model plus Einstein gravity, you don't need to go beyond. You just put lambda, a constant. And that explains acceleration. So that would be the simplest explanation of the accelerated expansion of the universe. But I'm saying that this is impossible. So since I'm saying that constant lambda has to be zero, immediately, this makes the question of dark energy much more interesting.

Because now we know that dark energy, whatever it is, has to come from some new physics, some new degrees of freedom. And so, therefore, it's extremely important observationally to really differentiate. And – first of all, cosmologically, all these types of energy have a certain equation of state. Equation of state for lambda is so-called w = -1. And so, we're predicting that w has to be substantially different from -1. So therefore, observationally, yeah, it's very important. Observationally, we should see deviation from exact constant, which means that we should see deviation from the equation of state of -1. So, with the present observational accuracy, we actually see it's not there yet. But yeah, the observations are extremely important. Because the prediction is that we should see a different equation of state.

Zierler:

A less technical question, one we're all dealing with, how has your work been affected by the pandemic?

Dvali:

Of course, there were negative things. There was a lockdown, and you cannot give lectures personally, and you don't see students. But there were also positives. You travel less, and you can give more talks with less travel. So I think some of this technology, we should keep. Because sometimes, you travel only for one day, you give a talk and then fly back. Now, it's much easier. So that part, I think, should be maintained. Less travel, less pollution.

Zierler:

Let's take it all the way back to the beginning. Let's go back to Georgia. Tell me about your parents.

Dvali:

Well, my mother was a scientist, a biologist, and my father was an engineer.

Zierler:

So you had it in the blood.

Dvali:

In the way I was raised, yes.

Zierler:

What were your family's experiences in World War II?

Dvali:

My grandpa fought in the War. And on maternal side the brother of my grandpa died, and another brother fought the entire War, came back OK.

Zierler:

What town did you grow up in?

Dvali:

I grew up in Tbilisi. Of course, when I was growing up, the memories of World War II were still around, but faded a little bit. Tbilisi is a pretty old city. It became the capital around the fourth century. The ancient capital is nearby. But it's a very interesting place. When I was growing up, especially, it was surprisingly cosmopolitan. And it has a very old culture and tradition of tolerance in all possible ways. And it's a very special place. I think it was a great experience, growing up in this type of city.

Zierler:

What memories stand out in your mind, growing up with Georgia being a part of the Soviet Union?

Dvali:

The Soviet Union was still extremely powerful, but certain thing were not so sharp. For instance, there was a certain level of tolerance throughout the Soviet Union. But Georgia was more shielded from this communist ideology. I don't know the reason. Probably because of, first, the cultural traditions. The area tried to sort of shield itself from the communist regime. Also, at that time, the secretary of the party was Shevardnadze. And this guy was really democratic already, then. He later became one of the main pushers of Perestroika. So he had some internal democracy. So therefore, there were things happening in Georgia, for example, in the Georgian film industry or writings, which in the rest of the Soviet Union would be unimaginable.

But somehow, it was tolerated. I don't know how to explain this. Because for example, one of the clear-cut anti-Soviet movies was made by Shevardnadze's request, apparently. And these things were known outside of the Republic. But I don't know the reason why these things were tolerated. But the point was, somehow, the atmosphere here was different. I don't know how to explain it. Maybe they wanted to keep at least one place relatively free or something.

Zierler:

Do you think the legacy of Stalin had anything to do with that?

Dvali:

No, the legacy of Stalin did the opposite. It's nothing to do with the legacy of Stalin. Maybe, again, because of this history of tolerance. It was a very special place. Maybe that was another reason. I'm still not sure. But the atmosphere was very different. Because from what I was hearing from my friends who lived elsewhere, like in Moscow or other places, it was very different. Also, for instance, TV programs. There was nothing like that in the sense of TV in other parts of USSR. There were educational TV programs about, for example, movies, and you would not see anything like that on any other channels throughout the Soviet Union. There was this communist sort of propaganda in the background, but I think it was already not so sharp. And I don't think anybody was taking it seriously, including communists themselves. So it was sort of going by inertia already by that time.

Zierler:

What was your family's religious background?

Dvali:

Well, the Soviet Union had sort of an atheist ideology. But my family's background is Georgian orthodox Christian. Georgian Orthodox church is a separate Christian church. My family was not observant, but they came from that background.

Zierler:

What languages did you speak growing up?

Dvali:

Well, we spoke mostly Georgian. I went to a public school because there were only public schools. There were private kindergartens, but schools were all public. But many schools, they had specialty in languages. There was a school that would specialize in English, a school that would specialize in Spanish. And the one I went to was called English School. So we started English essentially immediately. And Russian, we started in 3rd grade as an additional language. The main language at school was Georgian. This is not widely known, but the official language in Georgia was Georgian, not Russian. Of course, you could survive with Russian no problem. Especially in Tbilisi, people spoke Russian. So I was exposed to Russian pretty early. So my Russian is reasonably good. And I read a lot of books in Russian. But we didn't speak Russian at home. I had some Russian friends, and I spoke Russian with them, but mostly Georgian otherwise.

Zierler:

When did you start to get interested in physics specifically?

Dvali:

In Georgia, there was simply one school which incorporated education from the elementary all the way to a high school. You would go to the same building from elementary to high school. So I got interested in physics when I was in a middle school, so to speak.

Zierler:

And when did you realize that you wanted to pursue physics for college?

Dvali:

I was very interested in science in general. I was also interested in biology and chemistry. And so, I was undecided, but I wanted to do something scientific. I was also thinking about medical school. But then, I changed my mind at the last moment, basically. So the highest grade was 10th. I was preparing for the college exams because you had to pass exams to get into university. So I changed my mind at the last moment. Then, I decided to enrol in physics.

Zierler:

Was it always theoretical physics? Did you ever think about pursuing interests in experimental physics?

Dvali:

No, it was mostly theoretical physics. I was interested in experiments, but it was mostly theoretical physics all the time.

Zierler:

What were some of the research areas that your professors as an undergraduate were doing?

Dvali:

This was true throughout the Soviet Union. The bachelor and master's were together. We'd study five years, and then we'd get essentially something equivalent to a master's degree. It was called diploma. And so, when I was an undergraduate, I didn't have many research discussions with the professors. In the first or second year, I did sort of a review type project with one of my professors. Now, this professor was not very active in the research, but he was very knowledgeable. This was typical in many Soviet universities. In many top-level universities, actually. You would find professors who were extremely good teachers and extremely knowledgeable, but not so much publishing. There was a culture, I think, where there was sort of no pressure to publish.

So there was this project in elementary particle physics I did about the spontaneous breaking of symmetry. And then, I had to present it. I gave a presentation, 10 or 15 minutes, and I explained this spontaneous breaking of symmetry. You've probably seen it many times. There's this Mexican hat potential, and there's a ball on top of the Mexican hat potential. And if the ball is on top the hill, the system is symmetric, however, the ground state is not symmetric. So there is this illustration. And I read a few things about it, and I understood. Now, the reason I know that I understood is that now when I'm teaching the same thing, I think my understanding was more or less OK at the time. Of course, I know more now. But the concept, I understood.

So I gave this presentation, and there were these committee members, very nice people, who made these concluding remarks. They mentioned different things. And they said, "Look, this is great, this presentation about spontaneous symmetry breaking stuff. But these are very deep concepts, and undergraduates cannot understand it." [laugh] And I said, "OK, wait a minute, I understand it. What do you mean?" So it's interesting because sometimes, you tell people, and people say, "Yeah, everything you're saying is correct, but you cannot possibly be understanding this." [laugh]

Zierler:

Did you ever think about leaving Georgia for graduate school? Is that something that was available to you? Or you wanted to stay close to home?

