Gene Therapies That Work? - EP 86 Adrian Veres

0:00Adrian Veres: All right, we're talking gene therapies today. It has long been this much hyped, much heralded technology from biotech land where you send a therapy into your body. It is much smarter than your typical therapies. It's going in, fixing your DNA for what ails you. The massive problem with gene therapy tech is that it has been very expensive, hard to produce. It's been hard to send into the right places in the body. Often these gene therapies get all sucked up in the liver and don't go where we need them to go. Our guest today is Adrian Veresh. He is the co-founder and chief science officer at Dino Therapeutics.

0:48Adrian Veres: therapeutics. They have been working on making a better gene therapy for a long time. They were kind of doing this AI bio thing way before it was cool. They've been using AI machine learning technology to reshape viruses. Viruses are often how these gene therapies are kind of built and put into the body. So they've been trying to create a modified virus, call it a platform, a new way to deliver these therapies that is better, smarter, and faster than what's come before. They are a little different than many biotech companies. They're not trying to make just one specific cure for a number of diseases.

1:34Adrian Veres: They're trying to be what people call a platform play. They want to sell their technology to the likes of Roche and Novartis and other companies, some of whom are already their partners. Adrian is going to give us the basics on how gene therapy tech works. He's going to make you smarter on all this stuff. It's just fascinating. Technology is a tremendous lesson, which is what I always say, but this is true. I am Ashley Vance. This is Core Memory. Here we go.

2:21Adrian Veres: Adrian, thank you so much for coming. Excited to be here. I'm excited too. I always try to do my own research, reading things, but then it is the age of AI, so I always type the person's name in, and then I'm like, what would you ask them? and Elise OpenAI is a big fan of yours and it tells me I have a coup. It said, you're clearly one of the more accomplished young scientists in the gene therapy world and have stayed almost entirely out of the press and that I'm breaking new ground here by getting you on. Let's do it. This is, I feel like we're going to be talking about your company, Dyno, and gene therapy.

3:08Adrian Veres: I always feel like this is a difficult subject. I think people hear about gene therapies in the news. Even me, as someone who covers it off and on, I always have trouble understanding where we're at with this technology, sort of how it works. I mean, I think for the average person, you think you're getting some injection and it's rewiring all of your DNA And then, you know, we hear about these things, but then they don't seem to be out in the wild a lot. And so, yeah, I mean, I always like to just explain the basics of some of what's going on here. Does that seem like a decent place to start? Absolutely.

3:51And I think the other thing to realize is I think we're going to hear a lot more and more about gene therapy. Yeah. Like, I think we've been in the background phase. And if like what we and everyone else is working on, what's going to happen is we're going to be hearing more and more about it because it's just going to be a bigger and bigger part of medicine. So it's a great time to get the basics in.

4:11Adrian Veres: I feel like that's the thing. It's just stuff you hear about in the news for so many years. And it obviously would come with all this promise if we have a brand, you know, a whole new field of incredible, quite personalized therapies for people. And then it's just it always feels like a little bit of a letdown because they're incredibly expensive. They're rare. They seem to have bad consequences sometimes. And so, yeah, anyway, that's where, so, so let's, let's start here as one, one item, which I think kind of gets at what you're saying. The federal government recently seemed to kind of like open its mind towards gene therapies, especially in the case of rare diseases.

4:52Adrian Veres: And so for a very long time, these have been people who are this difficult population where it's just not usually cost effective for a big pharma company to develop a drug for someone with this rare genetic disease. And so they get kind of neglected. It turns out for some of these diseases, we call them rare, is actually quite a decent number of people. And if you add up all the rare diseases, it's a very large population of people. And so, you know, it's a difficult circumstance that is, I read about it really quickly, but it seems like this particular administration is saying, we've kind of come far enough with gene therapy technology where if you're in quite dire circumstances, you know, and there's enough evidence that something might work, we're going to be more lenient in letting you try this therapy without having to go through all the usual trials.

