This episode originally aired on November 4th, 2025.
Ali Nazar: [00:00:00] You are listening to KALX Berkeley 90.7 fm, and this is Berkeley Brainwaves, a 30 minute show that brings you interviews from across the Cal universe. I am today’s host, Ali Nazar.
Today we are pleased to have on Professor Saul Perlmutter, Professor of Physics here at UC Berkeley. In addition to teaching at Cal, Professor Perlmutter leads the Supernova Cosmology Project at Lawrence Berkeley National Laboratory and has won numerous awards and prizes for his work, including the Nobel Prize for Physics in 2011 for his work in proving that the expansion of the universe is accelerating.
In the spring of 2026, Professor Perlmutter will be co-teaching an [00:01:00] interdisciplinary course called Sense and Sensibility and Science that is aimed at equipping students with basic tools to be better thinkers. He stopped by Brainwaves to talk about that, his career, and more. Please enjoy our interview.
Saul Perlmutter: I think I was just one of those kids who just wanted to understand how everything works.
And I was, you know, always interested in how, what all the languages were from different fields, how people were thinking about the world. Um, I was fascinated by the questions of, you know, uh, how, how does the, how is this world constructed? We, you know, here we live in a world with no owner’s manual, and, uh, you know, what keeps us from falling through the floor?
What, you know, how does this whole thing put together? Um, and for that matter, how does the minds work and how, and, uh, how, how our, uh, brains function? And so all these things were always fascinating to me as, as a child. And they then continued, you know, as I went to college and I considered a, I was, I was always interested in sort of the fundamental questions that underlie everything.
[00:02:00] And I was considering, uh, you know, majoring in physics and philosophy. In the end, I, uh, realized that if you did that, you wouldn’t have any time for anything else. And I was also interested in the other fields. And, and in fact, with the, uh, with, with that choice, I, I decided in the end I would do physics first, figuring I could come back to philosophy. Maybe a little more readily than if I did philosophy first and then tried to come back to physics. And in fact, um, by the time I, I went on to graduate school, um, I, uh, realized I had not dived in as deeply as I would like to, to physics. And so I decided I would try physics grad school. Um, and I was hunting as a grad student for some topic for your, for PhD thesis, uh, work for the research that would be deeply philosophical that felt like it would get at something, uh, that you would really want to know the answer to, um, even if you were not a physicist. And I, uh, hunted around and eventually, um, was lucky enough to, uh, come across the possibility of using astrophysics techniques [00:03:00] to address these fundamental, uh, questions of, of the underpinnings of what are the basic forces and, and, you know, particles of the universe, um, that you, at that time you might otherwise have approached using particle accelerators and particle physics.
And, uh, it looked like there was a way to get at some similar questions by just studying these, uh, these objects out in, in distant space with telescopes. And that’s, uh, that’s how this, this particular, uh, work began.
Ali Nazar: I am curious how you set goals in a scientific journey like this. I’m assuming that you don’t set out to try to win a Nobel Prize. Is it just curiosity leading you to question after question that kind of builds over time?
Saul Perlmutter: So I’d say yes. Um, definitely I was very interested in many of these different questions. Um, but I will say that if I had a, a goal, it was probably to get at something that felt fundamental, get at something that was, seemed like as an underpinning of other things that we now [00:04:00] know. And, uh, and so that was, you know, the, the kind of questions I was most attracted to.
Ali Nazar: Well, the questions that you kept answering led to work that famously prove the expansion of the universe is accelerating. Can you explain for us non-physicists exactly what that means?
Saul Perlmutter: Sure. Well, when I was beginning this work, uh, what we knew, um, was that there was this odd behavior of the, of the universe around us that we’ve known since, uh, the times of time of, you know, 1930s of, uh, Hubble and CIPHER, that when you look out into the universe and you see distant galaxies, um, you quickly catch on to the idea that all the galaxies are moving away from each other and moving away from us. And that only makes sense if the universe itself is actually expanding. It’s a mind-boggling concept because you have to [00:05:00] imagine a potentially infinite universe.