Dvali:

No, it was available. But only within Soviet Union at that time. Perestroika already started, but it was a new movement, so borders were not open. So I had the possibility to go to Moscow or St. Petersburg. But somehow, I decided to stay in the group because there was an extremely good institute where I did my diploma. This was Institute of Physics. And so, the head of the department was Oleg Kancheli. He's well-known, especially in QCD. He did some important work there. And so, I wanted to stay with them. So I did the diploma first with them. First, I did a preliminary bachelor-type physics thesis with him, but he was sick, so I did it independently. And it was about spontaneous symmetry breaking, Higgs effect, and these kinds of things. And then, they invited me to do a diploma with them.

So then, I did my diploma physics with them. Meanwhile, they offered me to get a PhD position in their group. And I was very happy. So I accepted. So they had to ask for funding. And it was pretty standard. They would ask for funding, and funding would come in September. So they put in a request for funding. And when September came, it turned out that somehow, somebody forgot to send the request. So this request was never sent. And they had a problem. But finally, it worked out because they finally gave me a real position. So I ended up having a real position. So in the Soviet Union, you could do a PhD while being a graduate student or while having a research position.

Zierler:

Now, in graduate school, your work was largely in elementary particle physics. Were cosmology and black holes not yet on your radar?

Dvali:

Black holes, not so much. So what was on my radar, and still is since those times, is the so-called hierarchy problem. We have the Standard Model. We call it standard because it's so well-understood and it works so well. The Standard Model of elementary particle physics, which has three interactions: strong, electromagnetic, and weak. Weak is responsible for radioactivity, and strong is responsible for nuclear forces. So these three interactions are described by the Standard Model. The theory is extremely successful. That's why we call it the Standard Model. And the Standard Model has one ingredient, the so-called Higgs field. The Higgs field is very important because it's important for generating masses for certain particles. Basically, this Higgs field provides a condensate. So all of us, we live in this condensate. We literally are swimming in the condensate of the Higgs field.

And so, we swim in this condensate, and the elementary particles that compose us interact with the condensate, and they get masses. So that's why particles have masses. Now, this condensate is pretty interesting. From all the reference frames, it looks the same. So this condensate has waves in it, and these waves are Higgs particles. So we tried to discover the Higgs particle by colliding particles. Basically, by shaking the condensate and trying to produce this wave of the Higgs boson. And the point is, the mass of the Higgs boson is approximately 100 GeV, 125 or something. And at the same time, it's extremely UV-sensitive to the Planck scale. So here comes again this problem of UV-sensitivity. And this hierarchy is a big question. Why is the Higgs much lighter than the Planck scale?

Or you can translate it as, why is gravity so much weaker than the weak force or the electromagnetic force? For instance, if you compare an electromagnetic force between two protons with the gravitational force, the ratio is something like 10 to the 37. So gravity is much weaker by a huge amount. And the reason for the puzzle is this hierarchy. So I was working on this puzzle of the hierarchy in the context of grand unified theories. So grand unified theory's idea is to unify these three interactions of the Standard Model into one. And grand unification also happens at a very high scale, approximately 10 to the power of 16 GeV. And there, this problem becomes even sharper. Because there is this question to understand the hierarchy between the mass of the Higgs particle and the grand unified scale. So I was working on that in my master's thesis. And I had this idea that a Higgs field is like a sound wave, like a Goldstone boson. So this was one of the approaches I was trying to develop. But I also went into cosmology during my PhD. I also included things like cosmic strings. But black holes were not there yet.

Zierler:

In the international context, of course, the Cold War ended while you were a graduate student. Did things like the fall of the Berlin Wall or the breakup of the Soviet Union affect you personally, your day-to-day, the opportunities that you had?

Dvali:

Yeah, of course. For instance, the way I went for a post-doc–actually, I didn't even know about the concept of a post-doc. Because in the Soviet system, there was nothing like that. And I was visiting the International Center for Theoretical Physics in Trieste. And so, they offered me a post-doc position. So I interacted with a colleague of mine, Goran Senjanovic. So they offered me a post-doc position, and I got another offer from Pisa, from Riccardo Barbieri. And I didn't even know what it was. It was the first time I'd learned about the postdoc position.

Zierler:

This is to say that, had the Soviet Union not collapsed, post-docs beyond in Europe would not necessarily have been available to you?

Dvali:

Probably. Within the Soviet Union, it was a different system. Scientific ideas were made differently. There was no such thing as moving around. For example, people would start in the same institution, and you could work at the same institution all your life. Within the country, it was pretty normal. You could go for different seminars to different universities. But position-wise, there was no culture like post-doc positions or anything like that. You would start at an institute and typically would stay there for the entire career. So if the Soviet Union would not have collapsed, that might've been me.

Zierler:

Did you see your work as a post-doc as an opportunity to get more involved in cosmology?

Dvali:

Yeah. I started to be more involved in cosmology during my post-doc years.

Zierler:

What was your research, first at the ICTP and then at the INFN?

Dvali:

So at ICTP, the area I was working in, we can call it beyond the Standard Model physics. So anything beyond the Standard Model. And when I moved to ICTP, I had a few directions I wanted to pursue. One was this grand unification. Again, understanding the hierarchy problem in grand unification. And I came up with some mechanisms, some ideas along those lines. And another direction was these so-called topological defects, these structures that are formed in the early universe as a result of phase transitions. Just like when we cool down water, it doesn't freeze uniformly, there are always defects. Exactly the same when, when the universe was cooling down, it didn't cool down uniformly. There were always some defects formed, and some of them are very interesting. We call them topological defects.

For example, you can have stringlike objects called cosmic strings, or you can have domain walls between two different domains, and so on. You can have magnetic monopoles and this kind of stuff. So that was another direction. So I was fascinated with this idea of having topological defects around. And some of those defects are problematic. In other words, if they would be produced and stay around for too long, they would over-close the universe. And it requires explanation why they're not around. And many extensions of the Standard Model predict those defects. And so, there was this question, what are the possibilities? And usually, the standard explanation was inflation. But for instance, Goran Senjanovic and I wrote a paper about symmetry non-restoration. So this was the idea. Previously, it was put forward by Weinberg then by Senjanovic and Mohapatra.

So the idea was, under certain circumstances, when you heat up the system, symmetry is not getting restored. So if that's the case, then you don't have these domain walls and other topological defects such as monopoles. So yeah, we worked in that direction. And then, we (Goran and I ) wrote a couple of fun papers about so-called ``Alice” cosmic strings. I think they were put forward by Schwarz. They're a certain class of these topological strings which have a fascinating property that, if you take a particle around them, it can come back as an anti-particle. Now, those are not realistic, but it turned out that you can have some kind of flavor analogues of those so that you can convert particles of different flavors into one another when they go around the string. And so, we wrote a fun paper about that also.

Zierler:

Would you say that at that point, you were primarily operating in the world of string theory? Would you have called yourself a string theorist by the time you were a post-doc?

Dvali:

No, not really. I was learning some string theory as a framework. I was using string theory, and I was trying to make a connection with string theory. For example, with Riccardo Barbieri and Alessandro Strumia in Pisa, we even wrote a paper trying to solve certain problems with grand unification and string theory. But also, and this was something that became defining of my research, at ICTP, I was trying to understand two things. First, whether there can be counterparts of stringlike behavior in quantum field theory. So whether we could understand certain stringlike phenomena within quantum field theory. And that was really very interesting. So I was trying to understand the question of supersymmetry breaking. Because at that time, the idea of supersymmetry was very popular, very intensely discussed.

