Why Great Labs Need Breadth | Professor Hod Lipson on Robotics, AI and the 2030s

What if the real innovation engine in Deep Tech is not the spinout, but the lab itself?

Professor Hod Lipson is one of the world’s leading robotics researchers – recently the subject of a major Financial Times feature on the future of robotics. He directs Columbia’s Creative Machines Lab, where breadth beats specialisation, portfolios matter more than grades, and academia’s greatest role is to explore what industry cannot yet touch.

In this episode, Hod explains why he built a lab around possibility rather than narrow focus, why he believes robotics is the real frontier after AI, and why the 2030s could be the decade when physical intelligence begins to scale.

For researchers, founders and Deep Tech leaders interested in how breakthrough labs are built – and how ideas become companies.

Learn more about Lab to Market Leadership: https://www.deeptechleaders.com

Follow us on LinkedIn: https://www.linkedin.com/company/deeptechleaders

Podcast Production: Beauxhaus

Transcript

0:00

AI that can do things in the virtual world, the very same technology applies to robotics. It just has to work on a, um, on both more com-- um, more co-- it's, it's harder because it involves more channels that have to move at the same time. It's not just text-to-text. It's not just, uh, translation or generate a sequence of proteins. It's actually move fifteen, twenty, a thousand muscles at the same time. Mistakes are expensive in the real world. Time is slow. Real-time is slow. You can't accelerate the real world, and there's energy limitations. So the real world is difficult, but the same techniques work. And I can see behind the scenes the amount of effort to scale those techniques that worked well for virtual intelligence and make them work in the physical intelligence. Uh, the same techniques work. We just need to pump in more compute, and this is part of why you're seeing all this, all this buildup of infrastructure, of, of, uh, computers and, and energy to support that. But it's clear to me all the curves show that this will be solved in the twenty-thirties, early twenty-thirties. And when that happens, all bets are off. Welcome to the Lab to Market Leadership podcast. Too many advanced science and engineering companies fail to deliver their innovations from the lab to the market. We're on a mission to change that. My name is Chris Reichhelm, and I'm the founder and CEO of Deep Tech Leaders. Each week, we speak with some of the world's leading entrepreneurs, investors, corporates, and policymakers about what it takes to succeed on the lab-to-market journey. Join us Lab to market journeys start in a lab, obviously, and they're generally started by students, by PhDs, by postdocs, by researchers in seek of inspiration and, uh, you know, potentially thinking about starting their own lab to market journey, uh, innovating their own platform, maybe building their own company, maybe not. And so my question this week is really around whether those in charge of the labs, those who are hosting this different group of research talent, are they able to create the kind of conditions where some of these breakthrough innovations are increasingly likely to happen? That's the question for this week, and it's kind of a leading question because I kind of already know the answer. And, uh, and my guest this week is an example, is a great representation that it is possible. It's possible, and it's likely. I am delighted to be joined this week by Professor Hod Lipson. Hod is one of the most distinct figures in modern robotics research alive today. He directs the Creative Machines Lab at Columbia University, and he is the James and Sally Scapper Professor of Innovation in Mechanical Engineering. His research spans everything from evolutionary robotics, self-aware and self-replicating machines, digital manufacturing, and AI for science. He has over three hundred and fifty publications, over fifty thousand citations. He has co-founded four companies, and a number of different companies have emerged from his labs, like Empire Robotics, or students from his labs have gone on to form some really great companies. Formlabs is an example. Uh, just one month ago, the Financial Times even had a whole spread on Creative Machine Labs, where they reference just the topic we're gonna talk about today. I'm delighted to be joined by Professor Lipson. Let's get into it Hod Lipson, thank you so much for joining me. Pleasure to be here I wanna start out with, uh, a quote. I just referenced, in my introduction, I referenced the FT article from, I think, the 18th of April. And I just wanna reference this passage in the article."The young roboticists took seat in a dim corner of the room. They spoke with reverence about the lab's director, a Columbia professor of mechanical engineering called Hod Lipson." Hod rhymes with God."He is the reason they are all here. Lipson makes robots conscious. Lipson's students make robots dance.'Everything starts from Hod, but it comes through me,' Zhang said," Zhang being one of your students. That's qui- you know, that's a remarkable introduction. Wow. No, this is, uh, I, you know, I only read this, uh, couple, couple of, uh, uh, weeks ago, and, uh, quite, uh, I was surprised to see what it looks like from somebody else's eyes. You know, sometimes- Mm … uh, uh, you don't see things, uh, that are there. It's funny, I told him when I met, uh, this reporter, I said, "My name is Hod. It rhymes with Cape Cod, okay?" And I s- I remember s- saying that because I wa- because people, uh, don't know how to pronounce it. I also wanna make people comfortable. They don't have to call me Professor Lipson or something like that. Yeah, yeah. Uh, and I, it's, was interesting that's where he took it. Uh, but I can see the origin of that statement. Uh, but, uh, it's, yes, it is funny sometimes to see how, uh, how, uh, uh, people, uh, see things from a different point of view. That's, that's quite fascinating. Now, when you… Now, I understand from the little bit of research that I've done that actually upon leaving university, your first choice, your initial step was not to go into research or to do your PhD, but you actually started a startup. Is that right? Right. Exactly. You were involved in a startup. I, uh, I founded a startup, uh , that was my first company. It was a bit too early. It was a GPS tracking, uh, uh, a startup. We wanted to, um, uh… It's a, a friend of mine, uh, that, uh, from, actually from college, and we, we conne- reconnected a couple years later, and then we said, "Let's-- There's this thing, a GPS tracking. Uh, let's do something with it." And back, that was '94. So back in '94, most people don't know GPS was very, very low resolution. It was about 100 meters, and you couldn't do much. But we said, "Hey, we can, we can track, uh, uh…" This was for insurance companies. We can see how containers