Eric Betzig once walked away from science altogether. Years later, he and his best friend put $25,000 each of their own money into building a microscope on a living room floor, and it won him the Nobel Prize. In this episode, Méline sits down with Eric in his office. Eric is a physicist, a winner of the 2014 Nobel Prize in Chemistry for super-resolved fluorescence microscopy, and a Professor at UC Berkeley. He talks about his years at Bell Labs, the frustration that led him to quit, and the glowing jellyfish protein that pulled him back. He tells the story of building PALM with Harald Hess, a microscope that can see beyond the limit of light. He also explains why he believes biology is stuck. Life, he argues, can only be understood by watching it live, and we now have more imaging data than any human can interpret. The next breakthrough will need AI that can think in five dimensions. This is a conversation about curiosity, hard work and failure, and the biggest mystery in nature. He explains: ◼ Why we understand neutron stars better than what is happening inside our own cells ◼ How a microscope built on a living room floor led to a Nobel Prize ◼ Why AI drug discovery based on protein sequences alone will get nowhere ◼ What 5D imaging means and why we need AI that can think in five dimensions ◼ What Bell Labs got right about small teams, hard work and limited resources ◼ Why he believes prizes are corrosive to science 0:00 Intro 1:25 How Apollo and Star Trek Led Him to Science 2:00 The Limit Every Microscope Has Faced for Centuries 4:05 Why He Quit Science 4:59 Inside Bell Labs, the Legendary Lab of the Last Century 7:29 Why Small Teams Beat Big Ones 10:12 Price’s Law and Who Really Does the Work 12:04 Why Prizes Are Corrosive to Science 13:22 Harald Hess and the 4am Race to the Car Park 16:00 What Harald Taught Him About Humility 19:37 Out of Work and Out of Science 22:42 The Glowing Jellyfish That Changed Everything 27:50 The Scientist Who Funded His Work by Selling Ties 29:41 The Idea Behind PALM 32:26 Building a Microscope on a Living Room Floor 33:06 2005, the Most Miraculous Year of His Life 35:44 What 3D, 4D and 5D Imaging Really Mean 36:44 Why Biology Is on the Wrong Foot 42:41 Why We Need AI That Can Think in 5D 45:27 Why AI Investors Are Not Listening 47:31 Why Physics Has Stalled and Biology Has Not 49:06 Advice for Young Scientists 49:37 Why Science Has Grown Too Big 53:29 How AI Will Accelerate Everything 54:51 The Biggest Mystery in Nature 55:27 The One Habit That Matters Most ◼ Eric Betzig, UC Berkeley: https://physics.berkeley.edu/people/faculty/eric-betzig
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
SPEAKER_00All the AI companies that are trying to do drug discovery now, they will get nowhere if that's the only information they use. How can you understand life if you don't get it live? You can't. There's only one tool that allows us to look at life live, and that's the optical microscope. Life is by far the most complex matter in the known universe. We understand the interior of neutron stars better than we understand what's happening inside a cell in our bodies. It is the biggest mystery in nature. That caused me to actually quit science.
SPEAKER_00In those 12 years that I was out of science, a number of things happened. That's when I jumped back in, and that's what led to my Nobel Prize. We need somebody who can natively think in 5D. So we need to invent the next tool to get us to the next step. It's exciting and it's scary at the same time. Imaging has been stuck for biology. As much as I've worked on it, I've been a complete failure because we haven't moved the needle in any meaningful way. But if we can get over this last hump, it will be more impactful, would revolutionize imaging.
SPEAKER_01Hi Eric, welcome to the Humanware Project. And it's such an honor to sit in here with you in your office.
SPEAKER_00Thank you. Happy to be here.
SPEAKER_01Could you tell us more about your work and how does this all start?
SPEAKER_00I started uh um wanting from a very early age to do science and engineering. I was very much motivated by two things. Uh the Apollo program. I actually wanted to be an astronaut first, and then um also Star Trek. I didn't get really into microscopy as a field until graduate school. Um, when I went to Cornell, and there were a couple professors who had a crazy idea to make a microscope that could see almost down to the molecular level.
SPEAKER_00Um and microscopes have always been known for hundreds of years that there's a limit to how small you can see, which is about the wavelength of light. And to put in perspective, that's a hundred times smaller than a cell, so you can still see things in a cell, but it's about a hundred times larger than the molecules that make up the cell. So you don't really have full information down to level you want. So the, you know, the idea of being able to make a microscope that can look at living cells non-invasively with the resolution of electron microscope was really guided the first 20 years of my professional career.
SPEAKER_00The approach we took to see smaller than the wavelength of light is we would take a black material and we would use an electron microscope to drill a hole in it, which is much smaller than wavelength of light. Then we shine light on one side of this black screen, and the light that comes out of that hole, because of the hole is so small, has to be confined to a size as small as the hole instead of the wavelength of light. And by doing that, we would get very high-resolution images.
SPEAKER_00The problem with it is that twofold, and so I did this in graduate school for six years, then I went to Bell Labs and continued that work for six more years. And eventually I got very frustrated with it because while when we had perfectly flat samples to look at, um we could get good resolution. Cells aren't flat, okay, and cells move. Okay. The problem with doing it with this little light source is you had to scan it sequentially around the specimen to get your image point by point.
SPEAKER_00That takes a lot of time. And so the cell is in the meantime moving all over the place, and so you just can't really follow it along. And the roughness of the cell means that the light out of the hole spreads so rapidly that um that you don't have good resolution over most of the cell because the hole is too far away from the specimen. So that that caused me to actually quit science. Um, uh, that and many other aspects of what I learned about academia that I didn't like about academia. Um, and that's when I transitioned to working for my dad's machine tool company.