And what does it mean for a potentially Im infinite universe to expand? Well, apparently, um, every distance between every point and every object in the universe is getting bigger all the time. So it’s not expanding into anything ’cause it’s infinite, but it rather, it seems to be expanding from the inside out that, uh, there we’re adding space between any two points, any two objects in the universe.
And what was a key question at the time when I started, uh, doing my work as a grad student was the question of how much is that expansion slowing down over time just due to the, the gravity of, of every point, every object in the universe attracting every other object. And so you’d expect that that expansion would be slowing. And the more dense the universe was, the more it would slow down. And you could imagine it to be so dense that someday it would come to a halt and possibly collapse again. And then that would be the end of the universe if, if, if that happened. And [00:06:00] so, uh, I thought it was a, that seemed to be a pretty philosophically fundamental question. Do we live in a universe that lasts forever or someday we’ll come to an end and you could actually measure it if you could measure how much the universe was slowing down. And that’s what we set out to do.
Ali Nazar: And how did you actually measure it?
Saul Perlmutter: It turns out that we were really lucky that around the time that I was, uh, starting as a graduate student. Astronomers recognized that there was a kind of exploding star. It’s called a well exploding stars are supernova, and the this kind is called a type one A supernova. You can recognize it by what particular elements appear in that spectrum, and that particular kind of exploding star, um, appears to always um, explode very similarly, brightens to the same brightness and then it fades away. And that’s real useful because you can use that brightness then as a standard candle to tell you how far away that explosion is. And if you know how far away the explosion is, you know how long it back in time the explosion must have occurred because we know the speed of [00:07:00] light that is traveling to us from that distant explosion. And by the time it reaches us, um, you read off how faint it is and that tells you how far back in time uh, how far away and then how far back in time that event must have occurred. And all we had to do then was find a series of these exploding stars, that rep, that were at different apparent brightnesses and they would represent different distances and hence various different times back in history.
And we would then just ask for each of those explosions at from its time back in history, how much has the universe expanded? And that would tell us whether the universe is slowing down in that expansion. The last bit of the, of this story is that you can use the color of the supernova, uh, because when the, when the supernova explodes, it looks mostly blue, which is a short wavelength of light. And then while the light travels to us in an expanding universe, everything that’s not nailed down expands just with the universe. And that includes that wavelength of light. By the time it reaches us, it looks redder, and how much redder it tells you exactly how much [00:08:00] the universe has expanded since that explosion.
And so, you know, these are two big concepts. You know, one is you use the brightness to tell you how far away it is and how far back in time you’re looking. And you use the color to tell you how much the universe has expanded since that time. And, but it’s in the end, a relatively simple measurement. If you can find supernova from many distance, many different distances and different, different brightnesses.
And that was the challenge that we were faced with when we started the project.
Ali Nazar: And of course, you and your team were able to solve that challenge. And we’re speaking with, uh, Professor Saul Perlmutter of the Physics Department here at Cal, right here on Berkeley Brainwaves on 90.7 fm, KALX Berkeley. Um, of course that led to you and your team winning a Nobel Prize for Physics in 2011.
I’m curious, how did winning such a famous award change your life?
Saul Perlmutter: Well, the interesting thing about the Nobel Prize is that within [00:09:00] a scientist’s own field, a Nobel Prize probably doesn’t nowadays make a very big difference in what you’re doing because the people in your field who like what you’re working on still, like what you’re look on and people who um, have, you know, other things they would prefer to be doing, um, still are, are, are, feel that way. And so in your own field it doesn’t really change as that much. But what it does change is it provides a opportunity for you to talk across fields much more because people are, are interested in starting the discussion with you.