5:47Adrian Veres: Is that sort of where we're at and is that indicative of something new? That's part of it. I think that's tying into we're seeing a lot more new ideas on the regulatory side as well as on how people are framing what is a gene therapy. Like is it more of a medicine? Like is it more similar to a pill that you just take once? or is it more like a surgery where, you know, it's like a bigger intervention that you do once with a very specific intent that's actually different from person to person. And a lot of these ideas are pretty new. The reason we're also talking about them is the moment feels right.

6:24Like the convergence of technologies that's enabling this is just getting better and better. And it's helping, you know, what's going to happen now is a lot of the challenge that you mentioned, like cost, safety, how widely applicable it can be. All of these are really technological challenges at the core root, right? There's like really nothing that says, laws of nature don't say we can't do this. We just have to find a way to do it. And, you know, I feel that the progress we've been making as a field in the past few years has just been really incredible and it is going to continue. And, you know, that's why I think we're going to hear, you'll be hearing more and more about gene therapies.

7:05Adrian Veres: Okay. Well, yeah. So let's bounce off that. I mean, why don't we, let's define a gene therapy. Can you help walk us through the different forms that are available? And then some of what you're seeing that's starting to change. Yeah. I think at the simplest level, what a gene therapy is, is acting at the genetic level to improve health. That's really what a gene therapy is. The point is you're going to intervene at the genetic level to have the impact that you want to have. The contrast with this is if you take a small molecule drug, you can choose to act at many different levels of the biology.

7:48You could act at the symptom. You could act at the very last thing that's giving you a certain feeling of pain, for example. You can try to suppress that. Part of gene therapy is as a framing it's restricted to where you're doing this intervention. And what's helpful to know is we've been wanting to do this for a very long time. The reason for it is in understanding disease, if you go back about 50 years, we quickly realized that there's a lot of diseases that we could tie back where that disease is at the genetic level. Meaning once we started, we understood that there's this thing called DNA.

8:23It's this like stretch of letters. And we started asking like, can we find, you know, there's this terrible disease that exists. Can we understand its cause like tied back to the letters that have changed? And once we started doing that, the answer was, was like, yes, we can. It used to be really difficult. Now it's gotten dramatically easier because of technology. But very quickly, we were able to point to, hey, there's this gene that causes this disease. In fact, there's many genes that are named in that way, right? Like the cystic fibrosis, like implicated gene is really called CFTR. The CF stands for cystic fibrosis.

9:00And we've been wanting to, you know, as soon as we knew that, the next step was like, okay, can we fix it?

9:05Adrian Veres: Can we just turn it off or, yeah. Yeah, like, there's a mistake. Can you just go in and fix it? And that journey has been going on for decades now because it turns out that it wasn't so easy to just will it into action. There's a lot of components that we needed to bring into place. Even actually just the changing of the letters of the DNA, that's like the promise of CRISPR, which everybody's heard about CRISPR. Why was it such a big promise, such a big thing in our field and continues to be is because it brought in the ability to go make those changes. But then another problem was, say you're actually able to poke at the DNA, like change the actual letters with some tool, some technology like CRISPR.

9:47Well, you have to get the CRISPR in the cell at the right place for it to act. And that's another big challenge. And actually probably the bigger challenge that's held back gene therapy is the ability to do delivery effectively, go in and, you know, actually there's a bunch of different ways you can intervene on DNA and other components that are close to DNA. but actually getting the machines in in the right cells safely cheaply that's been a big challenge and that's actually the challenge that like we at dyno uh you know we founded the company to focus exclusively on that challenge that's what we focused it to work on that and so so like uh when

10:25Adrian Veres: we talk about small molecules like a pill that we take i mean is it fair to kind of talk about that more as like kind of mitigating a problem but it's not it's not fully turned something's not turned off or turned on like you want it to be you're always you're always still dealing with this condition in some way i think that's a bit unfair okay to amazing medicine yeah like i think the reality is it the small molecules give you more degree of freedom in terms of where you can go and you can use that to have a causal intervention like if there's a way to to do the intervention. There's a small molecules can do it on certain contexts.