are moving across the country, and if they're gonna make it on time or not- Mm and if they're going off course. So this is, this, 100 meters is, is, is fine for that. It's a big business model. We developed technology. It worked. And then we realized, uh, the hard thing is not developing the technology, it's actually getting it to market. It's getting somebody to pay for it. That is the hardest thing. The technology is the easy part. That was my first lesson. Uh, and, uh, and my conclusion from that, after doing that for a couple years, I said, "You know what? Uh, I, I'm done with this. I enjoy the technology more than the chasing the money." Yeah."Let me go into, into university and do just the technology." Little did I know that university is still all you do is chase money, and you chase idea- … and you try to get things into market. But at the time, it seemed like a, like a choice, but I think the reality is it never ends. It's always mixed. If you want to have impact, you gotta do everything end to end. That's… Yeah, that's right. You, um… Was it always mechanical engineering, and was it kind of always heading in the direction of robotics, or did that shift over time? No, that shifted. Uh, it, it was, it was, uh… I was really always interested in, in intelligence. Uh, and- Hmm … uh, in this dream of making it, uh, th- this maybe dishubris of trying to make a something as clever, as sophisticated as a human. Uh, and, uh, uh, it, it has a robotics angle. It has an AI angle. Uh, for me, for in the beginning, was a creativity angle. Uh, so it has all these pieces in it. Um- Yeah … but, uh, you know, mechanical engineering is, is one aspect. Probably, to be frank, mechanical engineering was the easiest way to get into this, this quest of, of, of making-- recreating life and, uh, uh, it's, it's sort of the body and, uh, that's, that's how it started for me. And so that's the-- And so that was the starting point for you. You know, people starting today, would you say they, you know, how many roboticists see it the same way? I, you know- Are more likely these days to see it as AI first? Um, it depends. As, as, um, I-- Well, again, this is perhaps where I differ, uh, from, from many people. I think the AI part is basically solved. Uh, and it's, it's, it's, it's taking care of itself at this point. It's moving forward. There's, there's many issues to solve, but- Mm-hmm … but the path is normal. When it comes to, to the body, uh, it-- we're still, we're, we're, we're decades behind, and there's a lot more, uh, a lot further to go, and therefore, there's a lot more for, uh, innovation at the bottom. Uh, when you're-- When the technology is mature, it's very hard to compete. Uh, it's very get-- uh, hard to get started. You have to join something that already exists. You have to… The, the bar is high, but robotics, the bar is low, and if you're gonna have, uh, start something, it's, it's better to start there. That's very, that's very interesting you say that. Um, I wanna come into the whys of all of that in a minute, but, you know, let's set this up a little bit. It-- In… You know, you and I have gotten to, you know, to know one another a little bit over the last few months, and, uh, and obviously in some of the recent research that we've done. If I think about your lab and the, you know, the environment that you've created or you appear to have created, it's very different from a lot of other academics or principal investigators that we've come across. The kind of traditional story or narrative, if you will, is I was in this lab, um, you know, I created this technology, I spun this out, I did this, and so on. You know, I built this team and, uh, you know, we eventually raised capital. With you, it almost feels like the lab itself is the real focus, is, is the real engine of everything. And, uh, you know, is that the case? Is it about the lab itself and then, you know, and then the other stuff is all benefit or the other stuff is secondary? Yeah, you know, it's hard-- this, uh, that's, uh, it's-- again, it's, uh, it's, it's, uh, interesting to see somebody else's perspective. So, you know, sometimes when you're inside, you're boiled into it, you don't see how it, how it's, uh, evolved. But I think, uh, there is a passion in the lab about doing something, uh, bigger. Uh, and you know, I th- and so-some of it comes from, uh, it's reflected in the name, I think. I've-- I notice a lot of labs are, are called in a very uninnov-innovative way. Uh, they're just called, you know, Bob's Lab, okay? It's usually the last name of the-- super, I think, dull and, and it, it just says, "If you can't even think of a name for your lab, I mean, come on, how creative can you be?" All right? This is to be always the number one thing with… Uh, so I, I really believe in, in setting a, uh, a sort of a, some people call it a, a transformative purpose for the lab. Okay, what are you going for? And this is when we, when we, uh, uh, you know, the lab, uh, in its current incarnation is called the Creative Machines Lab, and that's part of our mission. Uh, we-- I wanna make a machine that is creative. It's almost an oxymoron. At least for most of our, the l-last, uh, two decades, that was a, a, a, an oxymoron to think of a machine being creative. I mean, this, there's nothing can be more, uh, opposite. Uh, but that is our purpose, is to create a creative machine. And that-- so that gives a purpose to the lab, and then the lab has a-- it's almost as if it's its own entity. Uh, it, it's almost like a, a firm. It's alm-almost like a, like a, o-one of these re- big research labs that, that, that, uh, that, that have a transformative purpose, uh, uh, on their, on their flag to organize the world's information or whatever it i-it might be. And, and that I think gives people a, a goalpost, it gives them energy, it gives them a mean-meaning, and then good things happen. I think once you give people, uh, that power, good things happen. And, you know, a lot of people are talking about, uh, what's the future of work, and we might get into that, uh, later. Hmm. But increasingly, people are coming back to this idea that it all starts with a purpose, and once you have a purpose, the rest falls into place. But if you don't have a purpose, you don't know where you're going, you just want a nine-to-five job, uh, it's very hard to navigate. So I think that's, that's, uh, inadvertently, that's what I've created by naming the lab, by thinking about having a passion towards the impossible. Uh, but I can't say I planned it. It's not like I sat there and, "Oh, let me see how I'm gonna create a lab that, that spins off things." Uh, no. It just, it just, uh, I felt it. It just kinda happened that way. Yeah. I-is there a-- how do you- That kind of organized chaos. It sounds like, you know- Yeah, yeah… a little bit like organized chaos. It is. And, and you're providing this North Star. We're gonna create creative