Supersymmetry is a symmetry between bosons and fermions and predicts that every particle of the Standard Model should have a super-partner of a different spin. But since we don't see the super partners around, this means that supersymmetry is not an exact symmetry of nature. It's a broken symmetry.

Now, how to break supersymmetry is an issue. In particular, in string theory, we're still struggling to understand what is really a mechanism for supersymmetry breaking. And so, at that time, I was thinking in terms of field theory, and I was trying to understand whether we could break supersymmetry by imagining that we live inside a topological defect that is embedded in a higher dimensional space. These ideas were too naive at this point because many of the topological defects that appear in string theory do not break supersymmetry. And the ones that do are unstable.

But at that time, this was the idea, to understand, what we observe as supersymmetry-breaking, by somehow living inside a topological defect. And then, we made progress in this direction. It was really great fun to work with Misha Shifman. So when I moved to CERN as a post-doc, Misha was on a sabbatical from Minnesota. And we started to discuss this concept of supersymmetry breaking simply because we live on a non-supersymmetric ``island” embedded in extra dimensions. So the idea was this, that there are extra dimensions which are huge, even infinite, and there is an island in these extra dimensions, which is this topological defect. And in reality, supersymmetry is not broken, but because we live in this topological defect, we don't see supersymmetry. Because in order to see supersymmetry, you have to sort of come out of it and look at our world from outside, i.e., with the eyes of a high dimensional observer.

So effectively, this defect breaks supersymmetry on its world-volume. So this was the idea. So we wrote papers on that, and that was a great fun. We didn't know what to do with gravity, so we didn't discuss gravity in those papers. But for example, we came up with a mechanism for how to localize a spin-1 fields, gauge fields, on a topological defect. In quantum field theory, previously, this was considered an unsolved problem. And even after we wrote the paper, we had a lot of discussions with our colleagues because it took some time before they finally accepted this as a working mechanism. In particular, a colleague of mine, a great physicist, Valery Rubakov. He was very skeptical at the beginning. Finally, he accepted it, and he wrote a paper on that. So yeah, it was fun because this started my interest in extra dimensions and stuff.

Zierler:

And at this point, there was a lot of optimism that supersymmetry would be soon. Obviously, not at the SSC, but at CERN.

Dvali:

That's right. It's very hard to imagine that atmosphere. I really liked all my colleagues at CERN because they were great people to talk to. But the groups were divided into phenomenology and beyond the Standard Model, and then string theory. And essentially, when I came there, there was zero communication.

Zierler:

Because they were thought of as separate fields?

Dvali:

No, no. I never thought of them as separate fields. String theory is part of elementary particle physics. But it could have been the local situation there. Maybe in other places, it was different. The field is huge, and people were pursuing certain questions that were considered more formal. So there are questions that are more technical and formal. And the phenomenology people were pursuing questions that are closer to experiments. But that's what I remember. And I was very interested in breaking supersymmetry by D-branes and this type of thing. Of course, I was talking a lot to string people there, and I was also talking a lot to phenomenology people. But then, it was really great when Misha came because he's a great expert in formal aspects of quantum field theory as well as in phenomenology, and he was the right person to pursue these questions.

Zierler:

Did you interact or have contact at all with Joe Polchinski at this time?

Dvali:

During that time, no. Not during that time. But later, yes.

Zierler:

Tell me about your impressions of CERN. What was it like when you first got there?

Dvali:

Well, being a post-doc there was great. It was an extraordinary experience. First of all, there were several people from my generation who were very active. So the nice thing about CERN was that the post-docs were self-sufficient. So you didn't really need any senior person to work with. There were senior people. Fortunately, Savas Dimopoulos was there. And we started to collaborate. And of course, it was an extraordinary experience. But also, within post-docs, we would collaborate with each other a lot. And from that point of view, it was a great atmosphere. And all the freedom that CERN gave was great.

Zierler:

Was it useful to have access to the world of experimentation? Were you connected at all from the theory group to what was happening at the LHC?

Dvali:

So at that time, when I went there as a post-doc, LHC was in the planning stage. So I had some interactions with experimentalists. For example, Misha and I wrote a paper about breaking supersymmetry on a topological defect. We said in that paper that, together with super partners, extra dimensions will be discovered. And I remember, I got some messages from experimentalists, and they were very interested with what kind of predictions there could be and this kind of stuff. So yeah, I had some discussions with experimentalists.

Zierler:

When did you first meet Nima Arkani-Hamed?

Dvali:

More or less around that period. So I think Nima came to give a seminar or something. So we met there. And then, later, after CERN, I went to Stanford to visit Savas, and Nima was there. Then, we started to collaborate.

Zierler:

And tell me about the origins of the ADD model. How did that all come together?

Dvali:

Oh, yeah, that was a great experience. So Savas was at CERN. And we even wrote a paper about string theory compactifications, etc. Before that, we'd written other papers. And then, we started to discuss this project, which is why I came to Stanford. And then, we started a really intense discussion about this and about what to do with these extra dimensions, how to make sense of gravity with these extra dimensions, etc. So as usual, you're moving in the dark at the beginning. At the beginning, we thought it may not work. Then, we became very optimistic. But there were these new questions coming up all the time. Nima was extremely instrumental, and it was great to collaborate with him. It was a real advantage to have him. And it really resonated very well. It was a collaboration where everybody really brought something unique. Probably, nothing would've happened if any of us had worked on it alone. It was very organic.

So there were certain things you could not imagine without someone else. So from this point of view, yeah, it was extremely high resonance. It was great fun because, of course, this opened up a big arena for all possible crazy ideas that we had. But there were new questions coming up all the time. Before publishing the paper, I gave a talk. Essentially, there were three of us in the audience helping each other. So I gave a talk at Stanford, and during the talk–here, you have this extremely light tower of particles (Kaluza-Klein gravitons), and you have large extra dimensions, and therefore you have a huge number of gravitational excitations. And normally, when you have this type of a situation in beyond the Standard Model physics, you have a large number of weakly interacting particles. The most dangerous place is a star cooling. Because these particles can be produced inside the star, and they can escape.

A normal photon cannot escape very easily because it re-scatters many times before it has a chance to get out of the surface of the star. So for example, a photon created near the core of the sun takes millions of years to get out. But these particles, because they're extremely weakly interacting–like neutrinos. They are sort of like neutrinos, right? Even weaker interacting. So when you produce them inside the star, they escape because for them, the star is absolutely transparent. And they take away energy. And one of the most sensitive places for new physics is the star cooling. Because you don't want to affect too much the star cooling rate. Because star cooling rates, as they are, are pretty well-understood within standard mechanisms.

But the crazy thing was, we gave this talk, and after the talk, we realized that we completely forgot to ask this question about star cooling. But nobody asked in the audience. People were asking all possible questions. There was a lot of discussion. But nobody asked about star cooling. So there are certain obvious things that sometimes, you don't think about. But at the end of the day, we realized that everything worked out.

Zierler:

I'm curious how influential Ed Witten's advances on M theory were from a few years earlier on the ADD model.