11:07The problem is that it's just a very hard, like, hacking problem, right? It's like if maybe a programming analogy might help, it's you can't fix the soft, like, actually change the source code. And sometimes there's just an actual problem in the source code. Or even if the problem isn't in the source code in the first place, if you could go in and change the source code, that'd just be a very direct way to fix it. And instead with small molecules, sometimes you have to like find some other way of like, oh, this thing is happening here. So if I, you know, poke it here, you have a direct impact.

11:39And genetic medicines, you know, complement other traditional medicines, but they offer a different angle of attack, a different point of leverage. That's like very, very high because once you change the source code, the premise is you should be able to make a change that lasts forever for a very long time. And that's the potential of, that's what's such a big potential and why the field has continued and in the face of many, many challenges, just kept at it decade after decade.

12:10Adrian Veres: And over these couple of decades, I mean, I hear all the time about monogenetic conditions, the rare diseases that we're talking about. Often they fall into this group where it's a single gene that is causing a lot of mayhem and it's disrupting a very important process. As time has gone on, well, obviously there's other conditions that have many genes at work. And when we're talking about gene therapies broadly, are we in this moment in time just very, very focused still on these single gene problems? Or is there a thought? I could never tell if once we get to the multi-gene level, it's just too complicated.

12:52Adrian Veres: to even begin to address that? I think it's a matter of the order of events. Like it's sort of, it's very natural that where there is a single gene, the other thing that's very helpful in a monogenic disease is you understand the causal biology. Like you really know that if you, you have all the evidence that if you go in and fix, make this change, you're going to have the impact you want to have. And that's actually just really valuable in biology because often in biology, the trick is in biology is you can have a story that's like really beautiful. It makes a lot of sense. There's like a very clear argument.

13:33And then we've seen time and time again that when you actually test the argument, you're missing some component in the story. And so it doesn't work out. For monogenic diseases, that's so crystal clear. So that's why the focus is there. but nobody in the gene therapy field is saying like cool we're going to do gene like single gene diseases and then we'll have to find something else i think we're actually all excited about the the potential to to intervene at the circuit level like to have to intervene on on a combination on things to really shape cellular behavior and um you know that's absolutely coming in in pretty soon actually given where the technology is uh the pace of technological development right now.

14:15Adrian Veres: And so dyno is focused on virus, viruses, viral vector delivery. What are the methods at our disposal in 2026 if I want to deliver a gene therapy? What are my options? There's a couple, yeah, there's a couple options. The challenge, ultimately the challenge, it's helpful to understand the challenge is you have a piece of DNA or a piece of RNA, You have to get it inside of a cell. And there's really like two ways to do this. One is you have to put it in some sort of box that's going to get into the cells. And that box can be like chemically formulated. This is something like a lipid nanoparticle where you just mix a few components around and depending on the conditions, you can now achieve that goal.

15:04Then there's another set of things, which is what we normally do in biology is rather than invent new things, we look at what nature's made and we like borrow, steal it. And viruses do this by default. There's many different viruses that do it. And you can take them to do this delivery for you. So better yet is you can actually, which is what we do, is we engineer the system to actually make it do not just what it was naturally evolved to do, but we design it towards a specific purpose and really turn it into something we picked up from nature into a technology that's going to be able to give you the characteristics you want for any disease that you might want to be treating.

15:50Adrian Veres: This episode is brought to you by Brex. I spend all of my time inside of the world's most ambitious companies making rockets, semiconductors, robots. Very often the business is struggling and then all of a sudden things click and take off. Then the company's expenses spin out of control. They either have to clamp down, start counting pennies, or cross their fingers, hope for the best that the quarter or the year comes out okay. Brex was built to handle exactly these situations. It helps you move fast, spend smarter at the same time. They've got credit cards, banking, expense management all in one place.

