Transcript: "Anthropic's Wet Lab, Artificial Wombs in 24 Months, and Scalable Gene Editing | MOONSHOTS Live #297"

I got a little scared backstage when Palmer said, "There was a message on the screen saying I shouldn't say that ." I thought, "For God's sake, whatever it is, it's not about Colossal." Please don't say that. So, I remember at Abundance 360 last March, right before you came out, Elon was there. And you said, "Yes, I want a tame woolly mammoth." How are you doing with this project? Oh, we get these kinds of requests all the time. Yeah, yeah, but there was a great episode, I think, of " American Dad" or "Family Guy," where they explained... Did you see that? I don't know if you saw this. But they explained CRISPR there, and it was amazing. And that was cool, because at the end they say, "Yeah, and the woolly mammoths turned out to be the size of dogs." This is what they really were. So the fossils were just a mistake, right? It was cool because it explains CRISPR. But I said the second most popular request

[00:01:00] we get is for mammoths "in a cup." In a cup? Mammoths the size of a teacup. Like... Yeah, everyone wants a mammoth like that. They want to take the mammoth so they can put it in their handbag, right? So we can create new melanin, make it pink, and then Paris might want that. So we're not working on that right now. We're making good progress with the mammoth project, so... Amazing. So, let me begin. You founded four successful companies before Colossal: in gaming, defense, artificial intelligence, and mobile technology . And you knew nothing about synthetic biology. Me, yes. Before you started. The company is now worth over $10 billion. Why did the restoration of extinct species become your " billion dollar project"? And was the lack of experience in synthetic biology an obstacle or an advantage? Oh, I think that's a huge advantage. I think it gives me the opportunity to go into offices, Palmer

[00:02:01] talked about that a little bit too. For example, hiring people to replace you and hiring people who are significantly smarter than you, right? So what's great about this for me, from my perspective, is that I have to deal with a lot of the same problems, but at the same time I can sit down with, for example, Beth Shapiro and learn about isolating ancient DNA and how not to do certain things. But then, at the next meeting, I can learn about where we push the boundaries of multiplex editing and how many edits we can do at once, right? So I would say that, you know, I was able to, and probably throughout my career, I have been able to ask the right questions because I am very inquisitive. But I really stand by that old adage: surround yourself with the smartest women and men; just asked them a question, you know? So I go to meetings just to ask questions, and, you know, 90% of the time people probably think, " Okay, if you knew more about biology, this meeting would go faster." But 10% of the time they say, "We never thought of it that way , did we?" Because this is, by the way, such an important lesson for

[00:03:00] all of us entrepreneurs , right? Just because you're not an expert in a field doesn't mean you shouldn't pursue it if you love it. If you attract the best talent around you. Yes, George Church, and that's the only thing I would brag about, and something I'm very proud of. George Church says I'm the best student he ever had. Because I literally bombard him with questions: “ Hey, my favorite time of year is the holidays when no one works.” Because I can spend hours on the phone with George Church and just discuss the possibilities of synthetic biology . Who is George Church to us? Oh, if you don't know George, George is the father, you could say the father of synthetic biology. He's the head of the Department of Genetics at Harvard, and, you know, a lot of the next-generation reading and writing technologies were invented in Church's lab. And his laboratory is very fruitful. It has spawned a large number of companies worth billions of dollars. This is probably the most active startup biology lab in the world. And he too, he's under two meters tall , suffers from narcolepsy and he's

[00:04:00] terribly funny. And he's such a great guy. Yes. And, I mean, he's the most cooperative person in the world, right? And that's why he's also a fan of open source and strongly supports the democratization of technologies, such as genome sequencing. He wanted to bring the cost down from billions to $100, and he was very active in that area. So he's probably the most collaborative co- founder I've ever worked with. Dave? Yes, I'm very interested in the business model in biology. I think we all know that among all the future applications of AI, solving all the problems with disease and overcoming pain is the number one priority. Uh, the business model in biology is always... Yes. So, I'm really interested to hear your thoughts as a serial entrepreneur who came to biology from another field. My first question about the business model is about the basic models. We just saw that the Astra can drive without any additional settings. Yes, yes. Will it be able