SPEAKER_00But after I left that, um, it turned out in those 12 years that I was sort of out of science, a number of additional things happened within science that made possible a different approach to beating this so-called diffraction limit and getting high resolution. And so that's when I jumped back in, and that's when my buddy and I did this on his living room floor, and that's what led to my Nobel Prize. Yes.
SPEAKER_01Because when I'm reading about your story, it's I think it's more dramatic than many films.
SPEAKER_00It's it's it's it's like it's like a Hollywood movie.
SPEAKER_01Yes. Um, you know, Bell Labs is such a uh you know legendary lab back in uh the the last century. Do you feel like a genius when you were there?
SPEAKER_00Well, no, definitely not, because there were many, many people much smarter than I was inside of that building. When you first go there, you feel like you're you're like you're on probation and like, do I really belong among you know these people and and this history, right? You know, I when when you when you start looking for a job after graduate school, you kind of know what to expect when you go to different universities or companies about what kind of questions you're gonna be asked. Most of the time the people aren't expert enough or or um or daring enough to ask really probing questions.
SPEAKER_00But when I went to Bell to give my uh you give a talk, an interview talk, and they told me to prepare a 45-minute talk, it took me two hours to give that because I was constantly interrupted with questions.
SPEAKER_01Okay.
SPEAKER_00I loved it because it it meant A, these people were very sharp to get immediately what I was talking about. And they cared, right? In the modern world now, when you go to conferences or you give a talk, most of the time people just politely clap and great talk, you know, but they don't really, they they aren't combatitive. They aren't digging into what's at Bell, everything was just it wasn't personal, but everybody just wanted to get to the root of the science and the issues that were involved in Bell.
SPEAKER_00By far the best job I ever had was it was at Bell.
SPEAKER_01How is it like to work in there? Because I've been reading books about Bell Labs, and that was just such an honor that be able to talk to one of the biggest.
SPEAKER_00Yeah, it it was a once-in-a-lifetime experience. It had it had so many advantages. The first is um because it although when I joined, there was a decision that broke up the monopoly of the phone company that was the financial underpinning of Bell Labs, right? So they were I was that was early enough and close enough in time to when I joined that it hadn't really affected the funding of the organization, the organization as a whole. Um but um and so a key element of it was this very stable source of funding.
SPEAKER_00That being said, they did not allow you to grow uh large groups. So the beautiful thing about Bell was whether you had a Nobel Prize or whether you're a new hire, the most you could have is yourself and two other hires. Okay. Whereas on this campus, you'll find people have 50 people in their group, right? There is no way, no matter how smart you are as a professor, that you can be fully involved in the work of 50 people and know what's going on, right?
SPEAKER_00So it's a whole different organization where most of the people are in expert and beginners at what they're doing. And so it's incredibly inefficient. At Bell, everybody was an expert. And because the groups were so small, if you wanted to do something that was outside of your expertise, you had to collaborate. And because the groups were so small, they could have many, many groups, hundred plus groups, okay. And so every paper I did was not known by advance, but by just talking to people in the hallway.
SPEAKER_00Oh, we can use my microscope to look at uh light and optical fibers, we can look at semiconductor lasers, we can look at high-density data storage media, we can look at living cells, we can, and every paper I did in that period was not predicted in advance, and every paper was essentially revolutionary because of these sort of by chance interactions that you could not predict in advance about how people would come together. That was really its strength. And everybody worked like crazy.
SPEAKER_00I mean, I came in at four in the morning and I left, I left at 10 at night, and I did that seven days a week, and it was it was the best time ever.
SPEAKER_01What do you would say is the biggest lesson you learned there, uh, you know, like which influenced your work later?
SPEAKER_00Well, uh, many things. Um, one which I already knew is I'm a firm believer in hard work, really hard work. Um, I there's no substitute for it, because everything is so complicated in science and so difficult. The only way is to really fill your mind completely with everything you can so that even in your sleep or whatever, your subconscious mind can work on the problem, right? So so that was one. The other is is uh it's it's corrosive to have too much resources, okay?
SPEAKER_00Because because it it it means you're not you you can find creative solutions to problems without just throwing money at a problem or throwing more people at a problem and so forth. And um, and again, this this ethos of a small group, because um because then you can be fully involved with one another in the work and have a deep understanding of what everybody is doing, which you can't do under the normal university type system, right? Um there's something called Price's law, okay, which it which is valid across all disciplines, all fields, which says that the amount of work done goes with the square root of the number of people working on the problem.
SPEAKER_00So if you have a group of a hundred, roughly 10 of the people do half the work, all right. And and if you have four people, or three people in my case, square root of three is almost two. Two out of three are doing half the work, right? So, so um, so there's many, many lessons to be learned from Bell. And the sad thing is that there have been many efforts to recreate Bell. I was at one later on called Genelia Farm Research Campus, the Howard Hughes Medical Institute. And it was consciously designed after Bell Labs.
SPEAKER_00And for the first few years, it's it was sort of Bell-like. But because, again, most of the people were coming from academia, and they start screaming that we can only have a group of six, how can we do anything with six? You know, I used to have two. Um, but in the end, what happened is the group started to grow. It started to be less and less bell-like over time, and it became much more like academia. So many people have tried to take the lessons from Bell Labs. Nobody has successfully really, really gotten into creating a structure exactly like Bell Labs again yet.
SPEAKER_00The 12 Nobel Prizes. Mine is one of 12.
SPEAKER_01Yeah.
SPEAKER_00Okay.
SPEAKER_01And you are able to collaborate with all of them. Oh, yeah.