16:30Adrian Veres: They've got AI, technology, looking over everything, taking care of tasks for you. No more chasing down receipts. We use Brexit Core Memory. It has been a godsend in giving me back so much time. giants like Anthropic and Zoom use Brex as well, and so do tens of thousands of startups. You should use Brex, and you can find out exactly how to do that by going to brex.com slash core memory. Okay, and then tell me how badly, I butcher this, but so viruses have this kind of outer shell called a capsid, I think. I'll go. Some viruses do, but yeah, they all do in some way. So far, so good. And so nature is very good at this.

17:16Adrian Veres: Like you said, I mean, we look to nature, obviously, we all know once you get a virus, it gets in there and it goes to work pretty quickly sometimes doing what it's trying to do. And so you want to create a virus that's not just doing what it normally wants to do. You want to reprogram it to do what we want out of it, but to kind of use its mechanism. And so there's this outer layer that it seems like we can manipulate and try to get the genetic code that we're, the tweaks that we're trying to make into this outer layer and then have this system feed into your body? Yeah. So let's talk about that.

17:53There's a couple of interesting things. I think it's actually fair to, these aren't viruses anymore. Once you start using them as, you know, you take out just the shell. Okay. Like we've borrowed a component from a virus, but it's helpful to know like fundamentally what makes a virus a virus is it can go in. It's a machine that's a self-replicating machine. It's kind of one of the minimal. It's not the most minimal. That's probably prions, but it's one of the smallest self-replicating bits of biology out there. And its goal is to self-replicate. Well, we don't have that goal, right? Like we actually want to, in its process of self-replication, it has to deliver itself somewhere where it can, you know, take over the cell for its own purposes of replication.

18:36We're borrowing just the delivery components, so the capsid. So that's actually the component that we're engineering. We're engineering the capsid. And you're also hinting at something interesting, which is quite a bit of viruses. We know many of them, but we actually know a tiny fraction of what's out there in people, also in the animal kingdom overall. But what And what the field has done is actually there's a virus that has piqued a lot of curiosity because it just seems to, it's just kind of not a very good virus fundamentally. In fact, it can't even replicate on its own. Like it actually needs another virus around to be able to replicate.

19:17And it seemed to be that it's very simple. So it's protein only. And it doesn't cause much of a reaction to the body. And also it doesn't cause disease. So if you have, if you are infected with that virus, you actually never really know. And so that virus is called adeno-associated virus. It's AAV. And because of these set of properties, it was just very intriguing for the field for decades, being like, wow, if we had to take over the components and use this to make, to deliver gene therapies, this is a great starting point. And that's the story. Those are the giants whose shoulders we're on, is this progress of finding this one, testing this hypothesis, building it and building it.

20:03And what's happened in the last few years is a convergence of technologies, including AI and this ability to explore the sequence space in new ways than before that have converged, both in the lab and in AI, and enabling us to really just customize these capsid shells to do exactly what it is we want them. And that's been our focus.

20:24Adrian Veres: Do we know why this particular virus, like what is its purpose then in the world? I mean, it's hard to speculate on these things. I, you know, by default is this, it's, uh, it, you know, if I had to speculate, I'd just say like, like most viruses, they want to replicate. That's it. Their purpose is to self-replicate and continue. It seems like it's like bad at replicating or? I mean, you know, it's making do. Like one of these, it doesn't have to be the best. Like if it weren't able to do it at all, it wouldn't exist. Okay. So clearly it's good enough. And you know, that, that's good enough for me.

20:58The interesting thing is the, you know, it's, it's got this funny name, Adino. associated because people were studying adenovirus. And then when they were looking like in electron microscopes, they saw like, wow, there's this other thing that sometimes pops up. It looks like a virus. Could it be? And it took a while to figure out like, yes, it actually is. And it's interesting in its own right. And yeah.

21:19Adrian Veres: And so when you guys are, when you're talking about changing the shell, I mean, what are you doing? What is actually happening? And then what are those changes allow you to do maybe it's helpful to describe what this is because it's actually it sounds complicated but it isn't you can picture something very very simply so this capsid is it's helpful to picture like uh i said a box that undersells it a little picture a really nice like origami box okay and what i and what you want to picture is this like icosahedron thing and there's one piece that is repeated 60 times. That piece is actually very small.

