Will Scientists Ever Agree on Quantum Theory? | Sabine Hossenfelder & Matt O’Dowd (Full Talk)

If you measure one of the entangled particles, something physically changes about the other particle. This is exactly what they questioned. They were saying, no, this can't be because that would not be local. Therefore, there has to be an underlying element of reality which actually did determine the outcome before you made the measurement. So, hi everyone. You know that dream when you're like a kid, a five-year-old kid, and they put you on a stage and there are all these phases you're supposed to talk about like quantum mechanics or something. I think this is part of the design of this particular event because it's just going to be chaos and unconstrained, but I think that could be a beautiful thing. So, this five-year-old is Sabina Hossenfelder, who's vastly more qualified as a theoretical physicist and extremely wide-ranging in her thinking and her research. I, on the other hand, am Matt O'Dowd, just a regular five-year-old. I'm also an astrophysicist. Technically, I think about black holes and extragalactic stuff. We're both YouTubers though, so we have something very much in common. We do like to talk about the universe on the internet. Today, our brief is to talk about the quantum unknown, most of which is unknown to me. Quantum mechanics came into being 100 years ago, this decade, very close to 100 years ago. It's been one of perhaps the most successful theories in physics with massive predictive power. Yet, as they say, we still don't know what it means. One thing I actually really wanted to do because I'm curious about it is to ask Sabina about your first experience of quantum mechanics and your realisation that maybe this is something you want to do for the rest of your life. Would you mind sharing? I came to Quantum Foundations through a detour through quantum gravity. I learned quantum mechanics like most students those days, I guess, from a professor who was like, don't ask what it means. This is the math. Don't ask about interpretations. Stop thinking. It was literally this shut up and calculate attitude that we were given. Of course, some people thought there's something really deep going on that they're not telling us. But mostly, this didn't go anywhere. I did my PhD on particle physics, quantum field theory. All those foundational questions, they're still there, but they're covered up by all this mathematics. Then I became interested in the question of how to quantise gravity. I worked a lot with people who did string theory and loop quantum gravity and other things. This wasn't going anywhere. I arrived at the conclusion that there's probably a path that we missed early on. That led me to be interested in those foundational questions because I still think now that the reason we can't figure out how to combine general relativity with quantum mechanics is not on the gravity side, but it's on the quantum side that we haven't really managed to make sense of what's going on there. Yeah, that's not the consensus opinion with regards to gravity. My first experience with quantum mechanics was just an undergrad doing physics, not really knowing what I wanted to do. I just realised that there was stuff here, there was meat here to get into. That starts with a lot of Newtonian physics, very clockwork-y physics. You have these equations, they say how particles move, those particles interact with other particles, and you have this, as it's been called, this clockwork universe. Very dry, beautiful, but very dry and predictable. Then I took my quantum mechanics class. I remember hearing about the double-slit experiment, which maybe I'll describe in a second. It was like this little crack of light opened in this very mechanistic world. I realised at the time, I thought, oh, magic is real. I feel like this was my letter to Hogwarts. Of course, quantum mechanics is not magic, but it was this peering beyond the intuitive and the stuff that for all my life felt just this regimented thing that physics was. Then I went to do astrophysics because quantum mechanics turned out to be very hard. I think there is this mystical quality to quantum mechanics, which is why maybe it felt like that to me. There's this totally unanswered question at the very bottom of quantum mechanics that defies intuition. I think at the heart of that is the measurement problem. Would you maybe agree? Yeah, definitely. There are a lot of other things that people like to talk about, like entanglement. I'm always like, I'm not bothered by entanglement. We can deal with mathematically, but the measurement process, we just don't know what it is. We can't even say what the measurement is. It's kind of embarrassing, really. The interesting thing about the measurement problem is that we're getting there experimentally. Maybe to aid the audience, we should explain what the measurement problem is. You take some particles, it could be photons, electrons, whatever, and you throw them at a wall, a very thin wall. You see dots on a wall, and then you cut two slits in the wall, and you keep throwing. Some of the particles will go through the wall, and you have a screen on the other side. You see dots where your particles land on the other side, and you would expect maybe some to pile up because they went through one slit, some to pile up because they went through the other slit. That's not what we see, at least for a quantum particle. What we see is as you throw more and more particles, they form these bands we call interference bands that don't make any sense on the surface, but they look awfully like what would happen if a wave passed through those two slits. The wave would come out the other side, producing ripples that would interfere with themselves, build up or die away, and you would get this lovely pattern on the wall. This happens even if you throw, one electron at a time. The electrons don't know about each other. The standard interpretation these days is that that little speck of matter is simultaneously passing through both of these slits, reaching the screen, not as a particle, but as this