SPEAKER_00Yeah, Lou Bruce, who won it after me. I I had coffee with him every morning. Horse Stormer was who got the Nobel in '98, is the guy who hired me. I was in his department at Bell, right? Arno Penzius, the guy who found the black the Big Bang, he was the vice president, right? I mean, yeah, you're you're rubbing up shoulders with rubbing shoulders with some pretty smart people. Yeah.
SPEAKER_01And when you started, how have you ever thinking about you will win a Nobel Prize one day?
SPEAKER_00I mean, when I was a kid, I would, I would tell my sister, like, I'm gonna win a Nobel Prize and be a billionaire by the time I'm 40, right? So I I'm certainly wrong on both counts, one more than the other. But um, but again, that's just like an eight-year-old talking, right? So who was into science? Um, honestly, uh, that's another thing I would say is is um I think prizes are corrosive to science. I think, I think they're really a net negative. Um, because uh so many people, um, for every prize, it's you know, it's it's very subjective, right?
SPEAKER_00How are you gonna decide this person versus that person or this type of work or that type of work deserves? Usually what happens is there's just a lot of horse trading and backroom dealing about who's who's, you know, Harvard will put out, you know, 50 people to put in recommendations for their guy to get a prize or whatever, right? Or, you know, there's just so much silliness about it. And for everybody who wins, there's 50 people in the field who feel slighted, right? And it's it's just, you know, the idea of working towards prizes as a way to motivate yourself and say you're not you're not a scientist if you care about the prize, right?
SPEAKER_00You should care about trying to understand nature better.
SPEAKER_01Also, like of through the uh Bell left, you meeting your later best friend and buddies that you're can tell us the story about that. I heard you guys have very competitive character to working with.
SPEAKER_00We do, we do. So, so um, so uh actually when I interviewed at Bell, Harold had already been there for a couple of years, and of course, you go around and a number of people, you go to their labs, you talk to them, and then they might uh and then um they'll ask you questions and then take you out to lunch or whatever. So, so um I met Harold briefly and he immediately we resonated, even though he's a few years of just four years older than me, he was doing also microscopy, a different kind than I was doing.
SPEAKER_00But but so there was a lot of overlap there. He was from the Midwest, I was from the Midwest. I he wasn't on my schedule, but at the end of the day I said, Can I talk to that Howard guy again? Because I really liked it. And oh, you mean Harold? Yeah, sure. So so we talked. And and immediately once I got there, we hit it off because we both play tennis, we're both at about a comparable level. Um, he likes to work just as hard as I do. We're both just extreme nerds, absolute extreme nerds, okay? We're the type who, you know, walk down the corridor at super high speed and don't look at anything, they're just babbling away, right?
SPEAKER_00And so uh um uh so we would compete on everything. So again, I said we get in at four o'clock. If I got in, um, we would try to be there the the spot in the parking lot closest to our wing of the building. We tried to be the guy first there. So we would each get there about four, and if he beat me, I would say shit. And I'd put my hand on the hood of his car to tell by temperature by how many minutes he beat me by, right? And then and then we would work and we'd work until uh six, and then every night we would go.
SPEAKER_00Well, actually, at once the sun came up, we would then go to the high school to play tennis on the tennis court there. Then we would go back to work and work till six. Then we ate at the same Chinese restaurant two blocks away every night, and then we go back into the lab and we do that seven days a week until 10 p.m. and repeat. Okay. So, but yeah, we both got Harold. Harold is the smartest and most um gifted experimental physicist I've ever known.
SPEAKER_00He did things with his microscopes that I wouldn't even dream of attempting. Um, and so he became a lifelong friend. And, you know, once I left Bell, you know, we stayed in touch, but uh it was it hurt to not be able to see Harold on a daily basis.
SPEAKER_01So uh if you would say there's one thing you learn from Harold, from his character or his science side of um work, what do you would say that one thing you learned from him?
SPEAKER_00Humility. He is a very humble person. Okay. I mean, I have much more of an ego than Harold. Harold is much more of a pure scientist than I am. I came into science partially motivated by doing science, but partially motivated by ego, getting my career going, all of that. Over time, by being with Harold, I learned the humility that he learned. I learned that those artificialities of ego and and so forth are are, you know, silly, right? That that really it's all about trying to trying to do the best to understand nature, trying to use your gifts as as an experimentalist and as a physicist to reveal that stuff.
SPEAKER_00And that's it's you know, when you when you get your experiment working and it's getting data, there there's just no rush better than that. Far better than sex or anything else. It's just like, you know, man, this is just amazing, right? And and those eureka moments are things that that, you know, feel you you can call recall them in your mind like it was just five minutes ago, right? It's just yeah.
SPEAKER_01And what would you think um Harold would say the biggest lesson or the quality he sees from you?
SPEAKER_00The problem that he has as a scientist is he's too nice of a guy. And so once he gets once his skills are known throughout the building, um, he has more trouble saying no than I do. I tell people to fk off. I'm busy doing this, right? Harold is always very nice and tries to help everybody, but it means you get too diffuse, right? And I keep saying, you gotta focus, man. It focus on what you care about, right? I mean, because because and and he's been wildly successful there as well.
SPEAKER_00Um, but still I I think he's learned a little bit to uh kind of not be as as gregarious and generous as he once was, but to learn how to limit uh to the really important things.
SPEAKER_01Do you think it is a privilege for someone who is able to pursue their true desire to otherwise a science or privilege is the word.
SPEAKER_00Yes. Okay. So that's a very small subset of people that you're looking at. Like I saw you talk to Bob Lefkowitz, right? I mean, again, it's it's it's this, he's a true scientist, you know. I mean, he's really boring into G-coupled protein receptors, right? Um, and and and those are the type of people I love to work with, but they're they're extremely rare in in science in general. Okay. Um the the uh yeah, the majority of people are are are more self-interested than that.