machines. What are creative machines? And you talk about the oxymoronic- Yeah … kind of- Yeah … relationship there, and that's- Yeah … you know, and that's quite cool and, and deliberate. Um, how do you resolve that against the need for, you know, to, to satisfy different stakeholders in your world? Uh, you know, you've got, you know, grant funding. You've got the university. You've got all these different pressures that you might be feeling in a person of, of your position I think, um, uh, you know, when I started my career, people told me, um, "Focus. You gotta focus on one thing and be an expert in that, and that's, that's how you get tenure." And I did exactly the opposite. Uh- Yeah … I said, "Okay, this actually sounds to me a very risky strategy, okay? Because if you put all your eggs in one basket, what if you're not the world expert? What if the, if the world experts, uh, doesn't, you know, you, you don't get along with existing experts and they… You know, it sounds to me so risky. I'm gonna actually, uh, do a lot of different things, and I'm gonna see what sticks." Uh, and, uh, and this has sort of always been my, uh, my approach. Also, it's not just, uh, a political thing, it's also you just don't know what's gonna work. Uh, i- there might be some problems that seems to be easy, but it's gonna be hard. There's some gonna be some hard problems that actually turn out to be easy. So, you know, it's, it… I don't have this, this, uh, insight to say, "Okay, I'm gonna, I'm gonna think about this a lot and I'm gonna choose the thing to work on and I'm gonna crack it." Yeah."I'm gonna g- I'm just gonna try out lots of different things, and I'm gonna see what happens." And, and, uh, it's easier to attract funding that way because you have a broader net. Uh, it's easier to attract students because you cater to a lot of tastes, uh, and students come in wanting to do different things. Uh, it's, uh, I find it's actually easier to collaborate because, uh, you could do so many things that there's always something in your toolbox that's useful for somebody else. So it ended up being a much better strategy to be a generalist. Uh, and it's also, frankly, a much, uh, more satisfying journey because, uh, I mean, if you would see the range of conferences that I go to. I mean, one day it's about driverless cars, the next day it's about food. I mean, it's c- it's, it's a, it's a comp- And it's just a much interesting, more interesting life. Um, and finally, I've seen so many colleagues who, you know, spend a career, let's say, developing, uh, putting all their in- all their expertise in one area, to only to be, uh, uh, to be victims of their own success. They became so successful, they went into industry. Industry took it over. Driverless cars is a great example. And the moment industry takes it over, there's nothing for you to do in, in, in academia anymore. Industry moves so much faster, uh, and puts so many more resources to it. And if they don't have anything else, their career is over. So I think having that diversity is also, uh, mitigate some of those risks. So I think any way you look at it, it's a good thing to do, but you have to have that mindset, uh, that is opposite of what academia cherishes, which is focus, focus, focus. Then where do you draw the line for your lab? And your lab has had a really good deal of success, um, you know, in, you know, in terms of its output, in terms of the companies that are emerging from it, in terms of the, uh, you know, uh, uh, reputationally. So what are… If it's not about that focus, focus, focus, then what is it about? It's gotta be about something else. Yeah, it's all about impact. Then where is the bar for you? It's all about impact. Mm. So, so, uh, you gotta do something that's gonna have impact, uh, that people care about. Uh, and it could be long-term, but it's gotta be something that's gonna affect The world in some positive way in the future. And you can measure-- The, the-- Thankfully, uh, there are now so many ways to measure impact, so it's not, it's not a, it's not a, a some, some hypothetical thing. You can measure clicks, you can measure citations, you can measure publications, you can measure funding, you can measure where do your students end up going. I mean, there are numerous, uh, metrics for impact. Uh, you can measure s- how many startups you've created. You can measure how many, uh, you know, uh, uh, uh, how many, um, prizes you've received. I don't know. There's, there's so many ways to measure impact. A-and so choose one and, but don't… But you don't have to necessarily focus on a topic to get that. Mm. That's, that's, uh-- Focus on the output rather than input, and I think that's a, that's a, um, a, I think a better strategy. Uh, now, given that, uh, it's not that, uh, you know… I think what has happened in our lab is if, uh-- There's also-- I realized this pretty, pretty early on, that when you have too much focus in your lab, you have a lot of politics inside the lab because you have a lot of people who are working on very similar things, and they're worried they're competing with each other because they're working on the very same topic. So if somebody has a great idea, they wanna keep it a secret,'cause if that secret leaks, then somebody else will, will move ahead'cause they're in the same area. If you have a diversity of topics, uh, suddenly people collaborate more because they're not competing. Everybody has their own turf. Everybody's working on their own thing, and suddenly they c- they're, they're much happier to collaborate."Hey, I know this skill. I'll help you. You're not competing with me. You're in a different area. Uh, let's, uh, work together." So, so in a strange way, you diversify, you get more collaborations. Th-these are, again, not things that I've, I planned ahead. I didn't think about this, uh, and, uh, designed it that way, but it turned out to be a much healthier dynamic. Uh, and, uh, and again, if you read that, happened to read that article, Financial Times, it talked about all these collaborations going on. In large part, it's because people are not very close in terms of what they're doing research in. And so they share the purpose, but they're not working, they're not competing directly. So, so that's another area. And, uh, last thing I'll say about this is that in academia, we have a unique, uh, gift, a unique license, and that is we're not-- we don't have a bottom line. Uh, we don't have to create profit. Uh, we don't have to work on something that is, uh, useful in the next quarter. Uh, we have a license to innovate. That, that, that's the gift. That's the deal that we get. We get public funds, and our mission is to innovate and have impact. That-that's if I have to summarize what, what, um- Yeah … uh, public funded, uh, research is, that's our mission. And so, uh, I would say if a faculty isn't innovating Uh, or isn't doing the b- So, so that gives you a license to