[00:05:00] to work with biology in the same way, or will you say, "No, no, no, we need to create our own"? Still to come...So, on the base models now: we were lucky enough to collaborate with both OpenAI and Anthropic and get early access to their developments, which was great. For a long time, we used them as a kind of intelligent middleware, you know? If you look at it from a software development perspective . We linked lab logs and Jira. Perhaps the most difficult thing we did at the company was training scientists to work in Jira. This is sometimes more difficult than reprogramming sensors. So for a long time, you know, LLMs before the advent of advanced models were very good at writing term papers, right? But they weren't very good at ancestral state reconstruction or comparative genomics, were they? They just didn't work there, and that would have been extremely useful. So for a long time we used them as layers of infrastructure middleware so that we didn't have to hire Deloitte or Accenture to build all these systems and reporting, right? I think we've been very successful in

[00:06:00] implementing this for a long time. Now they have advanced far enough to be able to conduct simulation experiments. There are a lot of companies like Laila and others that are trying to build lab automation around this, you know? We're pretty cautious about how and when to invest in this category, but I think you'll find some interesting ideas about connections that are still really language issues, right? For example, if you search the literature for information about mice, let's say you want to create a new therapeutic company for humans, and you look at all the published works on mice, you will see that in many of them the same gene is called differently, classified differently, and the experiment is conducted slightly differently. So if you try to replicate the same experiment, you'll get a 40 to 60% failure rate, which is just awful, right? So I think AI will be very useful in the next wave of development, beyond

[00:07:01] small molecule drug development, and will help bridge this gap and unify terminology in research. There are times when we do work and later find out that there was already a peer-reviewed article on the issue we are trying to address, but because of the way we searched, we were unable to find it in PubMed or other databases. So I think in the short term AI is going to be the connecting link in communication, and you're going to see it—I think Anthropic announced it a few days ago, right? Where they, and Ginkgo along with others, are going to test this. You will still need to conduct laboratory experiments to verify and confirm the results. But modeling designs for a variety of different experiments and helping to creatively plan future studies is where we will see AI applied to biology over the next 3–5 years. How about this: if we're talking about text data and research results, analyzing thousands and thousands of previous tests is entirely the territory of large language models. But what if my input

[00:08:00] vector is just a gene sequence? If I upload this to Anthropic right now, I'm pretty sure I won't get anything useful out of it. Yes, but on a larger scale. So thank you for this eyeliner. This is something I fundamentally believe in—the importance of our global biorepository system that we are currently deploying. We are trying to implement a model, a kind of "Noah's Ark 2.0". It's not that Noah didn't make it through the first stage, but, you know, at least our approach is a little different. We do ATX sequencing and other things that he didn't do. So we're trying to work with governments around the world to create local biorepositories to get all this data so we can do T2T sequencing. And so, agreeing with you, I don't think that one genome—well, you can't just upload it into Mythos and it says, "Oh, well, if you make these six changes, it'll become immortal, right?" But I think if you look at, you know , thousands of genomes across all these evolutionary lineages in birds and you see that

[00:09:00] they're not susceptible to a lot of diseases, you're going to pay attention to that , right? We do this on a small scale, for example with P53, jellyfish and other things. We look at the known outcomes of species, and then try to trace them back through the tree of life. So we are already doing this in one of our companies, and we are very inspired, and the first results are positive . But, as you said, you can't just take a genome and get a ready-made answer on how to fix it and make it perfect. However, I believe that this amount of data from a comparative genomics perspective at scale will allow this to be achieved. Mm. But I also believe that I would say that the dataset is more valuable than the model itself. Yes. Yes. Yes. Because I have this dataset. We have emphasized this so many times. Yes. Alex. OK. Yes, I would like to talk a little bit about deextinction. I realize you have over four spin-offs now, but, despite everything, deextinction, to me, is what you and Colossal are, maybe Yes. And there are certain things

[00:10:00] we shouldn't talk about. And we won't talk about them . So, deextinction. There was a Russian philosopher, Nikolai Fedorov, at the end of the 19th century, the founder of a philosophical movement called " Russian cosmism," who argued that the ultimate goal of humanity in the development of science and technology would be to develop technology to resurrect every person who had ever lived. Russian cosmism, you can search. So , deextinction, in a sense, you're not doing it right now. Okay, I understand. So, in a way, by de-extinction, by resurrecting various model species, including the woolly mammoth, you are the first company on Earth, as far as I know, that at least has a plausible business model or a technical trajectory for trying to recreate the entire historical biosphere. Perhaps not just every human who has ever lived, but every

