Parents Will Choose Their Children's Genes | Kian Sadeghi, Nucleus IVF+

You dropped out of college. You started Nucleus at the age of 20.

The goal of life isn't to build a successful business. The goal of life is to live a great life. There's a necessity of genetics, right? Because if you do genetic tests and you find a genetic marker, oh my God, that has huge implications for my future daughter. I don't want to pass that down.

Where do you draw the limits of generational health?

The patient has the right to use or not use the technology. It's people who do IVF, they have no idea that this technology even exists. I'm not even just talking about the polygenic stuff like cystic fibrosis, PKU, breast cancer. You have technology that could save life that is not being deployed. That has If you're picking your child, which is their DNA for life, you don't want to do that randomly.

With like the most advanced AI model, I won't be able to specify precisely my designer baby to be like 6'5, IQ 183.

I have beef with your company because you guys do gene editing. What? We don't do that! People like just literally, they might read one tweet and then they build their entire worldview on that tweet about something or someone or some company. Height, low disease risk, those are proxies for living a good life. You can be super smart, super capable, maybe even tall and handsome, but they're not happy. Like, life is so much more than just the abstraction that is DNA.

Hello and welcome to yet another episode of AccelerBio. Thanks for coming on, Kian. I've been looking forward to this one.

Yeah, I'm very excited. Thank you for having me.

Nice. So let's start by getting personal. You fit the archetype of a visionary founder. You dropped out of college and you became a Thiel Fellow. You started Nucleus at the age of 20 in your bedroom, I believe. You raised $31 million. So before we get into all that, I'm really curious to hear your take. Do you think conventional education and career paths are outdated? outdated and failing young people right now?

Wow, it's interesting you ask this question because remarkably, I feel like people ask me this question much less than they used to. And maybe that's because I'm substantially more and more with each passing year out of college. Maybe that's because in terms of the zeitgeist conversation, it's shifted a lot in the last couple of years. I feel like it was much more of a contentious point, but then at some point there was AI, and then at that point it's like, why even bother about college anymore? It's like everything Like there's going to be AGI, so who even cares? But anyways, to answer your question more directly, which I haven't thought about in some time, I think that like, I really think to do entrepreneurship well, it has to be some sort of calling to some extent. Like I do think people like entrepreneurs, like exceptional entrepreneurs, they almost have like these really, really great ingredients. And sometimes some of them are good enough to actually take the ingredients and make into something. Sometimes some of them need a little bit of mentorship to actually take the ingredients in to make something, right? Sometimes you need your kind of master chef. You need somebody to be your apprentice. Other times, you can actually maybe even just figure out yourself if you cook enough times. I'm really of two minds about this. And actually, it's weird because I oscillate between two extremes. And usually when I oscillate between two extremes, it means that I have evidence of two very compelling models. that directly oppose each other. So I actually haven't built a model yet that integrates the 2 schools of thought. Okay. On the one hand, if someone comes up to me point blank and asks, should I drop out of school? I say, no, never. It's a really, really bad idea. But then usually what ends up happening is if I ask them 3, 4, or 5 follow-up questions and I actually get enough information about them, I can make a more tasteful assessment whether they should maybe actually drop out. But also there's this meta question of like, if someone's asking me if they should drop out, they almost certainly shouldn't drop out. You shouldn't need to ask me. You should just have dropped out actually already. I didn't ask anyone, oh, should I drop out? I was like, this is scary as hell, but I'm going to throw myself off the cliff. And so, there's that component too. Obviously, it's what you want out of life. I do think something that gets underappreciated in Silicon Valley discourse is there's this great PG quote saying that the goal of life isn't to build a successful business. The goal of life It's to live a great life, right? Building a successful business is ultimately a subset of the problem of how do I live a good life? And that, I think, is actually the more interesting question, which is actually what you're asking for. You're sort of implicitly asking for how do I live a fulfilling life and can it be done outside of startups? Of course, it can be done outside startups. Running a startup, I think, even at 20, even if it's your entire life, it's your entire being, your entire soul, as it is You know, when I started Nucleus, you eventually realize that a startup, that a company is one part of a life. And I don't know if that— I'm sort of rambling here, but this is sort of like kind of my idea stream as I think through this problem.

Well, thank you for your nuanced answer. And I think you're onto something when you say, if you have to ask someone else, should I drop out, then maybe you shouldn't. Maybe you don't have the confidence to do that. So that's fair enough. And let's just move on to Nucleus. And what I understand from your story, so you guys started by offering full genome sequencing for the individual. I believe then you moved into couples genetic matching, then into embryo selection. And now most recently, you are also doing the full IVF stack. So It seems like quite a journey. And what I'm curious to understand, first of all, was this journey planned or was it kind of something that just happened?

So one way I think about this is that whenever you create something, there's almost like a founding document, right? Obviously, the most famous founding document, at least in the West, is probably the Constitution. Right? This idea that like a bunch of people came together, they wrote some sort of like fundamental guidance for the United States of America. And then as much as possible, you know, us as citizens, we try to adhere to the Constitution. My kind of— I like to say that there's like, you know, local constitutions, right? There's like local like pacts. Each startup is like a new society. It's a new world, right? It's a new country within a country, if you will. And they each have constitutions. They each have kind of their fundamental ethos of what they stand for. And for me, the Constitution of Nucleus is my notebooks, right? So actually, if you go to my bedroom, there's a safe, and in the safe there's about 18-subject notebook worth of notes on genetics, on computational biology, on statistics, on prediction statistics, a lot of statistics. And on the first page of my first notebook, I basically wrote a vision for Nucleus. And on that page, it says something like, Through Nucleus, basically the goal is to use DNA, use genetic analysis to help, one, people live longer and healthier lives, and two, let people have healthier children. Those are the 2 fundamental kind of pillars in my mind, which is the genetic stack was never actually limited to adults or embryos or children. The genetic stack fundamentally is unique in that it's what we call generational health. It's a generational diagnostic. When you take a genetic test for yourself, you're also taking it for your children. You're also taking it for your mother. You're also taking it for your brother and your sister. People don't usually think about it in that way, but very literally speaking, that's what's happening. And so when I think about Nucleus, when I think about the vision of Nucleus, I think what I got inspired to do with Nucleus, obviously my cousin, she unfortunately passed away suddenly at 14. Obviously I wanted to make sure that didn't happen to anyone else in the world. And when I think about now realizing that vision, that vision of targeting both adults and also embryos and children, making sure that people have that entire stack, that entire stack, I think at first glance, like to a naive eye, to an untrained eye, they're going to say these things seem really different, but they seem really different to the person because from the outside looking in, they've never actually understood that what Nucleus is really, really fundamentally doing is we're building software To analyze DNA. And that software to analyze DNA, an embryo DNA file is the exact same as an adult DNA file, which is the exact same as a child's DNA file. In other words, the vision of Nucleus from the very beginning was to harness the power of DNA to build this genetic stack, which by the way, when I said harness the power of DNA, I'm not sure if you know this about me, but in high school, I actually did a lot of genetic engineering work as well. And so, for me, harnessing the power of DNA never was limited to from the very beginning, it was never limited to just sequencing, analyzing DNA, but we can talk about that in a bit. It was building what I call this genetic stack from sequencing, analyzing, and eventually engineering DNA. And, I would say that Nucleus really were the genetic optimization company. And, when you think about genetic optimization and the genetic optimization industry, it's not just about optimizing your health, it's also using our genetic optimization software to give your child the best start in life. So, fundamentally, the ethos is the same. That was a bit of a longer answer, but that's my perspective.