do crazy things. Like I- Mm. Industry cannot explore things that are too crazy. They cannot work on self-awareness because, you know, that's a, it's a philosophical thing, but it's not something that you, that you need for next quarter. We have this incredible unique license to do crazy long-term research that might be impactful. So we have to do that. This is our license. We are the only people on the planet that are the very small set of people that have been granted this gift, and we better take advantage of it to its fullest extent. And, and this is how you do it. I see it almost as a obligation to explore. But I wonder if that thinking kind of permeates throughout all universities. Do you think that's a common philosophy, or do you think… You know, I don't know… I know a number of universities in the UK and Europe, but I'm not entirely convinced they bring that, that approach. They may say that, and obviously that's gonna attract students and so on, but I don't see them behaving that way. I don't see, I don't see it either. Uh, and this is something I, I as, as department chair, I have to sometimes talk to junior faculty and say, uh,"Don't listen to the old-fashioned advice about focusing and being an expert. The world is changing. You gotta move fast. You gotta think big. You have to have purpose." Uh, but not everybody does that. So, so, um, you know, people have diversity of, of approaches. And, uh, and again, you know, that's also part of academia. You c- if you want to do something in a different way, go ahead and do it. Mm-hmm. Yeah. The, one of the bars that must exist in your, in your group, and obviously the FT article referenced this, and our additional research referenced this too, um, the bar for talent quality must be very high And so you reference impact. It's important to have impact. So I'm a young PhD, I wanna work in Hod Lipson's lab, and I have a, uh, you know, I have a proposal that demonstrates impact. Are you measuring me at all in terms of, you know, my ability to realize that impact or to deliver on anything there? You know, what is, what is distinguishing, let's say, my application from, you know, another four hundred individuals who are also applying to the lab? Um, I think, yes, the world is changing, uh, and it's becoming easier in that sense. And so, uh, the way-- Again, different faculty look for different things, but when I am recruiting people, I look to see previous impact, prior impact. Okay, so I don't care… I really don't care about the grades. I don't even look at the grades. It's not that I look and I don't care. I don't look at it I don't read reference letters. Don't believe in that either. Uh, I've written enough of them to know they don't mean much. Uh, I, uh, uh, the only thing I care about is what I call the portfolio, okay? This is very similar to the way people are recruited in arts, in architecture, in fashion, okay? I wanna see the creativity in your past work. That's the only thing I care about. So when somebody applies, and they're on the shortlist for whatever reason, and usually they're filtered by the university, and we get some kinda shortlist of, of maybe 100, I Google them, and I see what comes up. That's it. And if the-- a portfolio comes up with cool stuff they've done in the past, then they're a good candidate. And if nothing comes up or something boring comes up, then it's not interesting. Now, the, the reason why I say it's easier than ever, because it doesn't matter where you are on the planet, you can create a website. You can create a portfolio. You can, uh… E- especially in the area of robotics and engineering, in my area, it's very easy to, to, to get some, uh, components from, from, uh, online and build something interesting. Uh, it's very easy to write software. Anybody can do that. A baby can do it at this point. I mean, it's-- there's no excuses. So I'm not asking you to build, uh, uh, uh, a building. I'm asking you to build something interesting or, or to create some interesting software, and I think even people wi-without, uh, academic background can do this. So, so that's what I, I also tell undergraduates. I tell, "Look, my…" I tell them, "My classes, you're not doing it for the grade. I'm gonna give you a good gra-- I'm gonna give you all an A, okay? I don't care if it makes you happy, but I can tell you the A doesn't matter. What matters is I'm gonna give you opportunities to demonstrate your creativity, uh, so that you can launch whatever it is you wanna do next. And, uh, but you better think about what you're gonna do that's gonna demonstrate incredible creativity so that you can put it on your website, on your portfolio. You can, you can, you can do it." And this is-- also answer the question of, "So what do we do when we have a AI that can do everything?" Well, the bar is now higher. You take that AI, and you do something that is so amazing that would be impossible to do last year, but you, you can demonstrate how you can use this AI to do amazing new things. Write a book. Uh, uh, build something that would be impossible to design, but you, you've done it in one semester. Um, just, you know, reach for the sky. So, so I think it's actually easier for those who want. Yeah. But there's no, you know, there's no You're, you know, you know, good grades aren't, are no longer enough. It's not about that Oh, they're not, they're not enough. They're not, they're not even relevant Yeah. Yeah. Yeah It's only your portfolio No, it is… Yeah, that's, it's, that's very interesting we're getting back to that And it's harder to shake, it's harder to fake a portfolio, all right? It is I mean, you can fake grades Yeah Uh, it's just a, a letter. You can, uh, you, you know, you can cheat, uh, on this exam and that, but you, but to, to, to fake a project that you've done is very, very hard. To copy somebody else's pro- And you can't fake that Yeah, it's too much And you can't fake that passion. You can't fake that passion You can't fake a passion, and it's also too much- And the insight and wisdom and learning too much work to, to copy somebody else's project. It's very easy to copy somebody else's homework. So I think, I think, uh, uh, so I think, you know, people sometimes say, ask me, um, "Is all this AI making us dumber?" And I think the answer is the opposite. The AI lifts the bar now so that you really have to be creative in order to move forward. And I think that's, uh, uh, the bar is higher, uh, but you can also do more, so, you know, we're all just, uh, rising up. Yeah It's not enough to be good in math. Yeah You actually have to be super creative The, y- your involvement at that point, so you create this deal with these researchers who come in, and they get into your lab. And roughly how many new researchers are admitted to your lab every year? So I'd say we have, uh, on average, I would say I, I admit, uh, one or two PhD students. These are the, the expensive, uh… And a, and a PhD