[00:11:00] non-human organism that has ever existed. True. In the age of superintelligence, when great challenges are being overcome one after another, do you believe that the common task of humanity, as the 19th century Russian cosmologists believed, is that we will eventually have the technology to regenerate every organism that has ever lived, or at least every human being? Do you foresee this becoming possible? I am not. I think that with DNA synthesis, predictive modeling, and, you know, certain components of synthetic biology , we can get pretty close analogs. But I don't think so, with two big caveats. First, just so you know, Colossal does not use any of our Colossal technologies for humans. So we won't do that, even though I think our technology...... is not suitable for that? Is that what you mean? I think when you start working with people, you need different investors, sometimes and definitely different management, right? And you go through a

[00:12:00] different process, and... And more patience. And more patience, right? That was actually a really good piece of advice I got from Bob Nelson. Bob Nelson said early on, "Hey, don't apply any of this to people for a long time because you'll actually end up in code freeze mode with the FDA and it'll just be endless monotony." " Decide, develop technologies until they reach a plateau before moving on to another." But going back to your question, I think that fundamentally... most organisms ... remember, these are most organisms that we know of. Very few people realize this, but less than a hundred tyrannosaurs have been found ...I just want to go down that path or scare you. There are less than a hundred tyrannosaurs ever found, right? But then, you know, there were billions that were said to have lived on Earth at different times, right? So very few things leave fossil remains. So I think it's more likely that we'll be able to create life in terms of programmable biological systems the way we want it to be, rather than just bringing back everything that once existed, because I don't even

[00:13:00] think we know , do we? I think it's quite likely, based on future artificial superintelligence, that we'll get to the point where we say, "Oh, we're going to engineer that." And to your question, maybe it existed, but we'll just never know because there are no fossils left, right? So, I think we would rather create life for our own benefit than try to revive something. Even if Colossal or some of our technology, like cloning and genetic engineering, can revive something, as you know, there are environmental factors, epigenetics and all these other factors. For example, if we could clone Mozart, it doesn't mean he's going to show up and say, "Oh, I'm going to solve the problem of where the terrible trajectory of music has gone." Well, I mean, just maybe a follow-up question. So, in terms of Mozart himself, we really do know a lot about Mozart's life. So, let's say, if we really wanted to resurrect Mozart, we would have no problem recreating the conditions of his childhood. I...up to a point , right? I...up to a point, right? So, I don't ...by the way, just to be clear, I'm not in any way encouraging this

[00:14:00] line of questioning, but we take people and we grow them. It was like you were going to build a Michael Jordan- or LeBron-level dream team, and that scares me a little. So, I think it's quite possible that if we could... it's not entirely ... it's 100% accurate to say that we understand their genetic predisposition and ability for certain traits. And under the right conditions, if you want to go, you know, if you want to fully apply the simulation design to them and put them in the right environment , then your Mozart 2.0 could probably be even better than Mozart 1.0. But Ben, let's go back to the animal world, because the main thesis right now is that AI systems can design a genome sequence that reflects a phenotype. So if you want an animal that's bigger, it's not 100% ready yet, but... that's ... that's the goal, right? If you

[00:15:00] need an animal with a longer snout or an animal with wings. So, as I asked you on the FII stage, do you know if it's possible to create a Pikachu? Yeah, it seems like a weird fan favorite, you know. I think people are more accepting of Pikachu than these big genetic human camps that you think of. Brand types. Yes, yes, yes. Very true . So, I think people are more open to this model of genomic engineering than... But maybe to sum it up, it seems like this isn't a purely technical objection. It is more a concern about social, political and regulatory issues. Right. Right. Yes. And ethical. So, this is where I would like to intervene. You know , Stuart Brand made this famous comment, Peter, you wrote the title of your book. He said, "We're like gods, so we should start acting like them, right?" And he said this in 1968 . We are at a stage where any species can be revived or brought back . How do you consider the ethics of what we should give back? Which ones