There's already so much I want to double-click on in this answer, but maybe to begin with, when you say generational health, so okay, it's myself, then it's me and my partner, then it's my children as well. Where do you draw the limits of generational health? Is this what you described now, your full roadmap, or how much further can you This is never the

full roadmap. I mean, think about that. That's a good question. Let's break this down a little bit more, right? Let's say you're an employee of Nucleus and like how I'd explain this to you. IVF Plus, right? An end-to-end IVF experience. So basically if you're doing IVF, you can come to Nucleus and then we'll help you basically do what we think of as 2 key value props. The first is we give you complete peace of mind. We give you concierge service. We give you an end-to-end experience. You know you're in good hands with us. We have counselors that guide you through the entire IVF experience. A lot of patients feel stressed and concerned. They don't know. They feel like they're billing specialists. They feel like they're geneticists. They have to be scientists, right? They have to be sometimes physicians and self-advocates. We take care of all that for you. We match you with the best clinics in the world, right? In the world, not just in the States, in the world. And we then power your IVF experience with the most advanced technology. So it's a great example where peace of mind and kind of the utmost quality of care meets the most advanced genetic diagnostics. enabling you guys to ensure that you don't pass down any hereditary diseases to your child. And also you can optimize using our genetic optimization software. You can pick off of anything from height all the way to diseases. Okay? You can pick your embryo. And so now IVF+, if you actually think about it, is the culmination of Nucleus Health, right? Because if you do genetic tests and you find a genetic marker and it actually impacts your health, it will also impact your future child's health. And we've seen this happen, right? Where someone gets a result that, oh my God, I'm actually at risk for I don't know, breast cancer. And oh my God, that has huge implications for my future daughter. I don't wanna pass that down, but also now that saved my life. So IVF+ combines Nucleus Health, it combines Nucleus Family, obviously, right? Because you have to do genetic matching. You have to see if you're a genetic match to see what genetic diseases that you're at risk of passing down to your child. And it combines Nucleus Embryo. It's pretty remarkable if you think about it, right? Nucleus Embryo is obviously genetic optimization software. Once you have your embryos in IVF, you can actually compare and contrast your embryos. So now if you actually think about IVF+, It is the culmination for me personally of 5 years worth of effort in a beautiful use case where for both the patients and their future children, the necessity, there's a necessity of genetics. There's an absolute necessity. Literally speaking, what is an embryo? It's DNA. It's a string of A, T, C, and Gs. It's beautiful. And so the roadmap is always thinking about how do we build technology that actually serves patients first. One of my biggest gripes that a lot of people just don't have a lot of good intuition for because they never actually had to productize frontier technology is that building the technology is actually the easy part to some extent. I say that because once you can have the best models in the world, the best technology in the world, someone has to use it. Someone has to engage in it. Someone has to interface with it. Physicians have to use it. Physicians have to understand it. Clinicians, not just physicians, but more broadly, the IVF industry, right? Embryologists, right? The coordinators at IVF clinics, they have to embrace the technology. And so suddenly you realize that there's always this thing where it's like, YC, their catchphrase isn't make the best technology, it's make something people want, which is fundamentally very consumer-centric, as it should be, which is customer-centric.

So thank you for walking us through IVF+ and It sounds like a pretty interesting product to me personally. And just to be a bit more specific regarding what you said about the roadmap, so can you share any specifics about where you go from here?

Yeah, I think one thing that we're really focused on right now is just building out the IVF+ experience, right? Obviously, there's a lot of work to be done. One way of thinking about IVF is actually that Genetics is a key component of it. As I mentioned, embryo is DNA, but if you think about the IVF process, right, there's a couple, they want to have a child. Okay, first step, they have to match with the clinic. Okay, so right now we have formally announced one clinic. It's Beverly Hills Fertility. Fantastic clinic. The physician there, Dr. Ellen Goldstein, is one of the best fertility doctors in the world. She's amazing and exceptionally good bedside manner, which is really important to us as we look to partner with clinics. But we'll be announcing a lot more clinics over the coming months. And these clinics should not just be nationwide, but also internationally as well. And so we're very excited about that as we build out this kind of what we call our IVF Plus clinic network. Right. But then that's just the first step, which is the clinic matching. Next is, okay, well, you have to run a bunch of different diagnostic tests when you do IVF, right? There's things on like your sperm quality. There's different kinds of fertility tests you have to take and obviously the genetics as well. blood tests, et cetera, et cetera, et cetera. And so there's a question of how do you make that process easier, right? How do you make the process of running actually other diagnostic tests easier? And can you actually bring that diagnostic test data together with the genetics data as well? And then eventually you have to go through an IVF cycle. And so there's a world where you can imagine there's an app, there's an IVF+ app, and you think, well, each step of this process, obviously I want to keep track of where I am. Right? And so that really doesn't exist actually. There's not really like a brand in IVF. There's not really like the go-to like IVF application, right? These are clinics that aren't software companies. And so now you can start seeing how as Nucleus has gotten more and more into this industry, obviously genetics is fantastic. It's what we do and we provide the best genetic experience in the world. But you can see a much broader use case is forming around the entire IVF experience and serving patients who are doing this entire IVF experience. So that's kind of, I think, a good hint in the direction that we're personally very interested in.

By the way, if you're enjoying this, please hit subscribe. We drop a new episode weekly exploring how AI, biotech, and human enhancement are shaping the future. And if a friend would love this too, send it their way. It really helps us keep going. Now back to the episode. Judging by what you're saying, it sounds like right now it's a pretty messy and fragmented and maybe confusing market as well for people going through IVF.

Stressful. It's fragmented, it's decentralized. There's many different lab tests, many diagnostic tests the patients have to bear the burden. We solve all that.

Fair enough. And so next, I would love to double-click on the actual technology that you use. You have previously said that 23andMe was driving a horse buggy using a 20-year-old technology. I hope I'm paraphrasing it correctly, whilst you're building a Bugatti. So can you explain to me how different these 2 technologies actually really are?