Dvali:

You mean the compactifications with supergravity? Yeah, they were inspiring, of course. The fact that you can have the new scale in the problem. So the interesting thing about that was, that compactification clearly showed that you can have a new scale in the problem, which is related, but not directly related to the Planck scale. Because the Planck mass can be somewhat larger than the scale of compactification radius (usually taken as the string scale). That work was really fascinating on its own. It's a masterpiece. It was a great pleasure to see and read it. In addition, I think, from what I remember, it did give some additional spiritual inspiration.

Zierler:

And was the operating assumption with the ADD model eventually that the proton collision experiments at the LHC would bear this out?

Dvali:

Yes. Unfortunately, in physics, it's rare that we can really predict the scale of new physics. Usually, we put bounds. Now, the idea of ADD was to solve the hierarchy problem. So therefore, this was the motivation. Therefore, if the hierarchy problem is solved because of new physics which has to do with large extra dimensions, then it has to be nearby somewhere. So yes, the big hope was that LHC would see some of those predictions that we made, like quantum black holes, or Kaluza-Klein tower, and this kind of stuff. But the problem there is, there's a standard problem. For the hierarchy problem, the factor of five or ten is nothing. If the fundamental scale is a factor of ten higher, that still solves the hierarchy problem. Because we're talking about a problem of 34 orders of magnitude.

In a problem of 34 orders of magnitude, if the scale is factor of ten, a little bit higher, that is still a solution. But the problem is, with high energy experiments like LHC, there's a lifetime. And this is the standard problem that we're facing in particle physics, unfortunately. This is a standard problem in our field, in physics beyond the Standard Model. There's nothing we can do. We have to push the frontiers. Personally, I take this hierarchy problem extremely seriously. Especially in this light of cosmological constant. Actually, I just wrote the paper a few days ago just to comment. Because my point was, since these arguments about the s-matrix and quantum breaking, since in string theory, we don't have a de Sitter landscape, this also kills or makes extremely problematic the possibility to explain weak scale anthropically.

Because there were also these ideas to explain the weak scale anthropically. The idea is very nice, of course, because somehow, the value of the weak scale controls the stability of atoms and so on. So there were these ideas, but if there is no de Sitter landscape, then the anthropic selection is impossible. Because you need an actualizer. For anthropic selection, it's not enough that, for example, some parameter is crucial for our life. You also need an actualizer, a mechanism that actualizes all possibilities. (That is, a mechanism that can provide different values of the parameter in different parts of the universe). And without de Sitter landscape and eternal inflation, this is impossible. And so, this was my point. Now, the case for the hierarchy problem, for the new physics near the weak scale, essentially, it is stronger because we cannot rely on anthropic solution. So therefore, yeah, there must be some new physics in the neighborhood (of the weak scale) somewhere.

Zierler:

The ADD model, of course, garnered a lot of attention. What were some of the most relevant criticisms that stand out in your memory about the ADD model?

Dvali:

It's good that you're asking that question because this goes into the core of physics and what the hierarchies are about. Because at the beginning, maybe even now, some of our colleagues were telling us, "You have this construction, but you're trading one big number for another big number. Because let's say we have a hierarchy between the Higgs mass scale and the Planck scale squared. You have one hierarchy. And now, you have another hierarchy, which is the hierarchy between the fundamental Planck scale and the compactification volume because now, you are introducing a big number, a volume of extra dimensions. So you're introducing another big number." This was the sort of criticism. Now, this is a misunderstanding of what the hierarchy problem is about.

Of course, there is a technical problem about how to stabilize the volume of compactification. That's a separate story. But the question about hierarchies is a question about ultraviolet sensitivity. Once you express hierarchy in terms of non-sensitive quantities, then you are making some progress. And so, the important thing is that compactification volume is not ultraviolet sensitive. So we cannot compare these two cases. It's apples and oranges again. Because the hierarchy between the Higgs mass and the Planck scale is a question of ultraviolet sensitivity. But compactification volume is not ultraviolet sensitive. In nature, there are plenty of big numbers. We are not saying that there's a hierarchy problem associated with all of them. For example, if you take bacteria, it's much smaller than an elephant. But we are not saying there is a hierarchy problem. Not every big number is a hierarchy problem. The hierarchy problem is the problem that has to do with ultraviolet sensitivity with respect to quantum corrections. So this was important. So this was one sort of misunderstanding that I remember.

Zierler:

How long did you stay at CERN?

Dvali:

Well, first, I was a post-doc at CERN. And then, I came back as a permanent staff member later, in 2007. But I only stayed for three years.

Zierler:

But those were a very exciting three years because that was right before the discovery of the Higgs, of course.

Dvali:

Yeah, it was exciting. But we were expecting more from the LHC. The discovery of the Higgs is great, there's no question about that. But you have to understand, we are a generation of Higgs and Standard Model. So for us, Higgs was something so obvious that it sort of became boring at the end.

Zierler:

You sound like a real theorist.

Dvali:

Well, yeah. So towards the end, we started asking these questions. For example, Higgs is mandatory in Standard Model because of consistency, because it has to unitarize the theory. Without the Higgs particle, theory would not be unitary above a few hundred GeV. Then, we started asking these questions. Can theory be unitarized by some other mechanism? For example, we had this idea, Cesar Gomez, Alex Kehagias, and Gian Giudice. We called it classicalization. So the idea was whether the Standard Model can unitarize by producing some very high multiplicity soft particles. And high multiplicity soft particles is always like a classical object.

So basically, by producing classical objects, which would become more and more classical. So the idea was borrowed from black holes because that's what happens in black hole physics. If you collide elementary particles at very high momentum, at very high center of mass energy, once this center of mass energy exceeds the Planck mass, the outcome of these collisions are black holes. So the colliding high energy particles produce black holes. So we sort of wanted to explore the same idea in the Standard Model. So for example, it would have been very exciting to see that instead of a Higgs boson. But of course, the Higgs is the simplest explanation, and it works.

Zierler:

Tell me how the opportunity at New York University came about for you. Were you looking to leave CERN? Were you looking to have an academic position?

Dvali:

At CERN, I was a post-doc. So it was natural to move on. So I got an offer for the tenure track position at ICTP. And so, we moved there, our family. And so, meanwhile, I got an invitation to a few US universities that had openings in tenure tracks. So we moved to ICTP, and then I went on sort of a tour. So I gave some interviews and stuff. And then, NYU came at the last moment, actually. I was already considering other offers. So I flew there for a few days for the interview. It was June or something like that. And then, finally, I went with NYU.

Zierler:

And what was your research at that point with regard to extra dimensions in space?

Dvali:

At that point, we'd already written ADD papers. And then, I was continuing in that direction, and that fall was when Henry Tye and I came up with this idea of creating temporary de Sitter inflation using [D-branes] in string theory. So I was pursuing that direction. Extra dimensions, de Sitter inflation, these kinds of things.

Zierler:

I'm curious how Henry's experience in cosmic inflation may have been useful or your collaboration for him.

Dvali:

Yeah, that's how we started our discussion, actually. And we were actually going around the lake. I went there to give a talk about ADD. And then, we started to discuss inflation. And it was really very interesting. [laugh] In one lap, the idea came out. And then, we said, "Look, let's do it with D-branes, and orientifold planes, and stuff like that." And then, we wrote the paper pretty quickly. Because the idea was very simple. The idea was extremely transparent and simple. Because D-branes have this property that they give literally de Sitter-like source. And you have them temporarily, and then they collide and annihilate. Of course, we had to make some assumptions, like I said, about stability and stuff.

Zierler:

And was this research useful for broader questions about understanding the very early universe?