[00:16:00] should we not return? I'm a huge Stewart fan, you know, I love him, I love Ryan. This quote has been famous for a long time, like, you know, something from a famous movie about dinosaurs. But I don't know if I would characterize it as Stewart not getting on the plane. But from my perspective, I would say we waste a lot of time. From a synthetic biology perspective , it is not yet possible to bring back or create everything. So we try to be very balanced, and people ask us, is there a checklist? How exactly do you approach species selection? Because there are species that we have publicly announced, and there are those that we have not yet announced. Cut it, cut it. I'm not talking about what exactly you do. Do you mean philosophically? Generally philosophical, right? If we could bring back any species, who would be in charge? What arguments do you use to say: "this is what you should return, and that is not"? Just from a social ethics perspective, how should we think about this ? Speciesism, right? Yes. Yes, that's a great

[00:17:00] question, and since we're the guinea pigs in this, so to speak, we're wondering what their contribution to the environment was? What was their contribution to the food chain? Why did they become extinct? What do indigenous peoples think about this, right ? For example, for some of the species that we work with, there's a deep spiritual connection with indigenous peoples, you know? So you're not going to just come in and quote Stuart Brand to them, are you? Therefore, I think it is very important to take all these factors into account and weigh them . We may make mistakes, and I think society will make mistakes too, but we will keep trying. But we're also looking at whether there's an educational benefit to this, right? For example, when we were working with dire wolves, after consulting with indigenous peoples, the Red Wolf Coalition, and other groups, we also considered the pop culture component. We thought : is there a way to get all these people who are into science fiction , Game of Thrones, or Magic the Gathering involved in the cause of wolf conservation? Can

[00:18:00] we teach them genomic engineering, since we have brought back what they thought was only a mythical creature from their fantasy universe? So we try to evaluate all these things differently. There is no perfect internal algorithm for assessing the situation, but I am sure that as technology spreads and how governments use it, they will have to evaluate everything on an individual basis. So, Ben, no one should...... that we all make mistakes someday. Well, because there are so many different factors involved, right? Let's say I'm from some indigenous tribe and I worship the Tasmanian devil. This does not mean that it should or should not be brought back to life. There are a whole bunch of other factors. We will have to think about this very carefully, because the power you bring to this world is something we have never seen in human history. Yes, yes. I mean, it's quite... You're like a walking singularity. Yeah, we...I think society doesn't quite understand...I mean, they mentioned Clark too. And I

[00:19:01] think they all came up like Ian Malcolm when they talked about it. But I really think he was right. I truly believe that the technologies that are emerging at the intersection of synthetic biology, computing, AI, and eventually quantum technologies will be more powerful than any weapon created so far. Ben, we're talking about solving everything. We say that mathematics has already been mastered, physics, chemistry, and biology are next. Is there a tipping point, a singularity in biology, where suddenly there's a complete knowledge base of all the learned DNA, all the learned phenotypes, and you can literally design an organism ? CAD for biology. You can design, you can, you can, you can create a request. Yes. Give me an animal that does this. Yes, I think DNA synthesis is not quite at that level yet . But, but predict this for me. Yes, yes, yes. So I believe this world will come in less than 10 years. Good. So, you will be able to create your own

[00:20:00] Pikachu or other animal within 10 years. Whoever has the technology should do it for you, but I think the technology to engineer key phenotypes of basic organisms of different clades, to be able to synthesize them or multiplex edit them, and then grow them outside the womb—that's a decade away. Okay, outside the uterus. Over the course of a decade . Tell me, tell me about yours, yours, what is it? Over the course of a decade . Like this. Can you imagine what this is like? Well, let me give you the current curve, and that doesn't mean it will continue the same way, of course . But we celebrated success when we made 20 edits. And I think 99% of biopharmaceutical companies and research institutions would do that today. We consistently perform over 300 edits with an efficiency of over 90% . Isn't that right? That was, that was a year ago. Oho. Yes? We're testing thousands right now... So this is an exponential growth curve, right?