is about five years, so, so there's a… That's like the, the full-time employee if you, if you think about it. There's maybe, uh, a two dozen master's students. Uh, master's students are self-funded, so it's a lot easier for me, uh, to do that, and maybe a dozen or two undergraduate students. Uh, and frankly, I'm very lax in admitting people into the lab, uh, and they sort themselves out, uh, very quickly, uh, because the projects are so open-ended that if you're not self-driven, you're not gonna get very far. I, I don't have the bandwidth to babysit every student. So we'll meet, uh, sometimes when people in groups, sometimes… And, and some students, uh, do such incredible work that, uh, they have to sort of earn my attention. And, and it's not deliberate. I don't, I, uh, didn't plan it this way, but there's just so many people. Uh, but if they do cool things, I get interested and we, we develop a one-on-one relationship, and that turns into a paper. But if somebody's just coming in to do, uh, to, to, to do homework and, uh, to s- uh, I'm gonna tell them every week what they're gonna do next, and they might fizzle out of it and, and it just doesn't work. And I, and I've had students like that and, you know, I just, uh, don't renew their contract the next semester, so to speak. And they might get a good grade. I don't like to give people bad grades, but I'm just not gonna take them on next semester. The, and, and they're- Yeah … not gonna have anything on their portfolio. So, so that's, that's the, that's the bottom line. So it's largely up to them to find their own way It's up to them. It's organic. But- Yeah … but I do say,"Look," I say, "Here's some…" I, I have a running list of ideas. Uh- Mm … and I say, "Look, you don't have to do any of these, but here's things to spark your, your mind." And, um, I-- this is a running list. Sometimes it's in my head, but I try to put it on paper. I walk down the street and I say, uh, "Oh, wow, um, there's, uh, all these, uh, parents pushing, uh, their strollers, uh, uphill. Wouldn't it be nice if they could stand on a little platform and lean forward and, and have a electric stroller? Uh, yeah, let me just make that a project." Bam, it's a project. We're just launching one of those, uh, possibly a startup to push, uh, strollers around town. So, so there's, uh, um, the ideas are, are plentiful. And again, I'm in a very privileged position where if I walk down the street and have an idea, bam, I turn it into a project, I have three people working on it the next day. It's an incredible-- uh, again, no industry can do that. Nobody can create a project tomorrow, uh, and launch something and have… And, uh, and, uh, so some of the, some of the projects are like that. They're very super-- That's a super practical startup almost from day one. And some things are like, uh, very abstract. Like I say, "Wait a minute. You know what? I think I, I cracked self-awareness. I have this idea." And it happens to me every other day. It doesn't mean much, but, but it's still, "I have this idea." It happens when you work on, uh, when a robot has to do two things at the same time, then it has to differentiate between itself and this task. I have some, some great, you know, brilliant brief insight, spark of insight. Uh, I create a project the next day, and there might be a student who bites into that. So some students want to create a startup. Some people wanna study philosophical topics. Some people want to break a record in, uh, AI doing X. And, and so there's-- Uh, people come in with so many different, uh, things that turn them on in terms of, uh, research that you have to have a, uh, a smorgasbord of, uh, of things out there. And people come in, and sometimes they say, "You know, I love this idea, but how about we do it this way?" And then, and then, bam, it's off to the races. So it really, it's, uh- You can see my style is, is unstructured. I just, I just have a lot of options, a lot of people, good things happen, and that's it. It's, it's almost like a numbers game. You just do enough of it and good stuff happen. But I think that's it. But I think it's a… You know, from the way you're describing it, it's a numbers game, but it's also a real diversity game too. You need to have the breadth in there in terms of the domain, in terms of the different disciplines, uh, in terms of the different skills and talent in the team, and the different perspectives. You've got to have that diversity in there, not just to kind of ward off the competition and the political behavior that, that is unhelpful, but also to kind of, you know, to, to help create new opportunities, and new ideas, and new thinking. So you have this home, you know, this place where everyone feels comfortable engaging. And I'm guessing over time, if you've built it up and you have, you know, it, it, it, it's gonna result in a more chaotic but potentially much more productive- Right. There, there is a, there's a flywheel going on also here- Yeah … that once you start engaging in many areas, you, um… I also go and speak in a lot of different conferences. The conferences are very, very diverse in the area. Like I said, could be with food one day, the next day could be automotive, uh, and then another day it's gonna be, you know, some, some other, uh, uh, uh, topic And when I'm exposed to all these different people, all these different challenges, uh, I come back with all these different ideas. And so then I create even more things, and, "Hey, did you know that there's this, uh, connection between, uh, you know, medical diagnostics and this area," and, uh, "Hey, we can do something with food printing that, that…" You know, some crazy, crazy connections that, uh, spark even more projects. So I think there's a death spiral that happens if you're too focused. Yeah. You become very narrow in how you think. You now speak to a smaller and smaller group of people. You enter e- echo chambers where you're all thinking the same way. Now your projects are even narrower, and, and, and, and I can see this death spiral hit some faculty, um, and their research becomes… until it either dies because they cannot- Yeah get any more funding, or it continues, but their impact is negligible, and I also see that. I see senior faculty whose, whose impact is dismal because they've, they're so focused and so narrow and so, so s- um, rigid in the way they define what's interesting and what's not interesting, um, that, uh, they just, uh, they can't pick up out… They can't, they can't get out of that spiral. So- I get it … so it's helpful to keep it out, keep it loose, and not take yourself, uh, too seriously. I, I remember- Yeah … there's one, one, uh, advice that my postdoc advisor, uh, gave me, and I have to give a lot of credit to Jordan Pollack at Brandeis University for telling me this. He said… I, I had a, I had a, a big paper in Nature that kinda launched my career working in his lab, and he said, "Just