Dvali:

Oh, I think so. Absolutely. It started as thinking about de Sitter in string theory. Then, Henry moved towards this very nice idea about cosmic strings being produced by this D-brane inflation. And I continued more towards trying to do real engineering about trying to understand stability. So I wrote some papers, some of them alone, some with Qaisar Shafi and Slava Solganik (my student at the time), and others. And so, this understanding that string theory cannot tolerate de Sitter came from this experience. You know how it works. Naively, it looks like there must be a plentitude of possibilities. And then, you start exploring and try this, try that, and none of it works.

And then, some very general reasons emerge, underlying mechanics why these things do not work. And then, finally, you realize that there must be something more fundamental under this. This is the fact that string theory is formulated in the theory of s-matrix. S-matrix is this idea about having something in the infinite past, coming in, colliding, and going into something into infinite future. So there ares-called in-states and out-states. And I can give you a dictionary, a book, and in this book, you can find which in-states go into which out-states. And that's all you need. And so, that's the way the string theory is formulated. There's no other formulation of the string theory we have. And I think this is true in general for quantum gravity. It's an s-matrix theory.

Now, the problem is, in de Sitter, you cannot have s-matrix. There are simple scaling arguments that tell you that you cannot define s-matrix in de Sitter for gravity, even approximately. So part of this was known, of course. Actually, it was well-known that you don't have a globally defined time in de Sitter, therefore, you cannot define s-matrix. But probably what we were thinking was, "Maybe we can still define s-matrix approximately." It doesn't work like that. Because this is really fundamental. I don't know how to say this. The only time we do approximate computations, in physics, in life, everywhere–that's OK, you can always do approximate computation. We never solve problems exactly.

However, approximate computation makes sense if the framework in which you are is well-defined. If you don't know whether that framework is well-defined, the limits do not commute and you cannot really trust approximate computation. Approximate computation is no guarantee for the validity of anything. And this is extremely important. Another example of this is gauge anomalies. It's the same thing because you very naively say, "Look, let me introduce the fermion in the Standard Model with a tiny hyper-charge. OK, there will be an anomaly. But approximately, I can work with the Standard Model." It doesn't work because there is a discontinuity. There's no continuity making sense.

You cannot change the notion of making sense continuously. It's a discontinuous notion. Either something makes sense, or it does not. And there are many examples of this type. A slight change of the parameter, and the theory stops making sense. And this is what happens with de Sitter. And this is extremely important. Because of course, this still needs to be fully appreciated and understood. And I'm trying to push for it.

Zierler:

One of your most prominent publications with Nima and Savas, of course, was in Physics Today, Large Extra Dimensions: A New Arena for Particle Physics. I'm curious if part of the motivation in that paper was sociological. In other words, getting the word out to people who had a background in particle theory that these were some very interesting areas that more people should work in.

Dvali:

Yeah, this article belongs to this category of ``popular” articles. So we tried to sort of popularize the framework. I think we got the invitation to do that. Maybe Nima or Savas was the main connector. But anyway, in general, I think it's very useful to write popular articles to really sort of popularize the research from a firsthand perspective. That was the idea.

Zierler:

What was the relevance around this time of the discovery of dark energy in the accelerating universe for you?

Dvali:

So around that time, there was this discovery of dark energy. And we even have some comments in one of our papers that, "OK, because we now have these large extra dimensions, maybe this can change the story." And then, we independently started pursuing this direction. Actually, Savas, Nima, and I think it was Raman Sundrum, and maybe Nemanja Kaloper, they wrote one paper in that direction about this millimeter dimensions, how to explain the smallness of the cosmological constant in that direction. And I started thinking in a little bit of a different direction with my other collaborators, Greg Gabadadze and Massimo Porrati.

So at NYU, we went in a slightly different direction. And then, with Misha Shifman, especially. We said that maybe the cosmological constant was small because it was offloaded somehow, spread in infinite extra dimensions. So for that, large extra dimensions have to be infinite. So you really need an infinite volume extra dimension outside of where we live. And so, we pursued that direction. At the end of the day, nothing worked. So there are very interesting ideas, but there's always some issue with the constructions. There are always some assumptions one has to make.

Zierler:

What was some of your research on the cosmological constant at this point?

Dvali:

It's basically that. So there were sort of two approaches. One thing was, we came up with this idea to explain dark energy without the cosmological constant by self-acceleration of gravity. So we tried to modify Einstein gravity at large distances. And the idea was that maybe the observed acceleration is because of large distance modification of gravity. So this idea was put forward together in a paper with Massimo Porrati and Greg Gabadadze. One of our papers. And then, we tried to work this through explicitly. We produced a model of infinite extra dimensions. And then, we tried to work it through. And then, together with other people like Cedric Deffayet, who was a post-doc at NYU at that time–and indeed, we saw that there were such solutions. So it's an interesting effect. So you don't have a cosmological constant, but gravity gives self-acceleration.

So gravity itself, because of modification at large distances, it acts like a source for itself. It is sort of a graviton condensate, an extremely soft graviton condensate that gives you this acceleration and the effect of dark energy. Or you can say it differently. You can say that in this approach, dark energy is given by graviton condensate. So you don't need additional sources, so the gravity self-accelerates. And this general idea triggered a lot of interest. Now, there are many different approaches that are based on the same general idea of modifying gravity at large distances. Many interesting ones. That original proposal had issues observationally, and there were some consistency issues at the quantum level. I have seen many papers where people are working in this very general direction.

Zierler:

I'm curious of you got involved at all in axion research.

Dvali:

An axion is such a thing that it's very hard not to get involved in. [laugh] It's absolutely fascinating. The axion is an example of a field theory construction that is extremely useful from many different perspectives. For example, when I give effective field theory classes, you can use the axion for many different perspectives. It's a fantastic illustration of effective field theory. Because it brings together perturbative aspects, non-perturbative aspects, naturalness points, everything. It's just an extraordinarily interesting framework. So I was interested from many different perspectives. Like, for example, one idea is that the axion could be dark matter in the universe. So what we see as dark matter is basically an oscillating axion field. Many, many non-relativistic axions on top of each other.

So one thing was how to make the axion scale larger because there was a standard cosmological bound on the axion scale. And I had the idea to remove these bounds when I was a post-doc at university of Pisa. My idea was that, maybe in the early universe, QCD scale was different. So in the early universe, strong interaction scale was much stronger. Then, the axion could be settled to its minimum already earlier. And so, in that case, turns out this would relax the constraint on the scale of the axion decay constant. Anyway, the axion is sort of similar to the graviton. So the graviton interacts through the interaction that is suppressed by the Planck scale, or Newton's constant. So there is an analogue constant for axion.

So we're talking about the interaction constant for an axion. And then, later, I got more interested in the field theory aspects of it in the sense that it turned out there is a beautiful language in which you can understand axion based on so-called free-form language. In this language, you can understand the generation of the axion mass by QCD effects very transparently, and it's very similar to the way mass of so-called eta-prime meson is generated by QCD effects. So that is like more field theoretical aspects. Maybe you were asking me about experimental research of axions. There, I'm just an enthusiast. [laugh] Nothing else. At our institute, we had a retreat, and a colleague of mine, Allen Caldwell, was initiating this question about whether it would be interesting to look for the axions experimentally.

Zierler:

If you had to guess, do you think that axions is the best path forward for understanding dark matter?