[00:21:00] That doesn't mean it will continue to be like this, right ? We test thousands of options. We are working, we are now at the level of a thousand edits. That doesn't mean it will work. I've been working on this. Tripling basic edits every year? We don't know if it will work, but we have success. Initial results may be low efficiency, but it works. But there will come a time when replacing large fragments using DNA synthesis will simply be better. Unfortunately, people in this category don't really have a business incentive to synthesize DNA after a certain scale. So we just started doing it on our own. Math, tripling every year, now you have 1000 base edits. What does this give? Well, we, we...no, no, no, no. We're consistently above 300. We're testing 1,000. I don't want to ... I think it's very likely that we'll make it work. When will the moment come when everything gets really crazy and you can do whatever you want? I think synthesis... I think synthesis will replace multiplex editing sooner. Yes. So, essentially, a machine that generates all that gigabit code that you

[00:22:00] need. I think... I think this has a higher probability of success and is faster. Okay. So, Ben, you've created an artificial egg, not one that's eaten, but one that gives life to birds. How far are we from a woman in the room, young or old, giving birth to a child in an artificial womb? A person . From a technological perspective, I think the answer will be very different than from a societal, acceptance, ethics, and rights perspective. Okay, okay. So let's talk about an artificial uterus for mammals. Yes, I think within 24 months we at Colossal will have fully extrauterine mammalian birth. From conception to birth. Without a surrogate mother. Wow. Oho. For mammals. This is simply incredible. Yes. This is wild. I hope even sooner, but I think 24 months is very likely. Oho. I, I need to ask a quick question. Yes. Just raise your

[00:23:00] hands in the hall, is your brain exploding? Well, okay, just checking. Okay, I understand. Yes. I'm sure I'm not the only one . So what's the " iPhone" moment in Colossal? Is this a mammoth coming onto the stage ? Is there anything else that... I think that... I think it's ... I think it's a "wow" moment. Yeah, I think...Well, I mean, we've had a little bit of that already, but I think the next big tipping points will be when we show the world the next extinct species, right? I think it's always in the " zero to one" mindset. When will it be? Just kidding. Soon. So, I think demonstrating another extinct species through precise gene editing is number one. Number two, I think, is the artificial development and gestation of mammals, that's number two. And also, you know, we're working on something we have n't talked about yet, and I think it's just as

[00:24:00] exciting as direwolves. So we have more to surprise and delight you, if that surprised and delighted you. If this surprised and scared you, well, we have something for you too. Therefore. Can I ask you about the dataset? You know, protein folding came as a complete surprise to everyone . It seems like it was decided overnight. This has become a real goldmine of change for all of biotechnology, and my daughter uses it every day. Yes, we use, I mean, yes. Large-scale. So the equivalent of the genotype-to-phenotype mapping problem is when you say, " Okay, this sequence created...Oh, this is Alex Wiesner-Gross." Okay, this sequence...Oh, it created a llama. Perfectly. This...Oh, this is a dodo bird. OK". I like this final jump. A little... We like this jump, by the way. Uh, so is there ever going to be a point where, having a data set that you're accumulating, you can do an interpolation and say, "Okay, now I don't need to create this. "I know exactly what will happen." So, from a product perspective, we're currently doing it at the species level or trying to extrapolate to larger groups of animals.

[00:25:00] For example, size is an important factor, right? If you take the model species that we work with for the Tasmanian tiger , or the thylacine, is the fat-tailed dunnart, and it's 1,500 times bigger: from a marsupial mouse to a marsupial wolf, right ? Yes. So understanding that and extrapolating that, and what exactly regulates that—not just genes, but how and when it's regulated in development—how do you translate that to...well, can you make an orca the size of your goldfish? Probably not, but you will probably be able to scale within certain feature levels. Yes. If you look at certain species, like dogs, and also certain groups of birds, they have enormous ranges of scale, right? Which are well over 1500. So we have a 1500x magnification. We're looking at this not from a trait engineering perspective, but from a deextinction purist perspective, to look at the genes that

[00:26:00] caused X, Y, and Z. But separately, we're trying to extrapolate this to the clade level so that we can say, " Okay, how can we affect size, even within 20 to 50 percent of the variance in other species." So I think that coat color, size, skin folds , scales, feathers—all of these would be completely measurable. Yes. And this can be induced very quickly. Interesting. So you can... okay. Know everything, but we have an entire AI team that works just on patterns and stripes. In fact, this is a very difficult task. Patterns, tusks, stripes, muzzle length, hair, hair, yes, hair. Aha. By the way, it was quite significant. When Chris and I took our boys to Dallas, it was a real surprise for us to see woolly mice there. Aha. How many