because you had a big paper, don't think that all your papers have to be big. You have… Do- don't forget about the small, insignificant things." And, uh, I kept… I, I remember him saying that, uh, as, as I left, and I keep working on small things. And you know what happens is that some of the small things end up being big, and some of the big things that I think are gonna have a lot of impact have zero impact. So the truth is, even if I wanted to focus, I wouldn't know on what because I can't tell what's gonna have impact and what's not gonna have impact. Yeah. So, you know, I'm sort of, "Okay, you know, I can't tell. Let's, let's do it all, see what happens." Yeah. But in the context of a lab like yours, just to play devil's advocate for a minute, is there a danger of kind of cannibalizing some of your other ideas? Or is that just kind of, you know, natural Darwinian- Darwinian action and- Yeah… and it's to be welcomed? Yeah, it's welcomed. It's a free market, if you like. Uh- Yeah … uh, let's… yeah, uh, we, we attack the same idea different ways. I, I, you know, sometimes, uh, uh, you know, um- There, there's, uh, even a little bit of competition on, uh, maybe on equipment, on, uh… There's certainly competition on computing resources. I mean, yeah, there's a little bit of a Darwinian element. You know what's the biggest thing that people compete over in my lab? Space. Hmm. Okay, so we are… So in Manhattan, space is a premium. We're on Broadway. Uh, you can't just, uh, build this, uh, you know, use five desks to build your thing and then walk away and come back, uh, a week later. If you come back a week later, half the parts of your system will be gone, and, uh, your thing will be pushed over to the left, and the new thing will em- emerge next to you. It's a jungle, and that pressure, I think, keeps people also, uh, you know, there's people bre- breathing down your neck in terms of space, equipment, compute, uh, and that I think lights, lights the fire a little bit in a good way. Yeah. Yeah, it provides that sense of urgency to go- Yeah … in the way that New York can. Yeah, yeah. And New York does. It's very- Exactly, so I think actually, so, you know, moving from, uh, Cornell to, to Columbia- Hmm … uh, you know, that, maybe we'll talk about that later, but that's one of the immediate things that happens to you. Yeah. At Cornell, you can, you can think that you have time to do stuff, and you go to New York- That's interesting … you see you don't have any time to do stuff. Whatever you wanna do- Yeah … has already been done, and maybe you can do it. You know, there's a little bit more of that pressure- Yeah … and you cannot escape it. You feel it every day, all the time. I've got one more question related to this, and then I wanna kind of go through and understand how you kind of arrived at this position in life with this, you know, philosophy. My last question on this, you know, a lot of the different things with which you're involved, with which your lab is involved, you've got, uh, and I'm just reading from my list here, self-replicating machines, uh, metabolizing machines was one of them which, uh, which I'm interested in hearing about, food printing, uh, jamming grippers, uh, r- uh, you know, you mentioned robotic self-awareness, soft robotics, AI for science. You know, the, a lot of these, um, well, they each have a slightly different time horizon in terms of realization. And so I get the point about impact. Um, is there… But you know, when you're, you know, looking at someone that could, you know, or You know, when you're evaluating a proposition that could have an impact in, let's say, two decades from now, um, versus something that could have an impact in, let's say, the next decade, you know, in the next five years, does that make a difference? Is it more or less easy to evaluate or to deal with those things together? Are you trying to, you know, somehow group things in terms of time horizon? Yeah. I, I think, I think you're trying to grasp at a methodology that I might have, which I don't. Hmm. Yeah. I don't have any methodology. I just have numbers. Yeah. And, you know, as long as it's, um… I don't want to… I, I want to have, uh… So maybe, maybe if I have to think about it, if I really think hard, it's not, it's not a process I go to, but if I try to deconstruct myself, I think, look, if it's going to be long-term, it has to have a huge impact. Hmm. If it's going to be a short-term, then it can have, uh, sort of, uh, uh, lesser impact because it's immediate. There's less risk in it. The stroller is a short-term thing. Uh, it's gonna have impact, but it's not gonna change the world. But, uh, but self-aware machines, uh, or a new way to, uh, create food is a huge impact, and that can be a little bit more forgiving. You can go long-term. So I think there's a, the diversity is probably a long tail spectrum or something like that. Yeah. But it's not a- Yeah … it's not a calculated thing. It's just, uh- Yep. Uh, it's just- It's a feel. It's a feel. It's a feel. It's, uh, it's probably, I would say VCs have a little bit of that gut feeling. Okay. They might say, "Okay, this is a high risk, high, high payoff. Uh, this is a low risk but low payoff." And of course, everybody wants low risk, high payoff, okay? But that's the, doesn't exist. So, so there's, uh, the, the, all those low, low-hanging fruits are gone. So I think, uh, uh, we're looking for those things, and there's a… If you think about it, this, uh, this economic Pareto front, and yeah, I am looking for something that's huge impact and immediate. Yeah. Hmm. That would be great. That's hard to find. Sure, sure. Yeah. It's, uh, been a unicorn. Tell me a little bit, you know, just kind of walk us through the background, Hod, if you wouldn't mind, and kind of, you know, help us understand, I guess, you know, you know, what are the kind of key events or key individuals, you referenced, uh, you know, one earlier- Yeah you know, that have helped shape your thinking a little bit and you feel have helped you arrive at where you are today? Oh, okay. So, so, you know, hard to, uh, to, uh, go through that process. I would say, uh, academically, uh… I, so I grew up in acad-- in a, in a academic family. My father was a professor of physics. It probably shaped, um, my thinking very much. He, he was, uh, uh, engaged in physics, still is, is retired, but he thinks about it all day long. Uh, he, uh, did his PhD in Cambridge and, uh, was all about, you know, changing the world. He was involved in the race for liquefying helium and all these, uh, uh, grand stuff that, that people work on, and I got that. But he's also a hobby carpenter, and so he was, uh-- I-- all day long, saw him building things. He's a very, very much an experimentalist. I think that gave me a lot of, uh, uh, uh, a lot of, uh, sort of, uh, this initial, uh Fascination with science and with making things. And that's, uh, that permeates a lot