Yeah, yeah, let's do that. So In my head right now, I'm trying to think about whether we should give the whole history of genetics or whether I should start in the early 21st century. So one way of thinking about this is that, think of your DNA as like a 1,000-page book. Okay? There's 1,000 pages in this book. Historically, to read all 1,000 pages in your book of DNA, it was a billion dollars. A billion. Way too expensive. Obviously, you cannot possibly productize or sell a consumer product if the cost of reading all 1,000 pages is going to be $1 billion. And so, in 2006, the price had come down a bit, but it was still really expensive. It was about $10 million to read your 1,000-page DNA book. And I mentioned 2006 specifically because there was a company called 23andMe. 23andMe was a company that started in 2006. They basically focused on Providing some sort of direct-to-consumer genetics experience. Everyone knows 23andMe. I don't think you need me to really elaborate on what they do. And so, the thing was, imagine, and I really want people in the audience, whenever you're listening to a podcast, you should always intuitively imagine yourself in the CEO or founder's shoes with the restraints and restrictions and context that they have. That's a way of taking a podcast and actually being able to more kind of properly assimilated into your understanding, your worldview, right? If you were the CEO of 23andMe, you would look at the cost of sequencing a genome. You would say it's $10 million per sample. That's way too expensive. What would you do? You'd probably think to yourself, is there a way of getting, instead of a 1,000-page read of someone's DNA, maybe reading one page, maybe reading the most important sentences on each page, right? That's probably your intuition. And that'd be a good intuition because instead of $10 million, it would cost, let's say, $1,000, right? So, by reading a smaller set of actually your DNA, suddenly you're able to actually productize a consumer product. And that led to 2 things. One is it led to them getting to market, and they could sell the test. They could tell you about your ancestry, and they could tell you very limited health information. But 2, in the 20 years since 2006, what happened? Well, the cost of reading all 1,000 pages went from $10 million to a couple hundred dollars. And so now it doesn't really make sense to read this small one sentence in each page of your book, right? Because now it makes sense just to read all of it. I mean, it's only a couple hundred bucks. Obviously, you can imagine if you read 0.1% of your DNA versus nearly 100% of your DNA, obviously reading 100% of your DNA is going to lead to dramatically better genetic analyses. And so in terms of a pure technology perspective, They were under intense technological restraints because of the cost, but now the cost is actually cheap enough that then you can actually bring a whole genome test, what's called, and bring this entire genetic experience to the consumer. And let me make one more point here, which is important to say, which is any company, more inductively, let's always think inductively here, more inductively, any company is based off of what I call an intellectual foundation. So, if you think of them as societies, or maybe even more strictly to take this analogy home, if you think of them as castles, so each company is a castle, the very foundation of their castle is some sort of technological assumption, and the entirety of their castle is built on that technological assumption. So the whole thing about like the innovator's dilemma, the whole thing about like disruption technology, what you're actually doing is you're taking that bottom layer, you're taking that foundational layer of the castle, and you're ripping it out from underneath. Right. And so even though it looks so big and strong and kind of almost unstoppable, what you're really doing is you're taking the foundation of an entire set of castles that were built and you're pulling it out of the rug. That's what basically Nucleus is doing. But that's what's happening actually more generally. If you think about the transformer-based model, for example, the entire ethos of LLMs is based off of this research paper. And this research paper has now enabled massive industry strides. If someone were to go to the most upstream assumption, and they would pull out that, that is no longer the actual technical frontier, it would allow for a new crop of startups to come out. Then the best founders, they combine those technological breakthroughs with their fundamental constitution, with their vision of how society to be, and that creates generational companies. That's kind of the way that I think about it, both in the sense of like 23andMe and Nucleus, but that's how I also think about it more inductively if you want to get a more of a broad scope of like, what are startups fundamentally about?

Okay. So TLDR, you are leveraging this new foundational technology, which is whole genome sequencing, and that's what makes the whole difference. But what I'm really interested to hear about is, I guess, how accurate is this whole genome sequencing technology today? And I will give you a personal anecdote. if you don't mind. So, I've used not your company, but I've used CircleDNA, which also does whole genome sequencing. Right. I've got this whole report about different lifestyle predictions, food sensitivities, and whatnot. And, it always gives you a degree of confidence, which sometimes is low, sometimes is moderate, sometimes it's high. It also told me with high confidence that my eyes are brown. But well, they're blue in case you haven't noticed. So, that specific thing didn't seem that useful to me. Maybe yours is different, but is accuracy still an issue?

Yeah. Okay. First of all, a couple of things. So, CircleDNA specifically is actually, I'm pretty sure they don't use whole genome. I think they use NGS. I think they might use a whole exome technology. I don't think they actually use whole genome. But anyways, In typical DNA's case, we think DNA is—

It might be on a whole exome, yes.

Yeah, it might be. Yeah. And, just so people understand what I'm implicitly saying here, so there's different ways of reading DNA, right? There's the kind of 23andMe way, which is basically taking a sentence from each of your 1,000-page book. Then, there's whole exome, which is almost like if you may take specific chapters and read it. And then, there's whole genome, which is just everything. You don't stop at chapters or a sentence, you just read absolutely everything. Okay, I think one thing to remember is that the problem with historical genetics companies like the one you mentioned is that they will actually look at single SNPs. So, let me actually take a step back. One, to properly answer this question, we have to understand some genetic theory and then we can get to application. So, one way of thinking about genetics is the kind of classic way, which is that if there's a mutation. Mutation has fallen out of favor by the geneticists. People don't like to use that word anymore. But for the purpose of a broader audience, mutation, you can think of it as, you watch The Simpsons, someone falls into a radioactive pool and they get a mutation. Then based off that mutation, they become jacked or they become Spider-Man or something. Okay. In other words, a single genetic marker causes a disease or causes a phenotype, which can also be a trait. That is like the classic model of genetics. It's the model you learned about in high school when you were doing the peas and you think, okay, this someone's AA and this person's little a, little a, and you did the Punnett square. I think everyone's done that in school. That's like kind of classic genetics. And that is really focused on rare genetic conditions, things like cystic fibrosis, things like breast cancer, things like hereditary breast cancer, things like PKU, for example. In other words, for a large portion of diseases specifically that are rare and hereditary, they're driven by a single bad genetic marker. In your analysis specifically, I'm not even sure if that company gives out that kind of Mendelian variants in a clinical way, because I think they might actually be a DTC company. They might not have even given you this, which is really bad. You should just do Nucleus. I can set you up. But this is a kind of classical analysis, which is like, for me, for example, I'm a cystic fibrosis carrier, right? If I had 2 of those markers in my CFTR gene, I would have cystic fibrosis. So another way to think about it is like the model is very simple. Bad genetic marker, disease. Okay? That's classic genetics. Classic genetics has been practiced for at this point decades, even before the rise of a whole genome test. You could actually think about this analogy. So, when you talk about eye color, for example, which you mentioned, eye color does not actually follow that bad marker, disease. It doesn't. Obviously, from a prediction standpoint, bad marker, disease is much simpler. But it turns out many, many, many, many of the phenotypes that we're actually privy to, things like eye color, things like height, things like IQ, things like breast cancer, not the hereditary kind, breast cancer or schizophrenia or Parkinson's, or Alzheimer's or coronary artery disease, diabetes, for example. Many of these traits and diseases, they're not driven by bad marker disease. It turns out they're actually driven by millions of genetic markers. In other words, there's something called polygenic, many genes, instead of monogenic, one gene. And so when you think about polygenic phenotypes, now you need more complex modeling to basically actually elucidate and actually predict what someone is or isn't going to have. And one intuition for thinking about it is it becomes a lot more probabilistic. It becomes a lot less like, oh, you have this disease or you don't have this disease. It becomes much more of a probability. And so when you actually combine these millions of markers together at a population level, what you're doing is you're stratifying people in different risk buckets. Okay. And so I think sometimes when people ask, Like, for example, kind of your questions like, oh, my eye color or something, right? Think about something like height, which is a very simple thing that everyone's familiar with, right? You could probably imagine that something like height, for example, it's easily measurable and you can predict it. You should be able to predict it pretty well, right? Because it's very genetic. And that's actually true. You can. But if you think about height, for example, post-Industrial Revolution, the Persians, my people were Persian. We grew 8 inches in height post-industrial revolution. That wasn't because of DNA, obviously. That was because of nurture. So now, you're actually really asking the question, which is like, what is the extent of genetic prediction? Is it nature? How much is nature? How much is nurture? And in other words, my point here now is that people sometimes create this false dichotomy. They say, oh, it's nurture or nature, but it's actually nurture and nature. It's always nurture and nature. And depending on the phenotype, it's more nature, Or it's more nurture. In height, for example, scientists estimate it's about 80% genetic, which means that depending on your environment, even if I looked at your DNA and even if I could build the best possible genetic model in the world, you would still be wrong. There'd still be stochasticity because obviously you're not modeling the entirety of someone's nurture component. So what am I saying here? I think there's always this question of like, What is the endgame of genetics? It's very important to say that the final culminated model, once you get all the data in the world and you can build the best genetic models in the world, once that actually happens, you will still be wrong. But that's actually not a bug of the system. That's a feature of the fact that people have liberty, people have choice, there's a nurture component. And also, that kind of uncertainty and that error bar should be communicated in the prediction. So that's a kind of broader way of thinking about kind of the question, which is like technically speaking, if you had a different disease that was not genetically driven, you'd expect the error bars to be much larger, right? And so sometimes I think people want a yes/no with these, but they're actually much more probabilistic, especially for the many common phenotypes that you might have heard of.