[00:27:00] gene edits did it take to do this? The first generation we showed to the public was eight. Yes. Surprisingly. In one delivery, eight pairs of bases. Eight edits. Eight edits in total. Eight basic edits in one delivery. Got it. Surprisingly. Let's move on to audience questions. This is a question from Bruno. Maybe someday we'll have another version that ... Okay. I can't wait. Yes. This is interesting. We'll invite you back to Moonshots to talk about this. Okay, Bruno asks : does the "moon project" have to solve one of the world's problems? We had Shatner the other day , and I thought, " Damn it, if we have..." Oh, yeah. Yes. They appear everywhere. Yes, probably. Tribbles. That's right. We need tribbles. You could bring back the trebles. I think "return" is the wrong word, it's better to say that we could create them in the future. Create in the future. OK. So, Bruno asks you the following. Does a "moon project" have to solve one huge problem in one vertical, or can it be

[00:28:00] horizontal, solving everyday needs that span many industries? We'll ask Astro, the captain of the "moon projects," this question a little later. I think you can move across verticals, right? But remember, I have ADHD, and I think this lack of concentration gives us more wisdom in many categories, right? We view de- extinction as a systemic problem, but the same systemic modeling that helps conserve or restore species can be applied to many areas, including health care. So I think if you lay the foundation in what you're building, it can be applied to many cases. It does n't have to be so narrow that if you miss the moment, you lose the broader perspective. Here's a great question from Teresa. How do you plan to address ethical issues when creating mammals outside the womb? Mammals have emotional needs. And simply creating an animal does not relieve you of emotional responsibility for the living being. I was going to ask the same thing

[00:29:00] . For example, if it is born outside the womb, do you give it a family to live in? Yes, it all depends on upbringing. We do a lot of things , and I think most people don't know that because the media doesn't always cover everything we do. We have a foundation and we provide open access to all our technologies for nature conservation. So everyone can use our technologies for species conservation for free. We have 75 partners around the world, we are very grateful to them, and we also fund projects in this direction. Mammoths and elephants are extremely social animals, right? So we're not going to bring back one mammoth, we're bringing back their herds. We are working on 16 different lines at the same time. We have 16 different lines in operation. No, no, how many mammoths do you want to bring back? Do I want to return it? Tens of thousands. Tens of thousands of mammoths. Yes. Wow, in Los Angeles? Um, there should be Columbian mammoths or pygmy mammoths here. Augustine, Augustine, Augustine. The main question that everyone is silent about is: where to put these mammoths ? But back to that, but back to that ...Wait, wait, wait. Yes, I want to answer this because I think it's a really important and profound question. So you had, well, you know,

[00:30:01] California condors, there were all these different species that were close to extinction that people are working on and breeding. And we see that while there is a halo of positivity around this today for elephants, it also has broader implications for what Colossal is trying to do. In Botswana, we found an incredible group called Elephant Havens that works with orphaned baby elephants , right? Which are already born, not from scratch, but they were born and were abandoned for some reason, and they're working, using artificial intelligence and many other tools, to figure out how to create synthetic herds in a matriarchal elephant society, where they do n't get that right now. How do they raise these elephants and teach them to be elephants and also to work together in a herd, right? Because that's how elephants behave. And we're also funding and conducting research into everything from satellite imagery to drones to artificial intelligence, creating programs on different

[00:31:00] elephant migration routes and corridors so we can understand the social dynamics and movement hierarchies, right? And so all this technology and this data is impacting elephant conservation work today, right? Yes? So that you don't have orphaned elephants. You can return entire herds to the wild. But all of this data also tells us, how are we going to raise these animals so that if they're born outside the womb, they grow up in a social dynamic with the right hierarchy? Dave, did you want to say something? Here's a true story: Your colleague, George Church, was at a presentation we were giving at MIT, and the topic was global warming. And he said, "Well, I have, you know, something here." "I have a cure for global warming." We are going to return but yes. We're going to bring back the woolly mammoth. The woolly mammoth's native habitat is Siberia and northern Canada. Tundra. Tundra. And when they existed, there were no trees , because the woolly mammoth, yes. went around and felled all the trees. And elephants really do this in Africa,