of, uh, what I'm doing today. And I also… You know, when people talk about their early life as a scientist, frequently they refer to, "Oh, I always wanted to take things apart and understand how they work." I never did that. I never wanted to take things apart. I always wanted to build things from scratch. Uh, and I would-- Uh, Lego was my favorite toy, and I would be in the, in the-- my dad's, uh, shop building things. So building s- thing is my passion. I really wanna… And, and, and again, that, that, that recent, uh, uh, um, Financial Times article is all about building things out of building blocks. And I think, um, my Lego background is actually part of that, uh, idea that you can build a huge diversity of things from a small set of building blocks if you got the right building blocks. So I think, uh, so, so that's part of where it started. But then, you know, career, uh, meandered through lots of different, uh, paths. I was a engineer in the, uh, Israeli military, uh, for many years as a navy, in the navy as a navy officer, designing mundane things, uh, not necessarily weapon systems. And, uh, as a designer, as any engineer will tell you, after doing two, three designs of the same thing, you start thinking, "You know, there sh- there ought to be a better way to do this. I mean, there is no way that I'm gonna spend my life designing things, door handles, brackets. I mean, this is… Shoot me now. Okay? This-- There has to be a better way to, to go through all." And I said,"Okay, instead of designing things, I'm gonna design something that designs things for me, and design it once, and it will design everything. Instead of building things, I wanna build something that builds things for me. Uh, I wanna go one step meta on all-- on the design and build." And if you think about it, that's AI and robotics. Yeah. That's digital manufacturing. And I remember that spark, you know, I'm done doing things one by one. I wanna solve it. And what I didn't realize- When did that hit you, if you don't mind me asking? I- When did that kind of… When did you realize? I realized sometime during my early engineering career where I was just sitting there designing things and then, and just, uh, saying, "This, this-- I, I'm not gonna be able to do this for my entire life. There has to be something bigger." So some people go into management, which is one way to do it. They say, "I'm not gonna design things. I'm gonna manage 100 people. They're gonna design things." That's one-- It seems to me like it doesn't scale either. You can, you can only go so far. But if you can make an AI tool that designs things for you, not only you can design a million things a second, but you can solve things that you-- we don't know how to design. Like, you can design a, a solution to cancer. We can design a better battery that we can't design. So I didn't realize it at the time, but, but this is not just about doing more and making things faster, but we s- about solving problems we don't know how to solve. It's the ultimate leverage. Yeah. And I think that's, uh, that's, that's what it, uh, it, it how it sort of started for me. And since then, it's been a race from that. And different people, like my, my PhD advisor, uh, Moshe Shpitalni , uh, Shefi as we used to call him, uh, he was very open to those-- that kind of thinking, and that gave me confidence. Uh, and, uh, he was also… I, I, I, I learned from him the, the, the indispensable art of academic politics, okay? This is how to talk to other faculty, how to… I, I… He's a master, and I learned that from him. Uh, you cannot survive in academia if you don't understand that. Um, and, uh, from, uh, Jordan Pollack at Brandeis, uh, I, I also learned to, to be bold and think big. He, he sometimes, uh, said things that upset his, uh, academic colleagues but had, uh, bigger impact, uh, than, uh, most faculty. So I learned that, uh, from him. I, well, I did a postdoc at MIT. Uh, Nam Suh, department head in chemical engineering at MIT, he showed me-- I remember this, this graph. He said, "You can…" I remember one day, I, I burst into his office with a bunch of ideas, and he showed me this. He drew on the, on the, on the blackboard. Uh, they still had blackboards back then. And he said, "Look, here's the chart of impact versus, um, uh, impact versus difficulty," or something like that. And he said, "Look, you can do, uh, you can do very basic projects like E equals M, discover new physics, huge impact." Or you, uh, or you can do, um, very super applied things like invent a new, uh, you know, sliced bread and also have a lot of impact. But there's a ton of thing-- But both of these are very difficult. The easy part is to do things that are in the middle. They have a little bit of impact, and they're not very unique That's where everybody goes. Don't go there. You go to one of these two extremes. And I remember he showed that, and I walked out because all of my projects that I proposed were in the middle. They were kind of interesting, cool, a little bit of impact. Not gonna change the world. Ton of things to do, but not gonna have that much impact. He said, "Don't-- That's, that's where everybody ends up. Don't go there." So, so I went out of his office with all my ideas still in my pocket, and, uh, and I, I thought,"Okay, I'm gonna, uh, either… So I'm gonna do one of those extremes." So everybody gives you something, uh, and, uh, and, uh, Jordan told me, "Don't, uh, uh, don't, uh, uh, be afraid to do small projects." My dad told me,"Never decline a speaking invitation, no matter how small, because that's how you build a community." Uh, I remember that, um, and, uh, you know, everybody gives you a little piece. Yeah. The, um… So for your MIT, uh, uh, professor's, uh, point, what is the big, y- you know, what is that contribution or that, that, that project of impact that you feel best represents you? Or to your point earlier, is it almost a meta thing and you've created a group of people or a lab that are now having all of those impacts? Well, okay, so I think, I think that's, uh… If I tell you my biggest impact that I've educated people that have had impact- Yeah … that's a little bit of a cop-out, I think. Yeah. I'm gonna, I'm gonna revert to that- Yep … if I have nothing else. Yep. All right? Yep. Yeah, yeah.'Cause I'm gonna say,"Okay, they did it." Yeah. But if I have to list my own-- First of all, I'm, it's not over yet. So I don't wanna say, "Well, my biggest impact was…" No. Okay, I'm, I'm still going, and I have some ideas. But I think it's changed over the years, all right? So I started, I, I thought digital manufacturing, three-D printing was the biggest impact, uh, in the beginning. I developed three-D printing, um, when it was not cool. It became cool. People think it's passe, but it's actually growing twenty-five percent per year. It's, it's the new way of making things. Um, so that was a, um, big impact. I created a field called bioprinting, printing