Okay. So I understand you're saying traits are often caused by multiple genes. That's what polygenic means. I understand you're saying it's probabilistic. Nurture and your previous lifestyle also plays a role. In that case, it makes me wonder, does this actually make it much easier to predict things at the embryo stage? Because there it's basically 100% just potentiality. So is it at the embryo stage where it can have the highest accuracy predicting what's the potential of that individual?

Of course it's useful. I mean, of course it's useful, right? If I can determine if someone's going to be at dramatically higher risk for breast cancer or schizophrenia or Parkinson's versus an embryo that's 10, 15, 20 percentage points less, I mean, of course that's Very useful. It's almost like saying, I can't perfectly predict the weather tomorrow. So is it useful to do it? Of course it's useful. Also, there's different kinds of predictions. If I say it's going to snow tomorrow and it snows, that's great. If I say it's going to snow 15 inches tomorrow and it snows 8 inches, that's a pretty good prediction. Now, if I was like, oh, it's going to rain tomorrow or it's going to be sunny tomorrow, that's off. That's really off. People actually have good intuitions of good predictions, and they understand intuitively that a prediction cannot be perfect, i.e., the weather, but it can still be very, very useful. So, for example, if you can predict that someone's going to basically in their lifetime be at a dramatically higher risk for Alzheimer's, but then another embryo is not going to be, obviously, that's beneficial. You don't want your baby, God forbid, to have Alzheimer's. Or even in height, for example, if you can pick across several inches of height, I mean, I'm 5'6. I would've loved a couple more inches genetically speaking because you can't get it any other way. If my mom had just picked a taller embryo, that'd be great. Another great example of this is I'm in the 99th percentile for my schizophrenia genetic risk. There's probably some positive attributes to that, maybe creativity, and founders I know tend to be very high genetic risk for schizophrenia. They're very creative people. But also, there's a world where if I had a family history of schizophrenia, Obviously, I'm going to try to reduce my genetic risk as much as possible given my family, if I have one. I do not, but if I had a family history. And so obviously, if you're picking your child, which is their DNA for life, you don't want to do that randomly. That'd be crazy, I think. I mean, I don't say it's maybe crazy, but I would say that parents have the right to choose. They have the right to know this information and they have the right to make a selection. And I actually think the weather analogy here is really good. Because people shouldn't expect— and in fact, if you think about the weather, it's actually really good because sometimes if there's a snowstorm, people watch the weather and it's like, it's going to snow tomorrow. You prepare, right? You get ready for the snowstorm. Let's say it ends up being that there's 11 inches instead of the 15 inches, right? Are you upset with the weatherman? No. You're upset with the weatherman if they say it's going to be shiny, sunny, and 80 degrees, and then it ends up being like a hurricane, right? That's horrible. Yeah. Right. And so I think the idea of prediction is actually already embedded in society and people make predictions all the time. It's like the old quote, which is like, all models are wrong, but some are useful. It's like that old quote, which I think is very true. Anyways, that's kind of how I think I see it.

I hear what you're saying. And yes, probabilistic can still be useful. I'm wondering though, on the point of maybe us not having completely mapped what every gene pair or every gene combination does yet, is that a big blocker as well? And I guess subsidiary question, how much is AI going to help us solve that in the next couple of years?

AI will help a lot. The models today are already good. They're getting better very quickly, but AI is going to definitely help a lot. I think just to give an intuition here, almost one way of thinking about this is if you could pick between, let's say, an embryo that has a 40% risk of something versus a 60% or 70% risk of something, that's a really big difference. You're talking about basically lives for each of those percentage points being saved in aggregate. Or obviously, if it's a Mendelian variant, it's even more dramatic, right? You're talking about an embryo going from, let's say, 80% risk to baseline 15% risk. I mean, that's pretty unbelievable that you can do that just from the same set of parents. I think that one thing that's always really important to say is that the models will get better and better and better and better and better and better. They're going to be an amazing tool. Parents are going to use this technology to have healthier children, to give their children the best start in life. Usually people have a family history of something. Maybe it's schizophrenia, maybe it's Parkinson's, maybe it's Alzheimer's, and the parents want to do everything they can to basically mitigate their future child's chance. And maybe even a parent wants to optimize for their child being slightly taller or the like. And even when we apply AI, and this is kind of my key point because I think there tends to be this narrative, there's 2 extreme narratives about genetics, okay? One narrative is that There's something called what I call genetic determinism. It's the Gattaca model. Oh, you have this DNA, you're in the underclass. It's crazy, right? No, obviously not. Then there's the other end, which is like, oh, genetics is like horoscope. Kind of what you were initially asking, like, oh, it's horoscope. You couldn't even tell me my eye color, right? There's these 2 really extreme narratives of determinism and horoscope. It's like hocus pocus. It's actually in the middle, which is genetics is useful. It impacts. Yeah. It's notable. Everyone should know it. It could save your life. It could definitely help your future child 100%, but it's never going to be everything. But of course not, it's going to be everything. I mean, that'd be crazy, right? I mean, if you live life, you know that there's so much more. There's such a strong nurture component, environmental component as well to life. And people want to always— they want to say one or the other. It is a complex function of both. And so, I just think it's really important to say to the audience and to level set with people always to say that genetic prediction is going to get better and better and better, and it's going to help you have healthier children, give your child the best start in life. And it's also the case that even in its best possible state, it won't be able to perfectly predict something. I mean, it just can't. And again, that's not a bug. That's a feature of reality. That's a feature of nature.