[00:32:00] yes. Yes, yes. Forest elephants—yes, yes. They are incredible at this. So we're like, "Well, what the hell does this have to do with global warming?" He replies: well, grass grows without trees. Grass actually absorbs more carbon than trees. Yes, it is about six times more efficient, and the albedo effect from reflecting light into space increases by two to three times. Yeah, I don't think you guys have coordinated this with the Canadians to see if it's okay, but you're releasing them, releasing them in Canada. live in the tundra. So, I myself am originally from India. Seriously, where should we build Jurassic Park? Yes, no, wait, that's a really strong argument. So, in the early stages, I'm a big data fan, most of my experience is in software and a little bit in space hardware, but mostly I just want to follow where the data points, you know? So you have convinced, very intelligent scientists like George who say that if you have that density of mammoths in these places, in the tundra, they will have this effect on the permafrost—a drop in temperature of 6 to 8 degrees in the summer months, and it only melts up to a certain point.

[00:33:00] So you can extrapolate that. We actually did this exercise, and it's quite interesting. But it all comes down to comparing top-down and bottom-up approaches in terms of how they impact the environment. There are other people in the scientific community, particularly at Colossal, who believe they won't have that level of impact, right? But the good news is that overall, regardless of how they feel about how exactly 10,000+ mammoths in the Arctic will solve the problem of climate change, it doesn't really matter, because they have a net positive impact on the environment, helping to restore this ecosystem. So what I'm trying to do—and Palmer mentioned this earlier—is how do you get two scoundrels in the same room to agree? Well, that extrapolates to the tenth degree when they both have PhDs. In my experience, I don't have a PhD, and people sometimes think I'm waging war on academia, but in my experience, people with PhDs are pretty difficult to communicate with. Actually, they have a model, but I've noticed: when you sit them down

[00:34:00] together and tell them they're both right and help them figure it out, it really works. So I said that maybe George is right that this density of mammoths at this latitude and longitude would have this effect, but maybe others are right in saying that it would have a positive effect on flora and fauna, but it wouldn't fix climate change. In any case, it doesn't really matter whether this is a net benefit to modern elephants and the tundra ecosystem, which has been severely degraded. So everyone can agree that the ecosystem is in terrible shape and we need to make it better, right? And that's why , that's why I did it and, you know, I don't know. I don't want to claim that George is right, but I'll just say that I've studied the calculations and I think George is pretty smart. OK. Ben, you just started AstroMech, a company that's worth billions of dollars from the start. What does AstroMech do? We're looking at—that was kind of a lead-in to your question. Fundamental and

[00:35:00] animal models don't quite capture the entire tree of life, and they can't magically get the genome overnight and give us the answer. So we try to create what we internally call inflectional models. What are these inflectional models that you can plug into those fundamental models and say, " Okay, we've learned and we understand everything about this genome sequence , how it evolved, and more importantly, when it evolved and why it didn't evolve in related clades." And then we look at everything, like climate, what drove it, because we want to build, essentially, a predictive model to say, "Okay, where did this go and why?" Because I don't think that Astromech will have a magical, you know, entropic mythos, that $4 trillion or whatever final round that solves all the questions in everything. But I think we will have enough unique data sets to classify and understand to connect to them. So when you have global biorepositories and millions of

[00:36:00] samples to submit to Mythos, it can act as a kind of competitor, acting as a regulator, telling you where to go and what to focus on. Yes. Yes. You know, if you combine a few questions, what's the biggest problem that you think should be worked on, what's the " moon project" that no one is working on right now ? I think we will lose half of our biodiversity in the next 25 years. And everyone who loves us or hates us agrees with this. So, we need to do something about this. Governments, this cannot be solved by zoos or non-profit organizations. We need billions of dollars in federal funding from different governments working together to at least back up life. We back up everything else. We make backups, you know, we save our photos. We make backup copies of messages and emails. We back up everything. Although, in the end, I think most people make copies of messages and letters. Wait, some people use Signal. But