And so it, it has allowed for the possibility of all sorts of interdisciplinary work. Um, in my case, it also allowed the possibility of, um, starting to put together a, an approach to, to think about how do we teach scientific thinking, um, to everybody in the world, um, that it’s something that’s not just for scientists. This is something that can be useful for everybody in their day-to-day life. And I was able to pull together a, a group of, of [00:10:00] faculty to, to teach, um, this, this, uh, kind of material. Um, a philosopher and a social psychologist and myself all all started to develop this. And I think that was partly an outcome of the fact that the, there’s a sort of entree to talk to people if you, uh, when you win a prize like this.
Ali Nazar: And I know you’ve, uh, rode that conversation’s momentum into creating the Sense and Sensibility and Science course that’s gonna be taught again this spring. We’re gonna get to that in a little bit. Um, but first wanted to ask you a couple more questions. One is, I, uh, know that you taught a music and physics class in the past, and of course KALX is a lover of great music. So I wanted to ask you, can you tell us what the intersection between music and physics is?
Saul Perlmutter: Absolutely. Um, I, when I became, when I started as a faculty member in the, in the physics department at Berkeley, um, they, the chair, uh, approached me and asked whether I’d be willing to pick up a [00:11:00] course that they’ve had for, for many years, um, on, on the catalog, uh, which is called Physics Music, um, and the problem that they had had was that in the previous generation of physicists, there were a lot of musicians. Um, I think there was even the faculty string quartet, uh, at one point, and they would be all teaching that course over the years. Um, but they had all retired a number of years back, and it had been a long time since the faculty member, um, had a, a strong physics, uh, a strong, uh, music interest, um, who, who would also be willing to teach the course.
And they knew that I happened to be a, uh, amateur musician. I play violin and I sing, uh, um, and I, and so I at first thought, well, I’m not sure I want to do it, because I never took a course like that. I always thought it was potentially the most boring parts of physics meet the most boring parts of music.
And so I wasn’t, it didn’t have an appeal to me at that I originally, but then the more I thought about it, the more I realized that actually there’s a lot that you [00:12:00] can teach, um, uh, about how science works and how physics works. Um. Using music as your way in. And I, um, started getting together with a, a number of students, um, and we started to develop a concept for a course that we thought would be really interesting around physics and music.
And that would teach with every element of, of how the music works in, in, and, and understanding of it with physics. Also a little bit about how science works and how we learn about the world. Um, and, and what each of those stories tells us. So we began, uh, to develop that course. We developed ways of teaching it with all sorts of activities and games and.
Uh, exercises and it ended up being a real pleasure to teach. I think the students ended up really enjoying it, partly because for many people, physics sounds very intimidating to come to a boat. They don’t find music intimidating. Uh, you know, many, uh, you know, a big fraction of the, of the student population feels very much at home with, with music, and many of them are musicians themselves.
Um, but they [00:13:00] all, uh, listen to music. And so they brought a lot of themselves to the course in a way they would’ve, I, I think. It would’ve been a little harder, um, for a, for, you know, another physics course. By the end of the course, I was able to do one thing that I always thought would be fun to try, which is at the very last class, I taught a whole class about what is our current understanding of cosmology in the our Big Bang Theory.
And I was able to use, um, all of the physics concepts that they had learned. Over the course of the semester using music as the way in and all those show up again in an under understanding of cosmology so that I could teach a much more sophisticated version of how our picture of cosmology works than I would be able to teach a non physics class otherwise.
And so this was, I thought, a great example of, you know, how science comes together and, and allows you to do things that you never would’ve thought you could have done. Um, just because. When you learn something in one area, it has, [00:14:00] it has such an impact in other areas as well. Um, so in the end, I, I really enjoyed, uh, teaching it, and I was, uh, I was, I’m sorry, uh, at the end that I, that I, when I had to drop it, when I was starting to teach this other course, it’s so interesting to me how applicable physics is to so many parts of our lives.
Ali Nazar: And, you know, with that thought in mind, I know you used to lead the Berkeley Institute for Data Science and now it’s kind of a golden age for data science, as we’ve seen all these advances in artificial intelligence and large language models. So I wonder, you know, someone who’s thought about such foundational, critical things in our lives, how you think about AI and particularly this idea that we’re gonna create some artificial general intelligence that will be stronger or more smarter than us. Um, do you think that’s, it’s likely that humanity will create something that is smarter than we are?