SPEAKER_00The final direction things happen is almost always unpredictable. That's true for many startups, right? Many of them have to pivot, pivot, pivot in order to get to to something that finally works. Or or you know, a scientist pivot, pivot, pivots too, right? Um, but um, but it so it isn't so much um focus on one thing at the start, it's on the ability to kind of learn from your failures as you're going along, as to, as to, you know, there's there's a literal infinite number of things you can do.
SPEAKER_00And so the more information you get by failure is a great teacher, and the more you learn by failure, the more you're able to winnow the possibilities of what you might want to do going forward, right?
SPEAKER_01Uh yeah, as you mentioned earlier, you said at the age of 40, you'll win a Nobel Prize and become a billionaire in the middle of the year.
SPEAKER_00So here I am, so here I am at the age of 40, and I'm unemployed, no Nobel Prize, no billion dollars, right?
SPEAKER_01The kids and the family to the five years.
SPEAKER_00So my wife was still working, so it wasn't like we were going to starve, but still I need to figure something out. And and and so I reconnected with Harold, and so as Bell, Bell eventually imploded under the fact that they didn't have um that underpinning of the monopoly, right? And then there was the dot-com boom, and they tried to monetize Bell Labs as part of that dot-com boom, and so they spun it off in this Lucent technologies, and then Lucent had a big run-up in the stock market, and then in '99 that all crashed.
SPEAKER_00And so there was this diaspora of scientists leaving Bell Labs around 2000, and that was kind of the end of really Bell Labs still exists, but it's not like the Bell Labs that existed for 50 years stretch there. Uh I just knew that I had to do something, and I started to read the scientific literature, which I hadn't done for six years that I was in my dad's company. And I realized that while I don't like academia and I don't like how science is usually practiced, I missed doing science.
SPEAKER_00I miss doing science the way I want. And I wanted to find some way to get back to doing science. And so I started reading the scientific literature. And um, and the one thing there were two things I I knew is is I didn't want to do anything that would be anything related to machine tools at all. I didn't want to do anything related to microscopy at all because I've been there and done that. Okay. Um, and so I I you when when you get into these positions of wanting to pivot, it's really hard because your brain is full of all of these ways of thinking about everything that's impacted by your work, right?
SPEAKER_00And so you have to flush that all out, and that takes a while, right? You have to immediately say, No, don't think about a you know, a microscope, right? No, don't think of you have to, you have to keep so you you have to take the blinders off. When you're focusing on your goal, you have to put blinders on and just go. But when you're trying to figure out what you want to do next, you have to have as broad a spectrum as you can. You know, do I want to be go to go to a cooking school and be a chef? Do I want to do this? What do I want to do?
SPEAKER_00Right. And in science. And um and uh I started reading the scientific literature and I I saw that there was um uh uh one of the again, the big one of the big problems with with my first near field microscope for super resolution was it was just gonna be way too slow and damaging to look at living cells. And yet um uh uh while I was working for my dad's company, um three guys came up with this idea of taking um a piece of DNA from a glowing green jellyfish, and then you could splice it onto any protein you want in a cell, and then that protein would glow.
SPEAKER_00Okay. So one of the biggest problems of being able to do microscopy on living cells is that, you know, living cells themselves have very little contrast. It's basically like water on water, right? It's little bags of water. But fluorescence, this ability to take little glowing molecules and specifically attach them to specific types of proteins, cause them to just light up and glow. That's been known for a long time, but the methods that existed back in when I was at Bell sucked, frankly. Okay, it's introducing chemical dyes that would make the cell sick and it would go to places where it's not supposed to go in addition to the things it's supposed to.
SPEAKER_00But the this idea of just using genetics to be able to put on this glowing tag onto just the proteins you want in a very non-invasive way. My jaw was dropped down for about a week afterwards because I couldn't believe how that elegant was. And I said to myself, sh, I gotta do microscopy again, right? Because this was just too good. And the irony is that paper that about that glowing jellyfish came out one year after I left Bell Labs. If I had known of that paper, there's no way I would have left when I did, right?
SPEAKER_00Um, because I knew this would revolutionize imaging, right? And those guys won the Nobel in 2008 for that work. Okay. So um, so I I started thinking about what to do, and and um I started thinking not about near field, but okay, there are many things you care about when you want to image living cells. Um you're sure you care about resolution. And unfortunately, even to this day, most academics think, what is the metric by which you judge a microscope?
SPEAKER_00It's resolution. That's one metric you judge it. The others are, does it kill the cell? Is it damaging in any way? Can I look in three dimensions? Can I look fast enough if the cell is moving that it's not blurred out? So I need speed, I need 3D, I need time. So I wanted to make a microscope that would allow us to look at living cells with as much a gain in terms of the gentleness, the duration, and the speed of imaging compared to the current state of the art at that time, as super-resolution was a gain in terms of spatial resolution.
SPEAKER_00So I would back off from trying to get super resolution, but instead let the whole rest of the world march in that direction as super-resolution. I'm going to take what's advantageous about these new things and march in a different direction and instead try to do live imaging better than anybody has done at the normal diffraction limit. Okay. And so I came up with this idea of interfering beams from multiple directions to create this massive array of little concentrated spots of light that I would in parallel be able to image the cell because each one only needed to look at a little bit of the sample.
SPEAKER_00And then you build up that from each little cell, you know, each little region, what's going on to get the whole picture of the whole vine. Um, and so I spent two years of developing that. And um, and then I had this this, I called it my optical lattice microscope. And I said, okay, I have the theory all worked out. I have some designs designed for how to build this microscope. What I don't have is any money or any support or whatever. So what I did is I called Harold and I said, Harold, I have this idea.