Dvali:

No, that's very hard to say. Also, it's very hard to predict where the scale of the axion physics is. Could easily be as high as the Planck scale. In that case, these experiments would not see anything. For example, (with such a high scale) also this [MADMAX] experiment will not see anything, unfortunately. So it's very hard to predict the scale. It's a very elegant dark matter candidate. That's what we can say. So in general, in physics, as I said, we have very few guidelines. Now, once you distill only consistent theories, within consistent theories, you need to use some guidelines. Now, one of the guidelines is in cross-correlations.

So I would say that if we introduce some new physics beyond the Standard Model, if this new physics is motivated for more than one thing, that is more valuable. So that's the case for the axion. Because the axion is motivated by a completely different problem, a so-called strong-CP problem. And as a byproduct, it gives you a dark matter possibility. So from that point of view, it's highly motivated. That's what I would say. But dark matter is such an open field because the problem with dark matter is, we essentially only observe its gravitational properties. Since we observe its gravitational properties, we know that, with respect to gravity, it has to behave like matter, and with respect to ordinary interactions, it has to interact extremely weakly.

So it should not have ordinary electromagnetic properties, etc. So beyond that, we don't really know much. So therefore, pretty much anything that gives you cold bosons or cold matter particles would satisfy this condition. For example, black holes. They could be an extremely interesting component of dark matter. Primordial black holes. And there are other candidates. But certainly, the axion is one of the most elegant ones. That's for sure.

Zierler:

What do you see as some of the long term contributions from the collaboration with you, Gregory Gabadadze, and Massimo Porrati, the DGP model?

Dvali:

Yeah, that's the model I mentioned. This example is for self-acceleration. We wrote the paper simply trying to come up with a theory which would modify gravity consistently at large distances. And then, we used it as a prototype model for self-acceleration together with Cedric Deffayet, etc. And I think it's a nice prototype model. So it's simple enough to sort of illustrate the main points about the idea, both about how to modify gravity and how a modification of gravity could lead to self-acceleration. So from that point of view, it's a nice prototype model. Beyond that, as I said, it was later realized that it has problems. If we want to use it for self-acceleration, then it exhibits certain problems with certain instabilities. And in the form we wrote it, it cannot work. It requires some modification.

Zierler:

When you joined the theory group at CERN, did you take a leave of absence from NYU? Or was this a dual appointment?

Dvali:

No, I took a leave of absence from NYU, but I was going back and forth. I was also giving lectures at NYU.

Zierler:

And was this the time that you became more fully involved in black hole research?

Dvali:

Yes. I started to be more fully involved. As I said, we were involved in black hole business from the beginning of ADD because for example, we predicted this formation of quantum micro black holes in our second paper, with Ignatios Antoniadis, Nima, and Savas. So already in that paper, we predicted the formation of micro black holes at LHC. But I became more involved later. This was certainly before moving to CERN. For example, I got interested in black holes because black holes are extremely useful objects for putting consistency bounds on theories. For instance, I've shown, before I moved to CERN, that if a theory has a certain number of particle species, then consistency of black hole physics gives a bound on the scale of quantum gravity.

So actually, it turned out that this is more general than the relation of ADD. In ADD, the reason why the scale of quantum gravity is lower can be understood in this language. Because in ADD, because compactification volume is large, there are a large number of Kaluza-Klein species. Actually, the number of Kaluza-Klein species in ADD is 10 to the power of 32. And this is precisely the bound that a black hole gives you. So therefore, it turns out, it's much more general. In other words, if you give me any theory with 10 to the 32 species, that theory should solve the hierarchy problem because the quantum gravity scale in that theory cannot exceed energies around TeV region. Of course, that was research on black hole physics. So it was before CERN.

Zierler:

And generally at this point, are you operating on the basis that string theory is the best path forward to developing an understanding of quantum gravity?

Dvali:

No, as I said, string theory is a good reference point because there are certain things–for example, string theory is also good for testing certain ideas. For example, this theory about black holes that I mentioned. We have this theory about black holes, Cesar Gomez and I, that a black hole is a saturated, high occupation number state of soft gravitons. Gravitons of the same wavelength as the size of a black hole. Now, one of the tests of this idea was to look for a scattering process. It is known from a long time ago that if we scatter two highly energetic gravitons, we produce a black hole. For example, if you scatter two gravitons with solar center of mass energy, then the outcome of the scattering will be a solar mass black hole.

Why? Despite the fact that the scale of the center of mass energy is hugely larger than the Planck mass, you can reliably conclude this. Why? Because it's controlled by long distance physics. Because when these two gravitons approach each other at (approximately) three kilometer radius, which is the Schwarzschild radius of a sun, they form a black hole. And that's it. It doesn't matter what happens next. You know that you formed the black hole. So this was a good point to test our idea. Because our idea was that a black hole was a multi-graviton state. Then, if you compute the gravitational amplitude of two gravitons into many, this somehow should capture information about black hole formation. And we did this computation with several of our collaborators (Gomez, Isermann, Lust, Stieberger) . So we indeed recovered the (entropy) information about black holes. But then, as a consistency test, we did exactly the same computation within string theory. And we recovered the same results.

So in other words, string theory is a useful reference point for testing certain ideas because computational power sometimes in that domain in string theory is much higher. In other words, in physics, we always do what? In science, we always compute where we can. There is no single theory in the universe in which you can do all the computations. Even in quantum electrodynamics, we cannot do all the computations. Of course, we always have domains where we can be in control of computation.

So this is what we do. And sometimes, this domain is very well-understood in string theory. And then, we can use string theory as a reference point. Now, in more general terms, high energy quantum gravity is a theory of extended objects. Whether you want to call these strings or black holes is a different story. But there is some intrinsic stringiness in high energy quantum gravity. Now, if you ask me if the way we know string theory today is the final one, I would say no. We have a long way to go.

Zierler:

Scientifically, even emotionally, what were your reactions following the discovery of the Higgs when so little else has been seen at the LHC?

Dvali:

Discovering the Higgs was, of course, an extraordinary success of theoretical physics. It's a funny situation because of course, it's an experimental success, but at the same time, it's an extraordinary success of theoretical physics. Why? Because the existence of the Higgs was predicted theoretically just by consistency, by unitarity. So we knew that either Higgs or something even more complicated must come into play within the range of a few hundred GeVs in order to unitarize the theory. Because without this, the Standard Model would not make any sense. And this cannot happen because nature should make sense. So this was a great confirmation of that.

"OK, the simplest calculable possibility worked out." That was, of course, an extreme success of theoretical physics. I think it was unappreciated by the general public, the magnitude of the story. Despite the fact that there was a lot of publicity, I think it's still very hard to convey how fundamentally important this discovery was. That's my impression. Because it was mostly understood as, "OK, yet another particle has been discovered." No, this is not yet another particle. This is a thing where extraordinarily deep concepts come together like Goldstone's theorem, Higgs effect, unitarity, and so on. Most fundamental things in physics go through this. So it's an extremely special particle.

Zierler:

When did you start thinking about the connections between black hole information and quantum computing?

Dvali:

Oh, that was a few years ago. After Cesar and I wrote this theory of a black hole as a condensate, then we made the connection with calculable condensates, and we showed that certain properties there are very similar. So after that, I wrote a couple of papers, in particular, with a student of mine, Misha Panchenko, and there is this mechanism of how to store information in a condensate in a black hole type way. Never mind whether my theory (as a theory of a black hole) is correct or not. It's independent of whether this is applicable to black holes. So black holes gave me the understanding of how this mechanic can work, how a given system can enhance memory storage capacity. Now, once you have the mechanics, if it works in black holes, I'm 100% sure that that's the way black holes work, but for quantum computing, just take this mechanism, and try to use it in a laboratory. Because now it's an explicit microscopic mechanism.