Saul Perlmutter: I think in the long run, I think [00:15:00] it, um, it does seem that it should be possible, and it should be likely. Um, I can’t tell at all whether the current version of where we are is, is anywhere close, uh, as close as people are, you know, are fearing.
Um, my sense is that it’s getting at a big chunk of, of intelligence, but it seems like it’s not yet integrating in full picture of what, of, of what our intelligence consists of. Um, yet, and I, I think it’s, it’s possible to get there, but it’s very hard to tell whether or not this particular series of steps are, are right there, or, or still a bit of ways.
Um, I will say though, that, you know, there’s a, a real pleasure of watching, um, what you can learn from this particular round of AI, this, this, you know, large language model, [00:16:00] generative AI version. If we weren’t, you know, worried by it’s what it could be doing, its implications, um, I think it would be fascinating just to see, um, what combinations of things come together, uh, in this particular version of, of what, of AI, um, and what things you never would’ve thought you could separate out in intelligence.
So you’ll find yourself, you know, working with the, one of the current AI systems and realizing that it’s able to do certain things that you never thought it could do without also being able to do some other things that I can’t do. Yeah. Yeah. I think that’s just fascinating that you know, that you, that you can see that in, in action.
Ali Nazar: What are the use cases for this generation of large language models in your field in physics?
Saul Perlmutter: So the, I mean, so far, uh, the places that it’s seemed to me to have the biggest impact, um, have to do with making it easier to just [00:17:00] manipulate data and manage data. Uh, it’s, you know, a very useful tool when doing computer programming and, and, and analyses.
Um, and so just, uh, it’s much easier to do to, to, to pull things together using its capabilities than it is to go look up all the manuals that you might otherwise need to look up, uh, for, you know, just slinging the data around. And also just being able to pull together the programs to analyze it. Um. It has another really fascinating current use I find, which is that in any field where you need to talk across disciplines.
So for example, I’ve been working on, uh, you know, designing and building a space telescope, uh, and a concept. And there, um, you need people who are experts in, you know, electronics and people who are experts in, uh, oh, you know, fuel propulsion, and people who are experts in. Um, uh, optical engineering and others who are expert in, you know, quality assurance and, and management techniques [00:18:00] and all of these different, and, and then of course, the scientists who are, who are trying to get the, uh, in our case, you know, the, uh, the telescopes and the instruments to, to be able to measure things we wanna measure.
And each of the different groups has their own jargon and many of their own acronyms and terminology. And when you’re in those meetings, I found it, uh, really useful, um, to be able to take advantage of large language models capability of translating between the jargons. So you are, you know, somebody uses an acronym, you just type it in and say in the context of a space telescope mission, what’s a JJ XW, you know, or whatever it would be.
And it, um. It, that’s a place where the, that translation capability is really helpful and it keeps you in the conversation without constantly having to slow things down by asking, okay, let me pause everybody and I know at least half the group knows what you’re talking about, but what the other half doesn’t. And that’s, that can be very helpful.
Now, of course, the underlying, um, machine learning techniques that are underneath the, the current large language models, um, those, uh, are, have already been used and, and are [00:19:00] continue to be used by physicists in, in many ways. So, you know, cosmology models are, are, uh, you know, using the, some of the exact same techniques to train a, uh, a, you know, essentially an artificial system to predict what you will get in the next step of an evolutionary cosmology, uh, simulation from what the previous step was, much faster than you would get if you had to run the full simulation. Um, so there are many other applications of the underlying techniques. Um, but those of course, are not the large language model, uh, ones directly
Ali Nazar: That’s right. The, uh, large language models being kind of the new kids on the block in the world of AI, but them being somewhat controversial because we don’t really know how they work
Saul Perlmutter: Exactly.