thing we call a wave function, so a probability distribution of where that electron might be if you were to look. At the final screen, you look and the electron appears here, or appears here, or appears here, but the pattern built up is this wavy distribution that seems to only make sense if the electron was a wave passing through and then we say collapsed into a single point particle at the screen, at the measurement. The measurement problem is that how can the electron be in this fuzzy state of a probability space, a wave function, all the way up until we look at it and then suddenly resolve itself into one position? That's the double slit experiment, very crudely told. Can you say more about why it's weird that the electron would do that on measurement? Yeah, so there are two parts to this weirdness. One is what Einstein already worried about, so if you have this wave of the electron which goes through both slits, it spreads out towards the screen. But then as Jude just said, like when we actually look where the electron is, it's only in one place. So somehow the wave function needs to go away, like very suddenly, on the entire screen, and that's not local. So it has to happen instantaneously everywhere, which is faster than the speed of light. And this is what Einstein called a spooky action at a distance, because if you just look at the mathematics, it just goes snap, and it's in one point. How's that supposed to happen? We don't know the process. So we have mathematics, which actually just makes this snap, basically. But is that what is physically really going on? And the other thing that's weird about it is that what happens depends on what you look at. Because with the double slit, as Matt just said, if you have only one slit, you don't get interference pattern, you just get a pile of, you know, in one place, basically, it's one bump. So that brings up the question, like, what happens if you know which of the slits the particle went through? And the answer is, the interference pattern goes away. And that brings up the following problem, which is, how does the particle know that you will measure which slit it wants to go through before it gets there, because it needs to get on the way? Like, it needs to decide whether to aim at both slits, and then interfere with itself, or whether to not do that, because you will look at it. For me, this is like the weirdest part, like, how does the particle know that you will sit there and look at it? And that question is open to the audience. That's where we've been stuck for 100 years. So the standard interpretation is the Copenhagen interpretation, which really just says that things exist as wave functions until a measurement is made, and then the wave function collapses and precise values happen. I call the Copenhagen interpretation the non-interpretation interpretation of quantum mechanics, or the shut up and calculate interpretation, because it's not really an interpretation. It's not telling us what's happening when a measurement's made. It's just saying this crude heuristic of the wave function collapse just works. But we haven't gone further. There are a few different options for what's really happening at that moment of observation, and they include ideas about what this pesky wave function really is, what the relationship of the wave function is to the speck of matter, the particle that we know and love. I think the Copenhagen interpretation was basically Bohr's idea. I mean, maybe some historians would object, but I think this is basically where it came from. He said, let's just not ask what's going on. This isn't real. We're not supposed to ask that question. Shut up and calculate, basically. And then there are... Bohmian mechanics, I think, is the one that is easiest to understand in terms of good old-fashioned physics, because it brings back this particle idea. So Bohmian mechanics basically says there is some sort of guiding field that tells the particle where to go. So that's what's called a pilot wave theory. The issue with this idea is that you can make it work, but it will not be local. So you still have this weird influence from some particle on one side of the room to another particle on the other side of the room. So it's kind of... it's also an interesting question in how far those two are actually different mathematically. But yeah, let's leave this for another time. And so I think the best known interpretation after Copenhagen is probably the many worlds interpretation, which says, loosely speaking, that whenever you make a measurement, you create a new universe. So all possible outcomes actually do happen. So the particle goes to each point on the screen, but for each possible outcome, you get a separate universe. And, you know, personally, I think that this is really just an interpretation. It doesn't really solve any problem, because the measurement problem is really a mathematical question. So you want to be actually able to calculate what is a measurement. The many worlds interpretation doesn't do it the same way that Copenhagen doesn't do it. So I could push back on that, but maybe I won't just yet. So a broad difference between some of these ideas is, as you hinted at, the idea of locality. So can different parts of this wave function know about other parts of this wave function? But another one is what we call realism, which is just kind of like, what does the wave function really represent? Is it just a representation of our knowledge? Is it a physical thing? I'm trying to somehow get to a proper way to talk about super determinism here. And so the non-locality is key. So there are some experiments you can do. And when Einstein coined this phrase, spooky action at a distance, he actually wrote a paper talking about another quantum phenomenon, entanglement, where two different particles can have these completely fuzzy properties, so fundamentally undefined properties, but nonetheless be correlated with each other. So their reality exists in some correlation between them, not in the particles themselves. At least that is as predicted by standard quantum mechanics. And Einstein, Podolsky