So I see very clearly how saturation is connected with maximal capacity of information storage. Of course, I'm not an experimentalist, and most likely, to do this experiment, because of the large number of particles, is very hard. And I'm not necessarily talking about technological applications, although I don't see anything fundamentally impossible in that. But of course, this is (primarily) about understanding, so that you can do, literally, this type of information storage in laboratory systems. As I said, I even wrote a paper trying to write this (black hole model) as a neural network. So yeah, I had some fun conversations with people from biology and such. You can always imagine that somehow a human brain also finds some sort of similar mechanics for maximizing information capacity.

Zierler:

What was your reaction when the Event Horizon Telescope produced the image of the black hole? What did that mean for you?

Dvali:

Oh, of course, that's an extremely important step, especially for someone thinking about black holes. Now, that's a little bit similar to the Higgs story. In fundamental physics, nobody has any doubt that black holes exist. Because Einstein's theory is extremely well-tested at those distances. So we know, for example, that Einstein's theory works extremely well at earth-moon distance. Correspondingly, obviously, since it predicts black holes of earth-moon size, these black holes must exist because the theory is extremely well-tested at those distances. So in our community, of course, nobody has any doubt that black holes exist. But of course, it's always a different level to see explicit pictures. So emotionally, it's very intense.

Zierler:

Because even as a theorist, seeing it somehow makes it more real to a degree.

Dvali:

Well, I don't know. I don't know whether that's the satisfaction, that you see something is real, or actually it's the opposite, that you say, "Look, we knew this was real. You see? This is real." That you sort of confirm your longtime understanding. So it confirms your longtime understanding. It's not that you never thought about it, and now it's surprising. "Oh, look, there's a black hole." No. You knew that black holes were there. And, "OK, here you go." Something like that. [laugh] It's a triumph of theoretical thinking.

Zierler:

And what about LIGO? What did that mean for you, the detection of the gravitational wave? Did you understand it in the same way?

Dvali:

Yeah, absolutely. In the same way. Of course, this is an extraordinary experimental work, but it is a triumph of theoretical thinking. Because something that Einstein predicted 100 years ago, when this possibility was not even on the horizon, and now it gets confirmed. It was predicted theoretically without any experimental input, period. Newton says, "OK, there is a field." OK, static, but still, Newton has a messenger. Then, Einstein says, "This is geometry. There is a messenger, and it's geometry. And then, this geometry has excitations. And these excitations of geometry propagate as waves. This is all theoretical. Of course, Newton is not theoretical because the planets were there. But this thinking that the excitation of geometry must propagate as waves because you cannot have rigid geometries is an extraordinary theoretical discovery and thought. And this is what got confirmed after 100 years. So it's absolutely a triumph of theoretical and experimental physics.

Zierler:

Tell me about the Mad Max collaboration. How did that get started, and what are your contributions?

Dvali:

My contribution is probably next to zero, if not zero. [laugh] As I said, my only contribution is that, at the beginning, I was enthusiastic about it. And that's the contribution. Literally, my only contribution is that when my colleagues were discussing this possibility, I was participating in discussions, and as a theorist, I told them that it makes sense to look for the axion scale beyond the standard cosmological window. But you don't need me to say that. That's it. Once the data starts coming out, maybe later, I can contribute something more significant. But at this stage, it's mostly hardware.

Zierler:

On the administrative side, what have been some of your key achievements as Director of the Planck Institute in Munich?

Dvali:

Actually, the director position is not really an administrative one. It's a scientific position. Formally, yes, you are head of a department, and formally, you have to sign some papers now and then. And of course, you have to participate in decision-making and planning. But it's mostly scientific. So the main issues are standard things, like hiring post-docs and faculty members, supervising students, organizing things. The administrative part is not so heavy. Also, the administration is pretty efficient.

Zierler:

Now, up until 2019, you were shuttling back and forth between Europe and New York?

Dvali:

That's right, yes. So we moved to Munich because my wife is also a scientist. She's a biologist. And as you probably know, two-body problems are very difficult in scientific families. It's very hard to find two positions at the same university. And we got such an offer from Munich. So we moved to Munich. Of course, there were other reasons. Munich is a fantastic place with resources and a scientific atmosphere. It's great. So we moved there, and until 2019, I was going back and forth between NYU and Munich. And again, I was on unpaid leave, but got some advanced lecture courses.

Zierler:

And what have been some of your more recent interests in grand unification?

Dvali:

I came back a little bit to grand unification very recently, to supersymmetric grand unification. Because we're doing this project with a graduate student, Anna Jankowsky. And the interest is, again, this version of the hierarchy problem within supersymmetric grand unification. So I had this idea very long ago, when I was a student. In grand unification, there is the so-called doublet-triplet splitting problem, a famous problem. The problem is that one of the successes of grand unification is, the Higgs doublet, which comes as a doublet of weak interaction, becomes accompanied by a partner. Now, this partner is a colored state which has the quantum numbers of a down quark. Now, this colored partner is a triplet on the color group, so we call it a color triplet.

So there's a Higgs doublet and a color triplet. So the color triplet, because Higgs interacts with ordinary matter, this color triplet also interacts with ordinary matter. And it couples both to quarks and to leptons in the way that violates baryon and lepton numbers. So it can mediate processes of proton decay. Now, this color triplet, if it were as light as the Higgs doublet, it would mediate proton decay instantaneously. And so, this is called doublet-triplet splitting problem. And so, in the traditional approach, the way this problem is solved is that the triplet is made much heavier than the doublet. And my idea was to do the opposite. Instead of making the triplet heavier, to keep them degenerate (in masses) up there, but split their interaction strengths, so to split their couplings.

And so, now, we end up with a doublet and a light triplet, and this triplet is decoupled from the quarks and leptons. And so, the idea is to study the phenomenology of this triplet. It's quite exciting because this triplet can be a remnant of grand unification that (can be seen) at low energies. Surprisingly, and unfortunately, it's not very widely explored, this possibility. I don't know why. Probably, it's not that known. At that time, there was no archive when I wrote the paper. I published it in Phys. Letters. But there was no archive. And so, now I came back to it together with a Ph.D. student of mine.

Zierler:

And so, in your current responsibilities, do you have the normal responsibilities as a professor? Are you teaching undergraduates? Are you supervising graduate students?

Dvali:

Oh, yeah, I'm supervising an enormous number of graduate students. Actually, I counted the other day. I think 11 or 12.

Zierler:

So Munich is a great place to be a physicist.

Dvali:

In Munich, you get possibilities to supervise that many students. So yeah, I have resources to supervise that many students. But this number is too high, actually. I have 11 or 12 graduate students. PhD students, I mean. And then, also, quite a large number of master's students. And so, this is a really large number. So I think I'm doing too much. But yeah, you get possibilities for that. And it's great.

Zierler:

Well, now that we've worked all the way up to the present, for the last part of our talk, I'd like to ask a few broadly retrospective questions about your career, and then we'll end looking to the future. One thing we haven't talked about yet is the role of computers in your research. In what ways has the rise in computational power been relevant for complex equations and things like that?

Dvali:

Well, that's a great question. So I, myself, belong to people who prefer analytic computations. I always try to do some analytic estimates and computations analytically. But of course, there are certain problems that have to be solved numerically. For instance, we encountered those a lot trying to model black holes by these condensates. And there, the computational power is extremely important. So yes, we are using full scale computational power. This is very important.