So even with the most advanced AI model, I won't be able to specify precisely my designer baby to be like 6'5, IQ 183, and I don't know, whatever features I specifically want, I guess.

Well, there's always going to be stochasticity. So you could be disposed to be whatever, 6'4, or even have a high IQ eventually. But if you don't properly take care of your child, if you don't nurture your child, if you don't do all the things, or by the way, your child be really smart, really tall, but then they could be an emotional disaster because you abuse them or something. And then that could actually have some complicated impact on— because really what parents want, parents don't want their kid to be tall, they don't want their kid to be smart. What they really want is their kid to live a fulfilling, happy life. Those are proxies. It's kind of actually interesting because it kind of goes back to what we were talking about earlier, which is when you start a startup, it's like Do you really want your startup to succeed? Of course you do, right? But that's actually a proxy for— that's part of a subset of the problem of living a good life. And this is actually such an important point. Oh, this is such an important point. IQ, height, low disease risk, those are proxies for living a good life. I mean, I think if you want to take a good example of really smart people that are a little bit miserable, some of them are unfortunately in Silicon Valley. They can be super smart or actually in the Ivy Leagues and stuff. Yeah. They can be super smart, super capable, maybe even tall and handsome, but they're not happy, right? Happiness is so important. The fulfillment is so important, right? These are just proxies. And so, I think sometimes genetics is a good tool. It's a useful tool. It's going to help a lot of parents and maybe eventually help parents to do what you described. But ultimately, I think what people are going to realize is you're just going to have life. Life is so much more than just the abstraction that is DNA. Yeah.

So I actually don't disagree with anything you just said, but just to be the devil's advocate for a second.

Sure.

What it actually makes me think about, when parents have the power to select their embryos and optimize for certain things, is it going to lead towards convergence, at least on certain traits? And if so, does that reduce biodiversity? And over the long run, I don't know, maybe over a couple of generations, can that become an issue?

But embryonic selection will never do that. Let me tell you why. Ultimately, what am I doing is I'm combining mom and dad's DNA, and then based off of mom and dad's DNA, right? Because the embryo's DNA doesn't just spawn, it's based off mom and dad's DNA. I'm making a relative optimization. Let me give you a more intuitive way of saying this. I'm 5'6. I'm a little man. If I went right now and let's say I said to myself, oh, I want to have a tall kid. Okay. I go, I find a girl who's 5'4. I marry the girl. I do nucleus embryo. What do you think my tallest embryo is going to be predicted to be? My top embryo.

Probably somewhere near the middle, I'm guessing.

Right. Yeah, exactly right. My embryos, I'm 5'6, she's 5'4. Maybe they're going to be 5'6 if they're lucky, maybe, right? If you actually think about it, what ends up practically happening is that choice of who I married obviously is going to actually set the range of possible outcomes, right? Who you marry is actually, if you really are trying to optimize for something, you should optimize for that, which already people optimize for, by the way, right? That's why dating apps exist. People are so hyper-optimizing who they marry because these are all proxies for the health, the wellbeing, the intelligence, everything, right? And so, in that sense, no, there's not going to be some sort of everyone like, okay, even if I picked the embryo that was 1 inch taller, 2 inches taller, 3 inches taller, maybe it was 5'7, 5'8 if we're really lucky, right? It's normal distribution. Okay. So what? I mean, already that's been happening for generations where people are picking based off of what's important to them. What's important to people is actually all very different. Let me give you an example of this. This is actually really interesting. At the company, one time we did an internal activity where we took like 8 embryos, okay? And then I told everyone at the company to pick the embryo that they would want to implant, okay? There was about 16 people. The most number of votes that any given embryo got was 2. In other words, basically speaking, everyone picked a different embryo. And so, I thought that was really interesting because I was like, what? I was like, really? People pick this embryo, they pick that embryo. But when you ask people, they all have different value judgments anyways. And it's always based off the biodiversity already of the combination of mom and dad. And so, there's this narrative sometimes like, oh, everyone's going to do this. It's just not practically how it's going to happen. People have already been selecting for traits in their partner forever. If you actually think about embryos, they're possible siblings. Yeah. She's basically picking amongst possible siblings. Anyways, yeah, that's kind of my perspective.

I think this mental model makes a lot of sense. I guess one of the most obvious things that are not dependent on your partner and you could intentionally optimize for is gender. So in China, when they had one-child policy, suddenly they ended up with like 60% boys and 40% girls, and that was a problem. So, I don't know, maybe Western society has different preference. I don't know, that just got me thinking a lot.

Well, sex is actually a good question because there's a couple very interesting points on this. One is that you've actually been able to select for sex in IVF for decades, for decades. And what you found is actually people are generally, I think there's been studies, I was actually just talking to our new chief clinical officer about this, There have been studies that show that it's about 50/50 what people pick, which is quite interesting. I mean, there would be some interesting decision matrix probably that everyone's doing that probably converges on some 50/50 math, which I don't know, I just feel like that's intuitively, that's what I would guess. If you actually gave everyone a choice to be like, which sex do you want? There'd be some natural balancing locally in a family. I don't think generally, obviously there's going to be exceptions, but generally I don't think someone's like, I want 5 boys. Probably that's not true on average, right? Probably people want like great mix. I mean, I would guess, right? Probably, right? Or 3 boys, right? Right. I, I like, I, I don't know, like if I take a guess, like it probably, if you like did it, it's like some central limit theorem thing where you could like think of every person and ends up becoming like basically probably converges on like about 50%. So what I'm saying is that's actually interesting, but actually think about it, that's actually a good point too, because in China it's not like everybody in China is doing IVF, right? Obviously, still procreation naturally, conceiving naturally is the predominant way that you conceive, but there was still a sex imbalance. All right. And so, this is another point which I think is really important to make is you don't need genetic science. You don't need genetic technology to do some of the things that people are describing. Genetics will ultimately always reflect the creator. Technology will ultimately reflect the values. of the population. It will reflect the values of the consumer. It will affect the values of the government, right? It's never actually about the technology. Technology can be from the dawn of time, right? Fire could be used to burn down a village. Fire could also be used to make food, right? That's the nature of reality. And so, I think that there's this idea, it's like, oh, there's going to be IVF and then everyone's going to pick for sex. It's like, You could pick for sex for decades. And in fact, you didn't even need IVF to actually have societies try to control people's procreation. And so, genetics, as in any technology, there needs to be responsible use, there needs to be thoughtful use, and most importantly, in my book, it needs to be all about liberty. It needs to be about informed choice. The patient has the right to use or not use the technology. But in doing so and making that choice, it should be an informed choice. What's happening today is people who do IVF, they have no idea that this technology even exists. Can you believe that? No idea. Doctors don't tell them about it. Doctors aren't even educated on it per se. And so you have now technology that could save life, right? I'm not even just talking about the polygenic stuff like cystic fibrosis, PKU, breast cancer. You have technology that could save life that is not being deployed. Nationally. Why? It's not a technological reason. It's not a distribution reason. It's arbitrary. It's just that the physicians don't know about it. The consumers don't know about it. That's crazy to me. That has to be solved. This is life-saving technology. The difference between someone doing a genetic test in IVF and not doing a genetic test is a baby's life. Think about that. And it's a swab. I mean, this is a cheek swab. This DNA kit is a cheek swab. You open it up, you go, uh.