22:02So that's a protein. And that piece is defined by about 700 amino acids, which are the letters that define proteins. So it's just a string. It's really quite literally just a string of these 700 amino acids. Now, those amino acids are incredible, right? Because those 700 amino acids are able to together create this icosahedral origami box. And then that box is able to, you can put DNA in that box. So when we do the engineering, what we're doing simply is really just changing the letters. Of those 700 letters, if we find the right change, well, we can directly, you know, that can make it, that can change it if we want it to go to the brain because we want to make gene therapies for a neurological disease.

22:55If you find the right set of changes, you can direct it there. And now there's a lot of very interesting science for how we do that, but that's the core concept of the problem we're out to solve.

23:06Adrian Veres: And nature has had millions and millions of years to work on this, but the thought is that we can make these adjustments and still be, you know, good. at creating a change like even one-upping nature here? So there's a couple challenges, right? One is nature was never optimizing for our own goal, right? It wasn't trying to make an amazing delivery vehicle for brain gene therapies or muscle gene therapies. I mean, I guess we are now, so nature is finding a way in a complicated path, but the actual evolutionary process at the viral level wasn't aiming to create that. So that's why there's, you know, it's obvious that once we set a new objective, we're going to be able to do the engineering.

23:58Now that engineering though is pretty hard because what's happened is if you look at nature and it's a very, very finely crafted machine, that's that little protein shell. And we've actually run this experiment of what happens if you take those 700 letters and make for every one amino acid, there's 19 others. You can make the change to any other amino acid. And we've done this experiment. This is work done in the church lab at Harvard prior to the creation of the company where we made every change. And quickly that explains to you why the problem is hard. When you do that process, about 70 % of the changes just break the shell.

24:38So it tells you that if you're going to be changing it, you really have to understand, okay, if I'm making a change here, Got to make another change somewhere else to make up for it and keep exploring the space. That's one piece. To your question of also how come there's room to do the engineering, one thing that's really cool is if you look at existing AAVs in nature, there's a few that are very different from each other. So you can pick out two things that are clearly AAVs, like they make the same type of protein origami shell, but they differ in like half the positions. And that tells you that actually there's room to engineer.

25:16There's not one solution. There's many possible solutions. So if we explore that space effectively, we can get there. And then the other big piece of good news that we found was once you get that data, so for example, once you make every single change and find the good ones, you start doing combinations of changes. One thing we noticed, and this was before the AI movement that's happening now, this was around 2017, 2018, we noticed that actually machine learning techniques are exquisitely good at picking out patterns and allowing you to get traction on this engineering problem much better than people, much more aggressive, but also just able to pick out amazing patterns.

25:55And that was the core bet of our company was let's apply these technologies of let's measure many changes all at once, which was the sort of traditional church lab type of technology. And then let's combine this by applying machine learning and AI to take the information from all of those changes and put it together towards a very specific practical objective and put it together and see what happens. And now we're a few years into that journey and the results are really exciting to us in terms of the implications. of the impact of this output.

26:27Adrian Veres: Yeah, and I wanted to kind of come back to that a little later in some ways because you guys did start, the company was founded in 2018, I think. And so, you know, that's like transformers are just being written about in a paper. I mean, that was way before this big AI revolution. It's way before 8 million, it feels like AI biotech startups were peppering me with we're using AI to redesign proteins and all this stuff. You guys clearly were kind of on the cutting edge of all this. And maybe we could describe what I was reading about Dino. I mean, the way it's presented in the literature is that you're, for this capsid, this outer shell, that you're using machine learning to try millions and millions of different combinations of these, what, of these amino acids, of these proteins, and then try to get to a point where you have the most promising ones.