SPEAKER_00I'm really excited about it. I want to do this, but I don't know what to do and um how to get any funding for it. And he's and Harold looked at it and said, Yeah, it looks, he was now working for a disk drive company in San Diego after he left Bell. And and um I think he too was feeling, you know, you know, dissatisfied because he wasn't doing basic science anymore, right? And so um he he's I I asked him if he would come in and do this with me, maybe we start a company, do whatever to build this.
SPEAKER_00And he's and Harold is Harold is a proud man, and he said, I'll help you find something, but I don't want to do it with you because it's it would be chewing your cud. And and I don't want to do that. Um, but I will help. And so and so he together and I started going around um to places to try to sell this idea. And one of the places that Harold thought of was he was um friends with another guy at Bell Labs, um, Greg Bobinger, um, who when Bell imploded, he went to Florida State and ended up heading the National Magnetic High Field, High Magnetic Field Laboratory there.
SPEAKER_00And and he had been trying to recruit Harold to come there as a scientist. Um, and um, so Harold had made a previous trip and had learned, um, met another guy who was this weird dude who didn't fit in with magnetic fields at all. Instead, he was doing live cell imaging using these new types of proteins, right? With conventional microscopes. And Harold thought, well, it would be a natural fit for you to take your cool new way of doing live imaging and work with that guy to do it there.
SPEAKER_00So we went over to there, and when we were, when we were there, um, this guy, his name was Mike Davidson, and he was his day job was to look at the um the wires with microscopes to look at the the the quality of the wires that go into making these super high magnetic field things. But in most of the time, what he did instead is he used those microscopes to look at living cells because that's what he cared about. And in order to pay for his work, what he did is he took pictures of different things like cocktail mixes under a microscope and on a polarized light, and it looks sort of psychedelic.
SPEAKER_00He printed those on ties and he sold those and he made millions of dollars selling those ties, and he used that to support his work to do live imaging. Okay. And so we met this Mike, and Mike told us about, yeah, he knew more about these fluorescent proteins, these, you know, that I talked about from the jellyfish and all their variants than anybody I knew, right? But he said, hey, there's this new one too, and it's kind of weird because normally when you look at the cell, you don't see anything at all. But if you shine purple light on the cell and then you shine blue light on the cell, then you see the green glow that you should see like the jellyfish.
SPEAKER_00But at first you don't. So the molecules are first in an inactive or dark state. You put on the purple light and it becomes active, but even then nothing happens. But then you put on blue and it'll glow green. Okay. And so the idea I had when I was pushing my daughter in the stroller when in that first round of unemployment, was that you could do super resolution not with a small hole, but the idea was that if somehow you could take all the glowing molecules in the sample, the problem is that in a normal cell, those molecules are so close together that when you look at it in a normal microscope, they make a fuzzy blob.
SPEAKER_00And that blob is so much bigger than the separation between the molecules that all you see is fuzz, right? But with Mike's this new molecule, if you turn the violet light down so low, then statistically only a few random ones are going to pop on, statistically likely to be separated by more than the diffraction limit. So the so the glowing balls are isolated, and you can find the center of a glowing ball to much better precision than its diameter. So once you have those separated glowing balls, you say, ah, that molecule is there, that molecule is there, that molecule is there.
SPEAKER_00Fine. Okay, I got three molecules. That's not enough because there's 10 million molecules there. But then I turn on a different set and I do it again, and I take a series of hundreds, thousands, tens of thousands, hundreds of thousands of pictures. We'll get to that part of it, but you are absolutely right. But if we're not worried about time, if we're worried about resolution again, it's like, holy, and then you stack it together. And it's like, oh my god, this is the idea that I had 10 years before, but now it's doable.
SPEAKER_00In fact, it's trivially doable. I mean, it's it's like, why hasn't anybody done this already? It's just so obvious. And so Harold said, because we had come to this epiphany together, now it's not chewing my cud. Okay. So we drop my idea of this optical lattice microscope like a hot potato and say, we got to build this thing immediately, like immediately before anyone else gets onto it. So, so um, so all of a sudden Harold is invested, he quits his job, and we start trying to build this.
SPEAKER_00And it's like, you know, it will take too much time to get VC funding, to get a grant or whatever. Okay, well, we can do it ourselves. It's that simple. We don't need anybody else. So, so Harold, when he left Bell Labs, was able to take a lot of his equipment with him. So we had all of that in one of these storage sheds, you know, that's had been collecting dust for a few years. And so we pulled that out. Then we had to put about 25,000 each of our own money, which while I was unemployed for two years was a bit of an ouch, but we did it.
SPEAKER_00And um, and that's that's the story. That's the story everybody knows, where we built it on his living room floor. It was we had the idea in I think April 2005, and by um, and we built the microscope over the summer in September 2005. We shipped it to NIH and started doing experiments. And within a week, we just had some just just the molecules, not in a cell but on a surface. And we turned on the violet light and we saw just a few molecules pop on, and we said, We have it.
SPEAKER_00This is gonna work. And and it was like, it's really gonna work. And so uh then we start doing the cells and we get all the data for the paper, and then the paper appears, and the rest, as they say, is history. 2005 was the most miraculous year of my life. Not only did that fall in my lap, but when I was uh as I was going from place to place to try to sell my optical lattice idea, one of the places I went was Columbia. And Columbia, there was a guy, Marty Chelfi.
SPEAKER_00He was one of the three guys who came up with the idea of using the jellyfish to create this, right? And so um, so he won the Nobel in 2008. And um, and uh I told, you know, I he I gave a talk on that microscope. Um, and um, usually when you go to give a talk in university, they take you out to dinner afterwards, right? And so Marty was in the cab with me going to the air, going to the restaurant, and he said, he didn't understand a word of what I said because there was a lot of physics on how to create this optical lattice microscope.