and Rosen showed that if that's true, then you can take these two so-called entangled particles very far apart, light years, it doesn't matter, someone could make a measurement on one particle, and literally the way they choose to make that measurement, so the type of measurement they choose to make, causes a real physical change in the other particle. The other particle becomes defined in a way that it wasn't before. And so this is this spooky action at a distance. And I just got shivers right now thinking about it. It's weird as hell. Anyway, we've now demonstrated that exactly this reductio ad absurdum that Einstein proposed to discount the idea of that standard quantum mechanics has proved to be real. We've done these tests, and these are called Bell tests. As is usually said, they show that the universe can't have what we call local realism. The two particles that you're separating can't independently know what their own state is. The state can't be encoded within that particle. It has to be, according to the normal interpretation of this, that it has to be somehow distributed so that the resolution is able to move faster than the speed of light between these particles. Now, I know, Simin, you have a different interpretation of quantum entanglement. I have a different interpretation of what Einstein said in the paper. So maybe we'll start with this. So I think what the point was of the EPR paper was to say you have those entangled particles, one left, one right, and we assume that nature needs to be local because, of course, Einstein was Einstein. He didn't want anything to go faster than the speed of light. And by local, it just means that any influence or transfer of interaction has to travel step by step only at the speed of light, no faster. Right. So it has to go through space from one place to another, but it can't go faster than the speed of light. And so when you said if you measure one of the entangled particles, something physically changes about the other particle, this is exactly what they questioned. They were saying, no, this can't be because that would not be local. Therefore, there has to be an underlying element of reality, I think is what they called it, which actually did determine the outcome before you made the measurement. And this is what they called a hidden variable. So they said, so this entangled state, it just describes our lack of knowledge about what's really going on. But there is reality underneath. So in this sense, the wave function is not real in this interpretation. And this underlying reality is described by the hidden variables. And so now people have proved that these correlations actually exist. This is what the Nobel Prize in 2022 was for, right. And what Bell's theorem, so it's a mathematical prove about these type of measurements and correlations, tells us is that the only way you can make this work, what Einstein wanted to do, is if the theory is super deterministic. And that is unfortunately, a possibility that pretty much all physicists have thrown out, except a handful or so, of which I'm one, which I guess is why I'm here. So maybe I can just, why I think people tend to reject this notion, and maybe to give some more clarity to it. So these particles are separated. Let's say their correlation is that one has to have a spin that's pointing up, the other has to have a spin that's pointing down, but you don't know which is which. And you don't know which way the spin is pointing, left, right, up, down, forwards, back. When someone chooses to make, Alice chooses to make a measurement, the choice of direction that she chooses, left, right, up, down, influences what Bob on the other side is going to be able to see from his particle. Now, the spooky action in a distance is if that interaction happened instantaneously, but a hidden variable interpretation would say that this particle really knew what it was all along, right? And this particle really knew what it was all along. And it's just that we didn't know. And so the indeterminacy represents our lack of knowledge, not the particle's actual state. However, when you look, when you do this experiment and you look at the correlations, it really seems like the choice of measurement affects both particles. My understanding of a super-deterministic theory is that these particles have correlated spins, but because the observer on both ends is also part of the universe, their choice of measurement is also correlated with those hidden variables inside those particles. And so the person thought they were choosing measure up, down, measure left, right, but really they're all part of a universe that is going to give you the right answer. And you actually didn't have so much choice about which direction you chose to measure those particles. How badly did I mangle that? I want, and we shouldn't talk about Bell-type tests because no one would understand it. So you did it anyway. So I have to object to something that you said, which is that the measurement setting on one side does not affect the measurement outcome on the other side. You could use that for non-local signalling. They're totally independent. But I mean, in spirit, it was correct. But I think you see the need for super-determinism easier if you think back of the double-slit experiment. Because as I said, if you want to describe both, the situation where you have both slits open and you get the interference pattern, and the situation where you look at one slit and you want to describe that with a hidden variables theory, which was actually determined when the particle started travelling, then the path of the particle needs to depend on what you will measure. So when the particle starts going out, it needs to know whether you'll measure at one of the slits or whether you will measure it on the screen. And this is exactly what super-determinism is about. It says that the hidden variables, which determine the measurement outcome eventually, depend on what you will measure. And it's the same in the bell-type experiment. It's just that it's more complicated because we're talking about these different angles