It's not that hard to do.

No, it's not even a blood test. I mean, people, people, you know, inject themselves with stuff. They'll put in needles in their arm. They'll take all sorts of stuff out of their body. This is a swab and it costs what, $400, $500? Like, come on. Like, this is crazy to me. We must demand as consumers, we must demand access to the technology. And we must be educated about it. And that's one of my fundamental missions at Nucleus.

But by the way, I do agree with you, and I'm personally very excited by all of these technologies, and that's why I'm doing this podcast. And I think most of the people in the audience probably also are. However, the way I've selected some of these questions is I've just spoken to like 5 random people I know who are not in tech. And almost for all of them, their immediate reaction was quite negative, was like, ooh, bad biodiversity, oh, bye that, or oh, bad eugenics, which by the way, I think is bullshit. It's, it's completely different. We don't need to go into that. Uh, but it, it makes me think like there's definitely a lot of like cultural resistance to this probably right now for the average person. There is. So I wonder like, how do we flip that? And is it gonna take like, I don't know, one specific event for that to happen?

That's such a great point. And now let's actually think about this more generalizably because what we're really trying to ask is how do you introduce powerful new technologies to the world and show people like how beneficial it can be? That's what we're really asking here. And it's interesting because I actually think there's 2 industries right now that are the most important industries. There's AI/AGI, right? There's that narrative. And then there's genetic optimization and like basically designer babies. In other words, let's like break this down, right? There's like the broad societal narrative, which is like Terminator. Like, like it is as old as time, this idea that's like the machines are gonna take over. Like post-industrial revolution, it was like all the rage of like the machines are gonna take over. They're gonna come to life. They're gonna kill us. Like that's, that's a, that's an old narrative. Even older is, yo, if this girl and I get together, what's our baby gonna look like? That's like the oldest thing. Like, like the, like the oldest thing is like, what's the baby gonna look like? Like I'm, I'm sure from like the daughter, like what's the baby gonna be like? That's like the, The most inherent consumer demand ever. Okay? So, there's the designer baby narrative, which is a broad societal narrative, and it's basically sci-fi. Then, there's the kind of what I call AGI, which is kind of like the— or not what I call AGI, but what is sort of like the AI narrative, what I call the AI narrative, which is this AGI Terminator, the robot's going to come to life, which has also been a broad societal narrative. Okay? And then, if you take a look at those narratives, it breaks down In AI, as you know, there's the pessimists and there's the optimists, and then they are fighting. In genetics too, it's the same thing. There's the people who love the idea of genetic optimization and the people who very much don't like it. So there's the narratives. Then there's the product. ChatGPT, GPT-5, if you use it, okay, I don't think most people are going, that's AGI, that's Terminator. You go, it's cool, it's good, it's nice, it works. It's really cool and it solves really hard problems. But, you're probably not like, it's the Terminator. Okay? And so, the roots on the ground level, not like the narrative, the media narrative, not like the sci-fi, not like the cultural narrative, the actual product-level narrative, right? Because, again, there's no product, there's only really narrative. It's good. Nucleus embryo, IVF+, if you will, which is kind of like V0, the beginning of these visions, it's good. It works. You can pick across embryos, you can pick across IQ and height, and even eye color, and even breast cancer, and diabetes and Alzheimer's. It's beautiful. Patients love it. It's intuitive. They feel like they have more control of the IVF process, right? But you also intuitively, when you use it, when you see it with your eyes, when you actually experience it, you're like, all right, this isn't Gattaca. Come on, guys, right? It's the same thing with ChatGPT. You're like, come on, guys, this is not AGI, right? And then underpinning both of those, and you can see what I'm doing here, which is there's the broad societal narrative, there's the actual product level, Then there's the technological breakthrough. The technological breakthrough in the AI case is obviously this transformer-based models, which created LLMs. And then in Nucleus's case, it's obviously the rise of the genomics era, the drop in whole genome sequencing, the exponential increase in genetics data. So now you have fundamental technological breakthroughs, you have productizations of the fundamental technological breakthroughs, and then you have societal narratives. And then, the societal narratives are narratives that are both as old as time. And I have a big belief about this, which is fundamentally, the best businesses are satisfying narratives that have been as old as time. But then, these narratives have then different constituents that fight. So then, there's the product-level considerations, then there's the technological considerations, and then there's the most important thing, which is the narrative. So the question is, how do you actually show people what the technology actually is? The difference between an AI company and a genetic optimization company is that anyone can log into ChatGPT and have a good intuition and being like, look, say what you want about AGI, but this is a very useful tool that I think I'm going to derive great benefit from. Okay. And obviously there's still people that don't say that, but generally I think it's like intuitive.

I think most people that are using ChatGPT, they, yeah, they're pretty happy.

Yeah, right. They, yeah, they're happy and it's experiential. I can actually, instead of listening to some, narrative, I can go and actually use the thing. The difference, and this is something I think a lot about with genetic optimization, is, I mean, there's so many things you have to be doing to actually be in the position where you can actually pick your embryo, right? I can't just be like, oh, here's like, yeah, yeah, like you take the embryos out of your fridge and like, here, use the software. It's not like that at all. And so, that actually is one of the tactical things I'm always thinking about, which is like, how do you increase awareness in picking embryos? How do you literally show people, no, it's not what you think. It's not this crazy narrative that's existed of this Gattaca. It's actually just another tool in the toolkit that's going to help patients, right? And actually, it's already saving lives. And so that is one thing that I'm really focused on thinking about, which is like, and it's one thing that I think I'm actually proud about that we did at Nucleus. I think we've done a good job of bringing this technology out of the kind of frontier of academia and sort of this circle jerk and actually bringing it to people to saying, hey, this is what it looks like. This is how it works. These are the considerations you have to make. These models are better. These models are worse. Some models have greater errors. Some models have moderate errors. Instead of hiding and succumbing to this narrative, which is completely abstracted, you're actually focusing on helping patients every single day. Sorry, I'm getting wound up now.