SPEAKER_00He said, But I said it would be this much faster than a regular microscope, this much less toxic to cells, it's this much better resolution and speed. And he understood those things. And he said, So how are you going to get this done? I said, I have no idea, but a week ago I saw there's a you know trade magazine called Physics Today that comes out once a month. And there was some story about some crazy guy who wanted to create a biological Bell Labs. And I said, I love Bell Labs, and a place like that would be a great place to do it.
SPEAKER_00And he said, Well, you know, maybe someday, but uh, you know, first you got to re-establish your chops or whatever. And so I thought he was trying to politely say, you're not ready for that. And um, and yet then a week later, I get a call, and it's Jerry Rubin, the guy in the story who is building this thing. He said, I saw Marty Calfi at a party, and he said you might be interested. Chinelia, then the new Bell Labs. And I said, it's might be. And so within a month, not only did our super resolution technique palm fall in our laps, our next jobs now with Howard Hughes Medical Institute and a new Bell Labs fell in our laps all in the same month in 2005.
SPEAKER_00It was just, and and a month before that, I remember just being just incredibly low. Incredible, I'm never going to be able to get a job in science again. Complaining to my wife that, you know, life is black, you know, blah, blah, blah, blah, blah, blah. Anyway, it is what it is. And I didn't do it to win the Nobel Prize. I did because, you know, having been in that field for so long, when when you see that there is a path to what you had dreamed about doing, you know, it it's it's a compulsion. It's a, you know, you're you're you just have to do it.
SPEAKER_01There is one thing I find is extremely interesting about when you're imaging a cell, so you you put a 3D, 4D, even 5D. What does that even mean?
SPEAKER_00Yeah, so so I think everybody understands 3D, right? I mean, we live in a 3D world, right? And cells are three-dimensional too. They operate in three dimensions. Most of the images, even today, to my everlasting shame, that biologists take of cells is two-dimensional. Um, that's true in an in an electron microscope, uh, because you can you can only get stuff in two dimensions. And like I said before, a normal microscope, there's only one plane that's in focus at a time, right?
SPEAKER_00So while you can, you know, move the plane up and down and build a volume, and if we go looking in the microscopes later on, you'll see this kind of stuff going on. You can then build up a three-dimensional picture. But the hallmark of life is that it's animate. Okay. Um, the thing that defines life is its dynamics. I could rant for six hours about why I think biology is really on the wrong foot. Um, the problem is that biology is is um the victim of its own successes of the 20th century.
SPEAKER_00Those successes were biochemistry, knowing how in a test tube to figure out which proteins uh interact with other proteins, then structural biology, so things like electron microscopy that says, well, a mitochondrium looks like this or whatever. And and then um, and then genetics, um, you know, knowing DNA and RNA and the central dogma and all of that. All of those methods are exceptionally reductious. They're about taking the cell apart down to its constituent pieces and trying to understand how they look or or where they are or how they interact one at a time.
SPEAKER_00Okay. Um the problem is that the cell is it's not a machine really, but it's a highly dynamic system. Let's take a much simpler dynamic system, a car engine. Okay. It has several hundred parts. Okay. If I took that car engine and broke it down into all its pieces and threw it on the table here, please, please remake a car engine for me. Tell me how a car engine works. Okay. That is many, many orders of magnitude simpler than trying to understand biology.
SPEAKER_00And yet 90% of the biology that is happening on this campus or around the world is exactly trying to do that. All the AI companies that are trying to do drug discovery now based on alpha fold protein sequences, they will get nowhere if that's the only information they use. Because they're looking at one part in a billion of what's important, and they know nothing about the dynamics, right? The dynamics define life, right? You have to look. How can you understand life if you don't look at it live?
SPEAKER_00You can't. You can't. There's only one tool in our toolkit that allows us to look at life live, and that's the optical microscope. And particularly like my lattice light sheet microscope. So now we got 3D, but the 4D is the time, okay? So now you have 4D, right? So all of these, I don't know, it's still wiggling around there. There, those are lattice light sheet microscope images that you're seeing right there, right? So you're seeing essentially 4D, right? Because you're seeing, you're seeing projections of what's really a 3D data set, right?
SPEAKER_00And as it moves and so forth. So that's what life is like, okay? That every cell in your body is doing that. Inside of every cell, there are 40 million molecules of 20,000 different protein types bouncing around stochastically, randomly, seemingly randomly. And it's how they stick together that causes them to start to assemble into macromolecular assemblies, organelles, cells, tissues, you and me. It goes, there's levels of emergence from single molecules up to the entire biosphere.
SPEAKER_00Okay. And so life is all about that process of going from the inanimate to the animate to cover as film around the entire planet Earth. Okay. And so 4D is that. The 50 is again, I said, 20,000 different types of proteins in every cell in your body. So if we want, we can't see all 20,000 at once. I would love to, but with the fluorescent proteins again and the jellyfish and all of that, we can look at five or six at a time, right? And see how they actually interact.
SPEAKER_00That then gets to the icon story, which we can get into later if you want to get to. But you have to understand how to look at the dynamics of life to understand life.
SPEAKER_01Our mind or our physiology is experiencing a 3D word. If you add a timeline to look at the cells, but what exactly is 5D?
SPEAKER_00The fifth dimension is just the number of different types of molecules. So I can create an image that looks at only one type of protein. I can look create an image that looks at a hundred different types of proteins. So now that if if I think in terms of you know building up, okay, 2D, I've got all these little pixels and 2D, 3D, they're little cubes, 4D, they're little hypercubes, right? And with time is the fourth dimension. And now the fifth dimension is I want to know not just all of that for one protein, I want to know that for 10 proteins.