And of course, you know, the, the, the fact that anything that looks, that uses sort of neural network approach, um, where you’re, you know, adding together many, many weights, um, on, in many, many layers of, of, uh, of nodes. Um, and, and, [00:20:00] and that’s what these large language models are, you know, also based on, um, those are all very difficult to interpret, um, how it is that they’re getting their results. And then of course, once you get to a large enough system it starts looking magical. And I think that’s one of the things that was such a, I think people really enjoyed actually, um, before they got scared, um, watching what, what happens with the, the large language models as they pulled together a big enough dataset that they can do things interestingly, that you, that are interesting, that are hard to, um, know how they got to them.
Ali Nazar: Yeah. Isn’t it kind of ironic that our science is creating things that we don’t understand?
Saul Perlmutter: No. Absolutely. Now, of course, um, we, we already know that, that that’s, you know, that that sort of thing is the case. Our own brains are a great example of that, something that’s very hard to understand. Um, and I think it’ll be really interesting to see as we use these techniques to, um, help us with science, um, do we end up with some situations where we can predict things, but we still don’t understand them?
And, [00:21:00] uh, and you know that that can easily happen. Um, but you, uh, you don’t want science to get too far off in that direction because then it doesn’t feel like science.
Ali Nazar: Well, uh, you’re listening to Berkeley Brainwaves here on KALX Berkeley, 90.7 fm. I’m your host Ali Nazar, and today Professor Saul Perlmutter is joining me. He’s a professor of physics here at UC Berkeley. And also, uh, the leader of the Supernova Cosmology Project at the Lawrence Berkeley Lab. And, uh, Professor Perlmutter, earlier in the conversation we talked about, um, your, uh, passion for interdisciplinary approaches to learning and applying scientific thinking beyond science.
And I know that you’ve got a course that you’re teaching with a couple other professors this spring that, uh, pursues that line of thought. The Sense and Sensibility and Science class, can you tell us a little bit about how that came to be?
Saul Perlmutter: Absolutely. Um, maybe one way to think about it is that [00:22:00] I, it it’s probably really important for everybody to recognize that we live in a very special time in all of history and prehistory.
Um, that we live it in like the first generations of humans who have demonstrated that we have the capability of, of dealing with global problems, um, and, and, and solving them, um, and making, you know, and, you know, not everyone gets perfectly solved, but we are able to feed billions of people when we, you know, when I was a when I was a child, um, I think two thirds of the, of the world was going to bed hungry every, every night. And they were, uh, and that was when there was only 2 billion people on earth. And now we have 8 billion people. And I believe, you know, that same poverty index is now below eight or 9%. Um, and so we’re clearly able to feed a whole planet now that we never thought we could.
We’re, we can deal with pandemics. Just for the very first [00:23:00] time ever, we’ve understood how it is that you can, you can quickly respond to a pandemic and stop people from, you know, dying all over the world. Um, you know, and if we can, if we can organize ourselves and make these things happen, we even know how to deal with, um, climate changes.
I mean, obviously. Some people who are concerned, uh, that, that, uh, you know, that it’s human cause and others who are think that, you know, climate change is whether or not humans, uh, create the problem. But in either case, climates do change in history, but we are the very first generations to, I think, understand some of the levers of how you could maintain a livable climate on earth. Um, and you know, I, I, I think the idea that we are just at this moment in history where we could really turn to each other and say, okay, we have all these tools. What would it take to make a, a world where everybody was fed, clothed, housed, educated, given [00:24:00] opportunities to, to thrive? Um, that would be exciting.
But, we’re also very aware that, um, none of those things happen, um, if people don’t, aren’t able to think through problems together, aren’t able to solve problems together and to talk to each other. And I think what the course, one of the motivations of the course is this idea that we have this amazing opportunity and what are the tools that we need to take advantage of it.