and correlations and whatnot. Yeah. So I guess the one objection people tend to have is that the particle heading towards the slit needs to know whether the experimenter will try to measure which path it went through or not try, which means that the particle needs to be correlated with some fantastically complex system. In this case, the brain of the scientist who's making the choice about whether or not to measure which path it travelled. And it's hard to imagine a mechanism by which that correlation could happen. So somehow the particle heading towards the slit, you know, am I going to go through both slits? Am I just going to go through one slit? That choice on the part of the particle needs to be perfectly correlated with a decision that some scientist is going to make to measure or not measure. So I'm expressing my discomfort. I hear this all the time, but just mathematically, it's not true. Like the only thing you need is the measurement setting. It doesn't matter how it came about. Like, was there someone thinking about choosing this setting that they call their grandmother and ask what not? You can make this arbitrarily complicated. You can always do this also for standard quantum mechanics. You can ask the same question. Where did the measurement setting come from? It's something to do with the brain. It's correlated with the brain and what not, but it's totally irrelevant. Like if you look at the mathematics, the only thing you need is the actual measurement setting. It's this one variable. But don't you want to be able to sort of trace a causal chain that gives you that measurement setting and tells you why that measurement setting is correlated with the path of the particle? Because that causal chain is the complicated thing that it's hard to justify. So I'm not sure it's good to talk about causality because we know that causality gets really screwed up in quantum mechanics. So this is the assumption that superdeterminism is built on. Like this is what you start from. This is the thing that you don't explain. Like every theory needs to have some assumptions about what's going on, and superdeterminism starts from this correlation with the measurement setting. This is axiomatically, you say that's there and now let's ask what can we explain with it? And the answer is, well, we can explain all the measurement results of quantum mechanics, but it's local. So we've solved this problem with spooky action at a distance. Personally, I think this is obviously correct. Unfortunately, most physicists disagree with me. What do we do when several geniuses think that something is obviously correct, but disagree with each other? For example, I think Sean Carroll is as committed to many worlds as you are to superdeterminism. Well, you do experiments, obviously, I'd say. So this is the interesting thing about superdeterminism and also a few other quantum mechanics modifications, I should say, like collapse models. So they actually make different predictions from quantum mechanics in some cases. So for example, you want to use superdeterminism to actually explain how a measurement comes about. And there are just questions that you can answer with such a model that you can't answer with quantum mechanics. For example, in which circumstances do you actually get this collapse? Exactly what is it? And the problem we have at the moment is that this often, at least for now, in all measurements that we can actually make, it's covered up by decoherence. But this is a technological avenue, which is currently very strongly being developed with quantum computers and quantum sensing and so on. They try to do error correction, they try to keep the decoherence level as low as possible, and they're building bigger and bigger things. And at some point, they'll just get into this range where they have to figure out, like, is it still a wave function? Or is it supposed to be collapsed? Just exactly what happened there? I think we just experimentally go there. Is there an experiment in particular, or a prediction that superdeterminism makes regarding collapse? It's complicated. So the brief answer is, I think, yes, but I haven't written the paper. And so it's actually quite similar to Penrose's idea with the gravitationally induced collapse. The problem is that if you put in the numbers, it's really hard to get there. So I'm trying to, like, what's the experiment that you actually need to do in the laboratory? And all the estimates that I've done, they're like five orders of magnitude away from what can currently be done, so it's not great. But at least in principle, it's possible, and I hope that we'll get there in my lifetime. So that's the best I can do at the moment. Yeah, there are worse cases in theoretical physics. About orders of magnitude we are from a possible experiment. The graviton. So that's the interesting thing about it. For me, the interesting thing about Penrose's model is this link to gravity, because he basically says it's ultimately something to do with the gravitational interaction. And I think he's got a good point because, and this returns me beautifully to our starting point where I said I came to this because of our difficulty to quantise gravity. I think that eventually solving the measurement problem needs to explain why quantising gravity doesn't work. So I mean, it works in some sense, we can do it mathematically, but this theory breaks down if you get to sufficiently high energies, which is called Planck scale. And so I think you can use the collapse to solve this problem by saying before you get to that scale, all the quantum effects will go away because they'll just collapse. So to me, they're like two parts of the same problem, really. But if you keep increasing scale and you never see collapse, you have to start wondering if collapse is even real. And if it's not real, then stay tuned for many worlds at the next festival. But I'm delighted that you were somehow super determined to return to quantum gravity against your best efforts. I think I've made my case as good as I could. Please put your hands together for Sabine Hossfelder and Maso Dao.