Yeah. I like to hear that, how animated you become. And I think unfortunately you're right. For some reason, especially nowadays, it seems that some of the people with the loudest opinions are people that have the less exposure or the least amount of experience with the thing that they are passing their judgment on. Let me tell you a story.

Can I tell you a story? I was at a conference. Okay. I was at a conference and I was hosting like a discussion group kind of thing. And someone comes in and they sit down and they say, I have beef with your company. Okay. And I was, whoa, it's very intense things to say to me. It's my baby. I love my— I mean, we're just trying to do a good job. Okay. And I think I said, why? I mean, how— what could I do? Let's have a conversation. Let's figure this out. Right. And then they said, because you guys do gene editing. What? We don't do that. What are you even talking about? That's such a good example. I was like, what? It's like going to an AI company and be like, I don't like you because you built Terminator. What? Doesn't even make sense. What are you talking about? It's like so— and then we actually started talking and I explained to them we actually did. And they were like, oh, that's actually great. That's gonna help so many patients. I'm like, and then I'm like, okay, I actually am starting to think that like people like just literally they might read one tweet somewhere and then they build their entire worldview on that tweet about something or someone or some company. And so it's really important to me. I'm gonna continue for the rest of my life. I'm gonna push for this, which is like, just show people what it is. Don't hide, don't obfuscate. That's the thing that drives me crazy, which is sometimes in this industry people wanna hide Obfuscate. Oh, we're stealth. We're this. We're that. Why are you hiding? Come out. This is what it can do. This is what it can't do. This is who can use it. We want to help patients. We want to help you. We want to make sure that you guys are giving your child the best start in life, right? You can come ask me questions. You can go on the website. You can see the product, right? You can get it in the clinic. If you go ask your physician, we can have that set up for you. It's not some secret crazy thing. We try not to make it too expensive. Right. I mean, another thing that we didn't talk about is cost, which is really important for bringing technology to the marketplace. If you think about cost, IVF Plus, it starts at $25,000 for all in, by the way. That's for the cycle medications, that's for the genetic testing and the analysis, that's for implantation, that's for everything. That's end-to-end $25,000. Okay. There are genetics companies that charge $25,000 just for the testing of embryos. Think about that versus we sell $25,000 for the entire experience, including genetic optimization. There is no company that's worked more fervently on saying we must bring genetic optimization to everybody, no matter what you look like, no matter where you're from, and ideally, ultimately, no matter how much money you have, you should be able to access it. It should just be a good— it should just be a consumer good. That is so important to us, and I'm very proud that we're leading the mission that way.

I guess that debunks the other misconception that many people have that this is a tool for billionaires and everybody else is gonna be another client.

Exactly.

I agree with you 100%.

But here's the thing. Here's the thing. When people say that, it's because people have set that narrative out into the world. Why? Like, if you're a company and you sell to, you know, the wealthy of the wealthy of the wealthy, of course you're gonna say that 'cause you wanna close your deal. You wanna be, oh no, no, this is only for the tippity top. You're exclusive. No, this is for anyone undergoing IVF. Nucleus is for any parent that ultimately— and this is what I find so amazing about patients. No matter how much money you have, they want the same thing. They want a beautiful, healthy child. That's it. That's it. I mean, this isn't, you know, they don't want a baby to be born that's like jacked and no, they just want a beautiful, healthy child.

Some people maybe do, but they're probably in the minority. Maybe I do.

Eventually, eventually want their kid to be jacked. That's fine. They can go build muscle. They can go hit the gym.

I want my kids to be X-Men.

Yeah, exactly. But you know what I'm saying? People just want their kid, they want to give their kid the best start in life. And that is a universal need. And actually it's unique in that universal need, which is sometimes if you think about AI, people use AI in all sorts of different ways and applications. People have a huge range of why they use AI. Some people use it for very simple things. Some people use it for very complex things. What I love about Nucleus Embryo, what I love about IVF Plus is that everyone's actually using it for just really the same thing. They want, at the end of the day, they have a beautiful, healthy baby and that has hopefully lower disease risk. Hopefully it's a little bit taller if they want that. And anything else the parent's preferences and wants. That to me is such a beautiful thing and it's a universally common human goal. It's beautiful.

That doesn't sound like a crime to me at all. And by the way, so you mentioned gene editing right now, which obviously you don't do at the moment, but it makes me very curious. Is that somewhere on the roadmap 10 years from now? Does Nucleus do this, or what does Nucleus 10 years from now look like?

For us, Our key objective is to help patients. And so we would always evaluate the kind of stack of technologies. What do I mean by that? So we recently, there was a very famous scientist. I won't say who they are, but they're very famous. They gave a presentation recently and they talked about sequencing children at birth. And they said that, oh, if this baby's born with this severe disease, you can sequence their DNA and then you can design a lifetime of care for the baby. That's amazing. I would say 2 things though. If mom and dad knew they had that bad thing, they probably wouldn't want to pass that down to their child. I mean, you don't want your child ultimately to have some sort of severe genetic disease that albeit yes, if you can provide them with lifetime care, that's a beautiful thing and that's amazing. I'm just obviously an advocate of that. But if 9 months before you could actually just not pass it down by using genetic optimization, you probably would do that. And so, that's the first thing, which is if you actually consider genetics, You need to consider the entire stack. You need to say, okay, if we do preconception screening where mom and dad do those 3 things, and then we do embryo testing, and then we do embryo analysis, then we do genetic optimization, and then we do all those steps, would there be a step that's needed to use, let's say, genetic engineering at that point? Today, honestly, there probably really wouldn't be because if you actually use the genetic stack, You can actually already, with just embryonic selection, weed out a lot of deleterious mutations. Okay? Then, we can get to the very sci-fi thing. But, in terms of practically speaking, in the end-to-end experience, you can use, and it hasn't been used unfortunately, but you can use modern genetic science today to basically weed out Mendelian diseases from your future child. And moreover, actually, you could use it today to help reduce the incidence of polygenic predictors. No gene editing required. And that's the thing, which is technology needs to meet the use case. Technology shouldn't just be used for technology's sake. It's like applying a more complicated statistical model when a regression model works. You don't actually need to make things more complicated. Gene editing, in many cases, if you consider the genetic stack, would make it more complicated. Now, I think I know what you're going to ask. Go ahead. There's a question.