SPEAKER_00So now I've got a fifth dimension, which is color or protein type or whatever you want to call it. So that's where the 5D comes in. But it means you have to take crazy, crazy, crazy amounts of data. We take, we take, you know, uh, I mean, in a week we will take a petabyte of data, which is a million gigabytes of data, a thousand terabytes of data. Okay. And um, and so uh uh it's beautiful and it's complex.
SPEAKER_00And if you try to render it like those movies, you are compressing what is an incredibly complex data set into something which is interpretable to your eye, but really not explanatory about what's going on. Because all we can see, we evolved to see those are as much as I call that a 4D movie, what that really is, is that's a 2D movie. It's 2D plus time, 3D, right? Because that's all you're seeing, because it's a projection, right?
SPEAKER_00You're not seeing what's happening inside of that blue cell there. We have the data of what's happening inside of that blue cell. But if I tried to show that to you in a translucent thing, it would be like it would be like a projection through your whole body, all your organs and everything. You wouldn't see anything. So what we need, because we were evolved on a savannah to understand threats and opportunities in 2D plus time, what we need is we need somebody who can natively think in 5D.
SPEAKER_00We need to create minds that can see and think in 5D and then and then and then interpret it back in a way we can understand to us lesser beings. So that's so basically every microscope I've created in my career has been an answer to the limitations of the previous microscopes. From my hole microscope to the palm single molecule microscope to the lattice light sheet microscope to something we haven't gotten into, which is adaptive optics, which allows us to look deep when the light would otherwise get scrambled.
SPEAKER_00All of those techniques were specifically done because of bringing people in the lab, trying to do experiments and falling flat on our faces, embarrassing ourselves because of some limitation of the technology. So now we're at the point where we have incredibly performative microscopes that can reveal so much information that we can't interpret. Okay. So we need to invent the next tool to get us to the next step. And that tool is a five dimensional vision model that can interpret.
SPEAKER_00That data and interpret it and allow us to be able to understand. So we can ask questions through an LLM of the model. So what happens when a cancer cell is trying to work its way through a blood vessel to start metastasizing elsewhere in a tissue? And because the model itself would know what those words mean and know how to tie it together to this five-dimensional data set, it would be able to go through and pop up for us, pull away all the other stuff that's not relevant, and show us this cell doing this. And then I could say, what particular proteins are being expressed at the surface when an immune cell is attacking a cancer cell if you're doing immuno oncology?
SPEAKER_00And it would start to say, okay, well, here we saw these proteins. So I think it would revolutionize everything about it, imaging has been stuck as almost a non-starter in terms of being informative for biology. For it's as much as I've worked on it, I'm complete failure because we haven't moved the needle in biological discovery with these tools in any meaningful way. But if we can get over this last hump, it will be more impactful than determining the structure of DNA in the start of the molecular biology revolution.
SPEAKER_00And I have tried very, very, very hard to convince every AI person in the Bay Area and beyond to see that same vision and have failed miserably.
SPEAKER_01Why do thinking stop them?
SPEAKER_00I wish I knew. I I think I'm I cannot convey the problem well enough. That's part of the problem. Part of the problem is when they start to actually really understand the size of the and complexity of the data, they get scared because it really is, it's at a problem as big, if not bigger, than anything being tackled in AI today. And the and of course the other thing is there's a whole cacophony of voices saying, AI this, AI that, AI everything else, right? And so and there's so much AI going on in the VC world right now, and most of it is insane that they want to build virtual cells on the basis of just genetic information, transcriptomic information.
SPEAKER_00It's like having two screws of that car and trying to understand how the car engine works. It's doomed to failure. They are so full of ego and hubris and lack of insight about what how complex. Most biologists have no clue as to how complex biology is. If they sit and think about it for a while, they start to get a glimmer. But when they get into their weeds of doing, you know, their Western blots or their structural imaging or whatever, that goes away, right?
SPEAKER_00Because they're they're they're they're learnt, they've learned to be reductionists. Life is a one system that cannot be understood reductionist. It has to be understood holistically. And we have the tools to do it, but we all the tools accept the ability to analyze the data. And so that's a very, very, very frustrating step to be in. But I've been in frustrating points before. So it's just another one.
SPEAKER_01But I can see that when you explain it so clearly, Erica, I can really, really see the necessity of you know the tool of observation, because if you don't get the good data, how can you have a good Sure.
SPEAKER_00Tools tools is where it always starts. And you know, I a good example of how that's hampered other things is, you know, the 20th century was revolutionary for physics, right? Um, all the way up through maybe the 1960s, 70s. But now, now that basically, you know, you you have to have CERN and uh and the large hadron collider to to get to the the Higgs boson, the energies and the and the tools necessary to make the observations to put physics forward are so expensive, if not completely impossible under human scale, that physics is kind of stalled.
SPEAKER_00You know, I mean, you know, most of the physics being done on the other end of this campus are about applications of physics, not on trying to understand, you know, fundamental uh aspects of physics. And it's because they don't have observation to base it on, right? We're in the inverse situation. We have observations out the wazoo, right? But we don't have any way of of translate understanding it or starting to build models, theoretical models, on the basis of that information.
SPEAKER_00And so that's why biology is really, really exciting right now, because because it is so mysterious and the tools are at the are better than ever to try to unravel that mystery, except for the one we really need, AI, which is the one which is unfortunately inaccessible to us because the demands for other things in AI are bigger than ours. That if there's somebody out there who resonates with that message and has the resources to do it, I would love to talk to them.
SPEAKER_01Yeah. And what you would give advoice for like the younger scientists or people who pursue their career, like when they reach out to a frustration or, you know, the way they believe something could work, but then it turns out it hasn't existed yet. Is there any other voice for them?