[00:37:00] for the most part, I really think we need to invest in infrastructure right now to preserve these species, because if we don't, then... When we lose a species , it negatively impacts ecosystems. When we lose a species, it negatively affects animal food chains. So if you love ecosystems, you should make a copy of them. If you don't like ecosystems and hate the environment, but love animals, you should make a copy of them. If you hate the environment and animals but love [ __ ] people, you should copy them. You know, there's data there that will help humanity. By the way, as a quick aside, one of my companies with a great CEO, Bob Horry, called Celularity, has a Lifebank USA project where when a child is born, we store all the cells from the placenta. Yes. And you actually have the original boot disk. You have your child's stem cells, T cells, natural killer cells, everything. It's like if your child had an extra set of organs, would you throw them away? Probably not. So why throw it away,

[00:38:00] because the placenta is a 3D printer that creates a baby. So this vision of security, of backups—it's just amazing. So this is what you're doing in Dubai now? We do it in Dubai. We just announced a partnership with the U.S. Fish and Wildlife Service as part of Secretary Zinke's directive to preserve the natural resources that make America great. So , we are now doing it here, within the country . We have two more governments that we haven't announced yet, but we will do so when they want to, as they are part of our structure. And it's also similar to what George and I are talking about with open source—it's completely open. So any nonprofit , any academic institution, any large foundation, any private individuals, anyone from Palmer's secret club of billionaires who wants to put money into this. To be in a secret boys' club. Are you in a secret boys' club? So, therefore... Anyone who, for example... No comments. I'm not in any secret jet owners groups.

[00:39:00] I have a quick question. Okay, let's wrap up with your quick question. I want to be Ben Lemm with a mind as creative and crazy as yours to imagine these things. How do I do this? Oh, how do you do that? How do I go about... How did you become... How do I adopt your mindset to use technology as creatively as you did? Well, I was very kind. I guess I'm very curious, you know? I just like learning new things. And I think in a world, especially with artificial intelligence, where everyone has the answer to any question at their fingertips, I'm pretty good at admitting what I do n't know. So I approach things with childlike enthusiasm and just say, "Hey, I don't know that, but I'm sure I can find the answer." And I also found that most people—this is a big piece of advice I would give to everyone—people will help you. I am an optimist. I believe in technology, but above all, I believe in humanity. And people will help you. So I don't think people

[00:40:00] ask for help often enough. It seems to me that everyone walks, rides the subway, and so on, staring at their phones. But if you look around for even a moment and ask for help, people will help you. And so when I don't understand something, sometimes I call people, like some of our lead advisors at Colossal, our advisors, just because I wrote them a cold email: " Hey, I don't understand this. My teams tell me this. You are a world expert in this field. Can you meet with me ? This woman or man has no reason to talk to me, right? But they will agree to the call. So, I feel like I have this general curiosity that surrounds People like you. They want to help you. No, I think they're like everyone else. It really is. I believe in humanity: if people ask for help, nine times out of ten they will receive it. Most people ask. A quick round of applause for that, right? A worldview of curiosity, a worldview of purpose, and a quick question from Katie Wood backstage, who said yesterday that Anthropic announced their labs that are capable of creating something like CRISPR. Does

[00:41:01] it stimulate your work, or perhaps kill the momentum for development? No, I think this is an incredible confirmation, right? That's a great question. Katie is wonderful. I think this is very important, right? There are so many problems in biology that need to be solved. These technologies...we have n't even opened the door, the door is barely ajar. I think what they announced yesterday— people see it and say, "Oh my God, in five years we'll be dead because this is going to be Anthropic in our lifetime." That's not true, right? This is simply not true. So I think this was a huge turning point for the industry, showing that there are AI companies that understand biology, and they know that this is going to be one of the most acceptable uses for their technology, because people want to have healthier families and live longer, right? So it was a great thing for society. Very briefly. Do you have an assessment of the probability of a disaster, or do you not even think about it? You know, I'm an optimist.

[00:42:01] Yes, I am an optimist. I don't agree that... I think we're going to have scary moments, and I think that's okay, right? Because I believe in human ingenuity to overcome these problems, right? But I think it needs to be talked about. I think the media is exaggerating this a bit right now, to put it mildly . But I think we'll actually get to that point. We'll have some scary moments, but it's like turbulence on airplanes—everyone lands anyway, right? This is normal. Okay, welcome to the Oscars of optimism. Your applause to Ben Land.