I was going to say, I'm actually quite interested in the sci-fi scenario. I understand why you wouldn't start with that, but there is a strong current right now with Bryan Johnson and people really optimization maxing, health maxing for what you could say are extreme traits, like aiming for extreme longevity, near immortality, and probably optimizing for other Like extreme sci-fi traits, you could say. And I don't know where it's going to go in the future. I think there is like an opposite push as well. I don't know how it's going to resolve, but I'm kind of wondering, yeah, are we going to get at some point to the point where there is demand for like really sci-fi edits as well?

I think eventually, yeah. I mean, I think people eventually will just want it. I would say the thing to remember though is, again, there's never a silver bullet with these things. People think that they're all going to optimize for something. Nature's a balancing act. Great example of this is actually multiple examples. Height. The extreme version of height is a disease. It's Marfan syndrome. On the other end, it's dwarfism, right? Intelligence. On one end, you're cognitively impaired. On the other end, you could have some really severe autism, for example. Okay. And that's actually interesting if you talk to patients and they understand that if you increase your genetic disposition for intelligence, it may be the case that you saw the increase in your risk for autism. Parents don't want that. And so I think what we're going to find, human beings are going to find, is that there's no silver bullet here. We think that, oh, well, I'm just going to edit for this thing. It ends up going to become that nature's going to inextricably, by the nature of nature, everything's going to be balanced out. That's kind of what I think. So, I actually think it's going to be very similar to embryonic selection in that people are going to think it's like a silver bullet. It's not. It never will be. Genetics by its nature, as I've described, by its nature, it cannot possibly account for everything. So, I think it's going to be another tool in the toolkit of— The complicated, messy thing is that that is life.

Would it personally excite you if it's possible to engineer radical sci-fi things? And if you could choose one, like, I don't know, item in your future child on the menu that you could engineer?

I have something. You know what it would be? I hope they only need a couple hours of sleep. Oh my God, that would be so good. Imagine they wake up, they get 3 hours, they're like fully, they're That to me, it would make them happy, make them more productive. It would make them just, I mean, that would be awesome. I would love that. Oh my God, I get such little sleep, man. If I could actually get that mutation, not that it actually exists, there are a lot of candidate gene studies that turn out to be bogus, but I would love if it existed, some sort of short sleep, high effect genetic marker that I could just pop it out, 2 hours, boom, next podcast, boom, next podcast, boom, next podcast. That'd be cool.

Your future child will make a very good CEO if that's the case.

I mean, yeah, then again, it's like the classic thing where it's like if you push them to be a CEO, they're probably not going to want to be a CEO. You got to let the fundamental essence of your child— I'm not a parent, to be clear. I don't have a child. But I imagine if I did have a child, I would be like, okay, it's own human being that has this beautiful essence. And your job as a parent is to cultivate their beautiful essence and not necessarily inflict what you think is the best or worst thing for them. But every parent probably says that in theory, but then practically it's probably really hard. So I don't know, but these are some of the things I was thinking about.

I would love to engineer a Wolverine type of child who regenerates quickly, ages very slowly, has got super strength and agility. I love that.

I mean, Spider-Man would be cool. I would love that. I mean, if I could just like, you know, maybe what I should do is I should just start like, you know, putting like spiders, you know, around your kid and being like, maybe it happens, you know, maybe they get bit and maybe they become Spider-Man. Just saying, like, you know, it'd be really cool if it's like, pew, it's like, whoa, what? Like, no way. It's like, oh, did you nucleus that? It's like, nah, I just got bit by a spider, old-fashioned mutation.

Well, it does sound pretty sci-fi right now, but I don't know, maybe in a decade or two we can get there.

I think it's going to be longer than a decade, but maybe a couple decades. Maybe it could get more interesting

for sure. I asked you already about AI, but just briefly, do you actually believe in the exponential takeoff scenario when it comes to the intersection of AI and genomics that is really going to exponentially accelerate? What's possible?

I, I do think technology is gonna more rapidly accelerate as AI gets better and better. And even how good AI is, by the way, AI is already really good. Like, the technology is at a point right now where, like, society is gonna spend the next decade, decades assimilating the technology and actually using it and realizing its potential, right? Like, it's, it's like going back to the castle analogy. Every single company's underlying technological assumption has been rug pulled. So you have to rebuild everything. You have to rebuild every castle. But I do think that there's this idea of AGI. I don't think on transformer-based models you're getting AGI. I just don't think it's going to happen. I can go into a whole intuitive thing about this, but I just don't think it's going to happen. I do think you're going to need— I think AI, you're going to need probably new model architectures. There's something about spending a trillion dollars to build data centers, and then my brain just like kind of working, that I always find to be strange. It's like there's something like, you know, you know, like when you ever like do a, a math problem and like let's say you're doing a math proof and you're like, I could solve this math proof if I like, you know, did like 25, like if I did 100 cases of it, but that usually means like you haven't thought of the more elegant solution, right? There's something weird to me about how like efficient the human brain is versus breaking capitalism, like literally breaking capitalism to get enough energy to do something. You have to wonder, is there some upstream assumption that is missed here? I don't know. These are just my own personal thoughts, but I do think that AI is going to— I think it probably is the single greatest lever ever created. I do think that lever is going to be most consequential in the context of biology and specifically in genetics. Because genetics as a diagnostic, genetics as the language of life is probably the most high-leveraged thing you could possibly understand in the world. Why? Because it's the basis of humanity, right? So, if there's anything you could understand, I would probably be hard-pressed to find something that's as important as genetics. Maybe physics, maybe like— but even if you derive quantum entanglement or something like that, It's very theoretical. There's nothing you can really per se do. I mean, I guess you want to understand nature and its reality. That of course you want to do, and that's amazing. But in terms of practically speaking, I probably think physics, probably frontier physics, and that's actually not true too, because eventually the physical theory could become practical things. So that's not perfectly, but I think frontier physics and genetics are probably the most levered consequential things that AI is going to help elucidate. And when I say genetics, you can think of biology more broadly, but those are probably, I think, the most important. Meanwhile, B2B SaaS tools, I'm going to be honest with you, I'm less interested. I'm less interested. I'm more interested in the real insight. I want real innovation. I want there to be on somebody's employee's computer at Nucleus, I want to know there's something on there that's going to fundamentally change the world. Some piece of technology, some model. That's what I want. That's the vibe I like at Nucleus.

So on one hand, you seem very bullish. And on the other hand, I guess you're saying scientists will still not be out of their jobs. We will still need wet lab work, which is going to be somewhat slow and messy and not just going to happen within Meta's data center.

Oh

yeah.

Oh yeah. I think there is a real world, a physical plane component of this. I do think that if you could build the humanoids, that's gonna be very useful. I actually have to run though. I actually realized we're way over. I got way into this. I love the podcast so much that I went on so many—

Yeah. Sorry about that. It was a great chat. So I kind of lost track of time as well.

Yeah, so did I. This was great. I loved it.

Okay. I mean, thanks for coming on. Yeah, really appreciate it. And I won't hold you any longer if you have to run.

Thank you. Thank you so much for having me.