Zierler:

As you well know, there's a criticism of string theory that it has become decoupled from the world of experimentation. Do you remain hopeful that string theory might one day be experimentally verified and show us something about how the real world operates?

Dvali:

Exactly. I just gave an example of precisely the opposite because I'm predicting that in string theory, the cosmological constant is zero. I'm making a prediction of exactly zero. And my point is that string theory at least makes this exact prediction. Therefore, yes, that is an experimental prediction. And that's why I'm very hopeful that our observer friends will tell us relatively soon whether dark energy is there and how different the equational state is from lambda. That's an example of the complete opposite. What I want to say is that sometimes, it happens. You say, "OK, we have a theory. And there is a certain problem." Like, in this case, there's the cosmological constant puzzle. And so, it is true, for a long time, the idea was dominated by anthropic solution.

The anthropic solution is exactly what people are criticizing as lack of predictivity. Why? Let's say I want to predict something, and you are saying, "I have plentitude of this parameter," and somewhere in the universe, the value is realized, and that's the value where we happen to be because that's what is compatible with the existence of our life (forms). Even there, there are people who are pursuing this scientifically. For example, Weinberg illustrated that this idea can be somehow scientifically studied. But now, we have a theory of quantum gravity, and the question is, which is the path that this theory is taking? And yes, there was this assumption that theory's taking that (anthropic) path. So it has a plentitude of de Sitter vacua, and we happen to be in the right one. What I'm saying is completely opposite. I'm saying there are none.

String theory can not tolerate de Sitter because its formulation is based on fundamentally different principles. They're incompatible. And that's the predictivity of the theory. It's very important to somehow change this thinking. For example, I had many discussions with my friends who come more from a general relativity background. Because there are these two opposite views. I'm not talking about personalities, I'm just talking about points of view. I can put myself in the shoes of someone coming from general theory of relativity. And in the general theory of relativity, you say, "There is nothing special about Minkowski space." The space with zero cosmological constant. "There is nothing special about this space. Because in general theory of relativity, there are infinite number of solutions. And Minkowski is only one out of infinitely many possible backgrounds." But here, it's besides the point because quantum gravity is more powerful than the general theory of relativity.

The general theory of relativity is the low energy limit of quantum gravity. So obviously, it should not be surprising that more powerful theory is more selective. Therefore, not every background in general theory of relativity can be promoted into a consistent vacuum in quantum gravity. And that's the great thing about quantum gravity. It's very selective. It can only work for very special situations. And that's what predictivity is about. So therefore, I don't think there's a lack of predictivity. I think, simply, we still have to understand the theory. And once you understand the theory, it will be predictive. I take the same point of view about other parameters of the Standard Model.

Once we're really able to derive theory from more fundamental theory, whether it's quantum gravity without necessarily one of these known string theories, or some version of string theory, or no string theories, once we will really understand how to derive, that derivation will be predictive. That's my point of view. Because this is what I'm observing regarding de Sitter. Once we understand what it is, then we see that it's predictive. Lambda equals zero.

Zierler:

Do you think the term theory of everything is realistic and applicable?

Dvali:

Well, that's just a term. Physics is about understanding the nature of phenomena. We want to understand certain systems. And we always design a theory for understanding a that (particular) system. And theory always has domains of applicability and calculability. Now, of course, you can always define the theory of everything as a concept. But as far as calculability is concerned or predictivity, there will always be domains of applicability. Take string theory. In string theory, we know how to compute certain things in the weak coupling domain very nicely. In the strong coupling domain, we also know how to compute certain things using some dualities. And then there are regimes in which we cannot compute things.

Therefore, when you say theory of everything, that is just a concept. It requires a definition. In terms of calculability, I don't think we can–as I said, even quantum electrodynamics is a theory. It's a self-contained theory. But even there, we cannot compute everything. For example, the perturbative (series) are only asymptotic, so people are inventing new ideas (for going beyond perturbation theory) and stuff. Even within a theory, once you define one, there will be a permanent struggle. So let's say today, we all agree that we have a final theory. That doesn't mean that we can simply take and compute whatever we want. This will be a permanent struggle, even in a known theory, to compute things and make predictions. And that, I think will (never) finish. I hope it will never finish, that struggle.

Zierler:

In all the things you've worked on, do you feel like you know less about the subjects as a result of learning more about them?

Dvali:

Well, there is a component of that, that once you understand something, you realize that now, there's a new horizon. It's like seeing a new horizon. It's like climbing a mountain. You see a ridge, and you simply don't see what is behind. Then, eventually, you've climbed that ridge. You should not diminish your achievements. Understanding is always a pleasure and should not be underestimated. But of course, after every understanding, there's a new horizon. And typically, these horizons are larger. So I agree, there is this component in physics research. And that's also what makes it fun. Absolutely.

Zierler:

There's so much happening observationally right now with new telescope projects. What is most exciting to you? What are you most looking forward to seeing with all of the data that's coming out and will be coming out over the next decade?

Dvali:

Well, actually, all of them are very important. Because in fundamental physics, we really need observations at very different length scales. Unfortunately, one observation performed with one particular length scale, or let's say one gravitational wave experiment for some wavelength, usually is not sufficient for opening up new ideas beyond the Standard Model. Of course, one experiment can (also) be decisive if there's a prediction and the theory's already very well worked out, like what happened with gravitational waves. That's different. But typically, that's not necessarily what happens. For example, take dark energy. For dark energy, of course, originally the hint was given by this extraordinarily interesting observation about the supernovas.

So the standard candles. So it came from supernovas. But of course, just supernovas are not enough. You need CMB, cosmic microwave background observations and other observations. You have to put everything together. And so, from that point of view, we just have to keep exploring nature. Of course, there is a preliminary stage before an experiment gets approved. And of course, a lot of theoretical and experimental discussions are going into that stage. Obviously, the funds are finite, and we have to optimize our research. And of course, we have to set priorities. But once an experiment is approved, it's usually always useful. Usually, nothing is wasted.

Zierler:

Last question, looking to the future, you mentioned passing horizons and seeing new mountains. So as you look ahead, what are they? What are the most important mountains that you want to see in the rest of your career?

Dvali:

Well, usually what you see are just directions. I think it's also typical for others. You always think the most important thing is what you're doing at that moment. Because you're excited about that particular thing. So at the moment, I mean, (I am) excited about a few things along those directions, like black holes and stuff. For example, one thing that I really would like to see are the simulations of this mechanism that was proposed for black hole information storage within a laboratory setting. So that's one thing I'm looking forward to. I'm discussing with some experimental colleagues that possibility. So that's one thing. And the other things are, I want to work through the implications of this absence of de Sitter in string theory and quantum gravity because it has many implications also for early cosmology.

Because it predicts new imprints from the early epoch of inflation, quantum information imprints, into the present fabric of the universe. And we need to look for those. And this has to be worked through. So that, for example, is an important technical project. Also, I'm trying to understand the implications of the saturation I mentioned. As I said, this striking observation that for all the saturated systems, their entropies are like black hole entropies. It scales like area, and they behave like black holes, etc. and I want to push forward with this to understand the quantum and theoretical implications of this. And many other things. [laugh]

Zierler:

You'll find new things that you haven't even thought of yet.

Dvali:

Probably. I hope so.

Zierler:

Gia, it's been so much fun spending this time with you. I want to thank you so much for doing this and sharing your insights.

Dvali:

Thank you very much.