SPEAKER_00Um find something that really interests you. Focus on that. Um, if it doesn't work out or if there's frustrations, the world is very, very large. The world of problems you can pursue is very, very large. Science is far too big now. Up until World War II, science was done by individuals. It was not done by institutions. Government realized that science was kind of critical. And so uh in the wake of World War II, this science advisor came up, Vannevar Bush came up with this endless frontier, this idea that we need to inject lots of research money into universities to strengthen science.
SPEAKER_00And that works to a degree, but when it's taken too far, what we've seen as a result is literally an exponential growth of science. But um what that means is that it's poorly spent. Okay. And there's far too again, the the groups of 50, right? If every professor replicates himself by producing postdocs and graduate students who want to become professors, and he does this 10 times over the course of his career, at least wise 50 times, then um then what does that mean?
SPEAKER_00Every one of those people want to be, so you get this, right? And all along the way, they complain about how they have no money. Well, they have no money because it's being, they have the same amount of money they ever had, but it's because the funding is going up at the same rate as the population of them are going up. But but the the problems are not greater than before. And so people get more and more into esoteric things or unuseful things, or a huge thing in biology now is is making atlases, okay?
SPEAKER_00Making like knowing where all the neurons in the brain are wired, right? And thinking you're gonna understand how the brain works on the basis of that extremely limited set of information, or these transcriptomes where they figure out what RNAs are are produced in different cells so they can understand different cell types, it's still not broadly broadly interpretable or not really pushing them. I always talk about moving the needle. You want to move the needle in biology. And because there's so many people doing in physics and biology and chemistry and everything, there's too many people.
SPEAKER_00And so what happens is is like I said before about the every in the good paper I publish allows a hundred pieces of shit to be published, right? Is is that is that the noise increases faster than the signal. Okay. And so as most scientists talk about how impoverished they are, how we have to, how we have to force put more emphasis on science, fund science more. I say the inverse. Um, I say that that the hundred people in the physical research division at Bell Labs produced more in a 20-year period than half of the country does otherwise, right?
SPEAKER_00So, so it everything's gone down the wrong path. I think if you were a young physicist or engineer, you would be insane to do a 10-year track position versus going to work on making life multiplanetary. Do you know how difficult a problem and how multidisciplinary a problem that is, and to be able to do it with the resources they have versus being a professor? I mean, you'd be nuts. But people, people have such blinders on about, you know, particularly in China, uh, every every every kid who comes to me from China, they want to be a professor.
SPEAKER_00And I'm like, why do you want to be a professor? It's a horrible job. You know, it's like, you know, and and it's because they've been taught from birth that there that there's honor and in it and and in and it makes your family proud if you're a professor.
SPEAKER_01It's like I really think that it's about storytelling to know that it's okay to not make other people proud. Yes, you need to be exciting life.
SPEAKER_00I I find over the course of my career that the pace of everything has accelerated. Um, you know, the the the information that's available online or whatever obviously is accelerated. Um, the innovations of various kinds have accelerated. Um uh the AI is going to, in ways we can't predict, increase that rate of acceleration enormously.
SPEAKER_00Uh it's it's it's exciting and it's scary at the same time. Um uh but um you you just look at already just, you know, it there's the good and the bad. I mean, everything in the end usually tends toward the good because humans are able to eventually learn what's bad and make corrections. But you know, uh the the the sort of polarization that happens in the country today. I I say to my kids, you know, um, you know, presidents, um uh uh, you know, uh revolutionaries, Lenin, Mao, whatever, they don't matter.
SPEAKER_00Technology, honest to God, everything that's happened to humans since agriculture, every major change, every major thing that has happened has happened because of technology. Life is by far the most complex matter in the known universe. We understand the interior of neutron stars that are thousands of light years away better than we understand what's happening inside a cell in our bodies. Um, and and so it is it is the biggest mystery in nature, right?
SPEAKER_00And so that's what's so exciting. And again, it will take innovations I can't even conceive of, but based on continued growth of technology and compute, you can't say that we won't necessarily get there.
SPEAKER_01And if you would say there has one cornerstone, good habit that incentifies you or make you centered to solving good problems, uh, what do you say that one It's hard to say.
SPEAKER_00I get again, the the habit is hard work, right? I mean, everything everything we do is is non-trivial and the only, you know, 99% of what you do um in any endeavor where you're doing something somebody's not done before, whether it's in a startup or or at work or if you're an athlete or whatever, is about screwing up. Okay. You you try something and it doesn't work. You try something and it doesn't work. So the only way you can get through all of the bad ways to the one way that works is by trying them all.
SPEAKER_00And the only way you can try them all is by working hard, right? So um, yeah, I'm a firm, firm, firm believer in the power of hard work.
SPEAKER_01Thank you for our time for today, Eric. Thank you for watching this new episode of Humanware Project. So that we can bring out more amazing content like this one to you. Please like and subscribe. Your support means the world to us.
At 28, Richard Branson bought an island he couldn’t afford. Decades later, it is still his home, and the place where he lives by the principles that built the Virgin Group. In this episode, Méline returns to Necker...
As a young teenager, Randy Schekman cycled to his local police station and told the duty officer he was running away from home. His mother kept borrowing the money he was saving for a microscope. Half a century later,...
Most microscopes have a limit to what they can show us. Eric Betzig helped to break that limit, and won a Nobel Prize in the process. In this episode, Méline goes behind the scenes at Eric’s lab at UC Berkeley. Eric is...
What if some of the best health and longevity secrets are not found in laboratories or supplements, but growing on trees right in front of us? In this episode, Méline travels to Hainan, China’s tropical island, with her...
At Golden Door, one of America’s best-known health retreats, guests are served a light tomato broth in the middle of the day, after a morning of hard work and before lunch at one o’clock. It looks simple, but there is a...