And some of that is something, are the tools that science allows us to bring to asking how do we solve problems together? So science isn’t just a collection of, you know, finished products of, uh, of we understand physics, we understand biology, we understand chemistry. It’s also the techniques to think about problems and in particular the techniques to think about problems with people that we disagree with, that science has really been able to have [00:25:00] all these successes because people have learned as scientists to work with other people who they will disagree with and where they need each other to think through rigorously where the mistakes are in our thinking and we’ll always making mistakes in our thinking.
And so we always need people who disagree with us to help us figure out where we’re going wrong and then to do better. Um, if I think we can get to that point, then I am actually relatively optimistic that we’ll be able to build the kind of planet we’ll be proud to live on. Um, the course itself then is walking through, oh, some, was it like 24 concepts I think it is, uh, where they provide a vocabulary of ideas, which helps you think a little bit more clearly about the world and also helps you think about the world with other people and how, and how you can figure things out together. So some of the ideas are just the ideas of, um, probabilistic thinking or how is it that we don’t think of the world as everything is either true or false, [00:26:00] black or white, that we have a probabilistic sense that, well, this one we’re 99.999% sure about. In fact, I bet my life on it, but this one I’m only 85% sure of. But that’s probably good enough for me to make a decision for certain purposes, not good enough for other purposes.
Um, other ideas of the course include the ways in which scientists have learned to stick with problems much longer than most humans would. Uh, that they, they’ve developed a culture that, um, that I, I think, gives them a sense of optimism that they can actually solve a problem. And that’s crucial because you need to be optimistic to, uh, so that you can stay with the problem long enough to have a chance of solving it. Most really interesting problems that are worth our salt do take longer than, than, than most people think. And so that’s one of the tricks of the trade.
And then the course finally gets around to the questions of what happens when people try to work together and [00:27:00] where do they tend to fall into a herd thinking, which is, I think, dangerous versus where do they, um, get some of the wisdom of crowds where you’re able to take advantage of the best of what everybody knows. In a positive way so that you learn from each other. And what techniques have we invented, um, or different groups, um, uh, try, uh, experimented with over time that look good at this kind of, uh, capability, um, of getting people to think through problems together. So that it’s a rather interesting combination of concepts that we’re teaching in the course, and, uh, and that’s why it really brought together faculty from humanities, social science, and, and natural sciences. And all three of us are always in the room, uh, together when we’re teaching the, the part of the course, uh, with, you know, with the full class.
The funny thing is, we, we started realizing that we needed it. Um, oh, now, it was like a dozen years ago, uh, um, as we were looking at society, making very practical decisions about, you know, what’s the appropriate level for the debt, uh, in, in a country. Um, and it be, [00:28:00] it sounded like the arguments about the, the debt limit were, you know, more just religious arguments, not actually, uh, people trying to think through, well, practically speaking, what do we know? What would work, how would we test it? Um, which is just a very different style of, of discussion. And so we began, you know, trying to think of it way back then. But, um, over the last dozen years, of course, we’ve watched our society struggle more and more and more with being able to talk with each other and especially be able to talk, um, across what appears to be in surmountable differences where in fact my read of what the actual, you know, not the leaders, but the people, uh, in, in our, in our population are much better at thinking through problems together if they’re given these tools and opportunity, um, than the current leadership is giving room to do. And so I think that it’s a, for me, it’s a real passion project. You know, it’s saying that I think we’ve got to do to pull ourselves out of this, this, uh, this, you [00:29:00] know, conflictual mode, um, which doesn’t really serve anybody.
Ali Nazar: Well, it sounds like an amazing course, and it’s available to all UC Berkeley students, this spring Sense and Sensibility and Science taught by Professor Perlmutter, um, and colleagues.
And I want to thank, uh, Professor Perlmutter for coming on Berkeley Brainwaves today. And, uh, we appreciate his time and hope you guys will check out his class. And this has been Berkeley Brainwaves here on KALX Berkeley, 90.7 fm. University of California and listener supported freeform community radio.
I’m your host, Ali Nazar. Talk to you next [00:30:00] time.