with live cells. When I started, people said it's crazy, it'd never work. I couldn't get funding. Uh, then, uh, created a company. Now there's a field of bioprinting. There's journals. There's conferences This is the way we're gonna create replacement organs. That's gonna be huge impact. Um, and I think it's gonna play out. Food printing hasn't happened yet. I've been at it for 20 years. People are not buying into it. I'm telling you, it is going to be, uh, eventually Uh, you know, robotics and food are gonna come together. There's no doubt food is a big part of our life. Manufacturing and software is a big part of our life, eventually gonna come together. We'll have automation in, in our kitchen. Hasn't happened yet. I know it's gonna happen. Uh, self-awareness, uh, is my long-term biggest, uh, sort of thing that I've been at for a long time. It was a joke in the beginning. It was not even a joke, it was a taboo in the beginning. Like, I couldn't tell people, I'm interested in this question of can machines be self-aware? I mean, it was-- it's more of an oxymoron than creative machines. Uh, the idea that machine can, can think about itself, I mean, the, the-- where do you even start thinking about something like that? But now it's, it's actually not so crazy, and it's going to-- And before you know it, uh, we'll start thinking about sort of machine psychology and things like that. So this is all gonna happen. Uh, I f- I, I feel that's a, a big impact. I mean, uh, there, there's a, there's a bunch of things. Um, automating science was another one. We were almost the first to create a machine that can rediscover physics laws from observation. We've done some work since then. Again, now it's a whole community and, and so oddly, once a whole community flocks in, I leave. I kind of-- I know when a-- it's-- it becomes too crowded. Uh, it's frustrating because you peop- you see a lot of people reinventing things you invented in the past, and you're saying, "But…" And then I, I become this old guy who say, "Ah, I've already done this before." I, I don't wanna be that person, so I just gotta move on and do something else, and I'm happy just to see some citations come out of it. Yeah. Um, I've got one last question for you. In one of our very early conversations, uh, you mentioned-- You know, we were talking about AI right now and, uh, and what we might see in the 2030s, uh, as it relates to robotics. And so if we think AI, I-- and, and I think you said something along the lines of, "If you think AI is big right now in the 2020s, wait until the 2030s and robotics." Can you expand a little bit more on that? Yeah. So, so, so, um, most of our world is physical. Uh- Mm … most of our things we do all day long is we bump into the physical world, we move things around, we do stuff, we, uh-- and by l- by, uh, you know, by large, this has remained, uh, analog. We are just, uh, we are not, uh, competing with machines in the physical world, and, uh, uh- When that happens, uh, our lives are gonna change so dramatically that it's hard to imagine right now. And what I'm seeing is that the same techniques that are being used to create virtual intelligence, machines, uh, AI that, that agentic AI if you, if you, if you like, AI that can do things in the virtual world, the very same technology applies to robotics. It just has to work on a, um On both more com-- uh, more co-- it's, it's harder because it involves more channels that have to move at the same time. It's not just text to text, it's not just, uh, translation or generate a sequence of proteins. It's actually move fifteen, twenty, a thousand muscles at the same time. Mistakes are expensive in the real world. Time is slow, real-time is slow. You can't accelerate the real world, and there's energy limitations. So the real world is difficult, but the same techniques work. And I can see behind the scenes the amount of effort to scale those techniques that worked well for virtual intelligence and make them work in the physical intelligence. Uh, the same techniques work. We just need to pump in more compute, and this is part of why you're seeing all this, all this buildup of infrastructure of, of, uh, computers and, and energy to support that. But it's clear to me all the curves show that this will be solved in the twenty-thirties, early twenty-thirties. And when that happens, all bets are off. Uh, it's all bets are off because so much of our world is physical, so much of our supply chain, our livelihood, everything is physical. And, uh, and it's not gonna happen overnight. Uh, there are, uh, b- Knowing how to use your hands in, uh, construction is different than being a dentist. It's different than, uh, folding laundry. There's a huge diversity of ways in which we're physical, uh, so it's gonna take time to unfold. But, but, uh, I see it as… I, I, I can see in your eyes that you are, uh, thinking it's bad, but, uh, it's act-- or terrifying. But I think it's actually the, the again, from I-- from my point of view, the, the leverage that we'll have as a species is enormous. We will be able to build new things that would have taken us a century. We can do it in a week. I mean, this-- we, we'll be able to build things to solve physical problems that are ailing us, uh, at, at an incredible rate. Uh, and, uh, we'll, we'll have no excuses. So I, I can't wait for that to happen, but what we need is the self-confidence to do that. And I find myself, uh, the most important thing that I teach, if you like, as a professor at this point, is con-- self-confidence. I say, "All these are happening, all these tools are gonna be at your disposal. Students that graduate now, you're gonna live your entire career, your next fifty years, if not more, you're gonna be in a world where you just say what you want, a machine will do it. So you better have the confidence to make the right choices. If you cower and wait for somebody else to tell you- Uh, it's gonna be game over. In fact, the machines will decide for you. So you better, you better wake up, have-- think about what you want, and be able to articulate that with confidence because you'll have the tools to, to, to, to make that happen. So th-this is true for everybody, but, uh, but, you know, not everybody's ready for it. So I think, uh, I think we have to, to get ready. That's a-- And I don't think we could have designed a more perfect ending to this show. That's perfect. Wow. That's exciting. It-- That's hugely exciting. Hod, thank you so much for joining me today. Thank you. This has been great. The ple- Really, this has been fascinating to listen to all of this. So much to pick through here, so I'm so grateful. Thank you again. Thank you, Chris. You've been listening to the Lab to Market Leadership Podcast, brought to you by Deep Tech Leaders. This podcast has been produced by Beauxhaus. You can find out more about us on LinkedIn, Spotify, Apple, or wherever you get your podcasts.