27:27Adrian Veres: And then I don't know if you test on animals or not. I presume you probably test on mice and things like that along the way. And then so you're starting with these millions of things. Machine learning is kind of helping winnow it down. And then you get to sort of animal testing. And then you move forward with the most promising ones. Is that the general idea? You have the right pieces. But let me put them in the right order. Um, one thing that, that the field of gene therapy learned pretty early on was, especially for delivery, it's really valuable to get to the ground truth as much as possible.

28:04And this means working in animals from the beginning, especially in the most relevant animal species, for example, like primates. Okay. And the focus there is, is also actually learning, like, these are challenging experiments. These are ethically weighty experiments, making sure we learn the most from each one of these that we have to do. And what we're able to do is we have the ability to make a collection of these different capsids, call it about a million. It's not that large of a number, but I'll get into why that's actually better in many ways. It's actually we chose that number because then we can get measurements on each one.

28:45And we can get that in vivo. So in one experiment, you can measure for this, you know, half a million to a million capsids, how many went to the brain? How many went to the liver? How many went to the muscle? How many went to all the different organs? And allow you to see how are you changing from the natural propensity of the capsid, which...

29:05Adrian Veres: Okay, so you put many in at the same time. Yes. Okay. Yes. And then there's markers on there so you can track them. Yeah, so the way to think about this is since you have that protein shell and inside of it is DNA, Every time we make a change on the outside of the protein, which is what we're trying to do, we record the nature of that change on the piece of DNA that goes inside the shell. It's using a barcode. We're essentially encrypting it inside the box. But then what that allows you to do is when you actually, you can mix several of these together. They're labeled by the piece of DNA that's inside the shell.

29:43And when you put that in an animal and it goes to the brain, if you then in the brain take out the DNA and say like, all right, what genes have been delivered here by AAV? And then you read the barcodes and then you can say, okay, so which capsid did that? That's a new, the natural one didn't do this or did it very poorly. What about which of the new ones did it? And that gives you exactly the type of data that you want to understand. Now, the problem is the first time you do it, it just doesn't work very well.

30:12Adrian Veres: Okay. Like you barely see any signal, right? Like you'll see a wisp of like, wow, maybe it went a little away from the liver and a little into the brain. But as you iterate, as you take that data, feed it into machine learning algorithms and ask it, okay, learn from this, find the combinations which are going to be most promising in silico. And when you're doing this design, you also get to specify exactly your wish list. Like you can say, go here, go here, don't go here, don't have these features, be easy to make so that this doesn't cost a ton of money at the end. Like you can actually specify all these features and have the machine learning compute stack just like run through billions and billions of possibilities and then come down to, all right, the next million you should test is this one.

31:01And then we do the experiment again and again and again. And that's the, you know, that's the approach that we've been taking and we found to be very fruitful.

31:09Adrian Veres: Are the markers, do they fluoresce or how do you find that barcode? So the barcode is a piece of DNA. And because it's a piece of DNA, the two things are going to happen. And we actually measure both because we care a lot about this data. DNA gets expressed into RNA. That's how genes turn from their DNA information into a protein that can do stuff. And both DNA and RNA are information carrying molecules. Like their literal purpose in the cell is to just ferry information around, which means they're actually kind of easy to read and write. So we just read them off with sequencing. You just sequence it.

31:46Yeah. And we do this in a pooled fashion. And because everything is barcoded, when we see a certain sequence, we know that that came from a certain virus we designed in a certain way. And then we can use benefit from that information.

31:59Adrian Veres: So you give an animal an injection? Yes. And then it acts for a while and then you just take a blood sample or a tissue sample? So we take out tissue samples of every tissue so that we can learn as much as possible. And see what happens where. Exactly. Sequence each of those tissues. And then ultimately this just becomes this like giant table. I see. Like literal, you could open it in Excel. That's like every tissue, every virus you tried, did it, you know, go, is it more than the baseline or less than the baseline? And then from that measurement, you can look at it. Actually, sometimes you see like, wow, that's really promising.

32:34saying we should just validate that one because it's, you know, maybe it does what we need it to do. Or at least it looks like a step in the right direction. We can confirm that. But then you can then also repeat this process.

32:48Adrian Veres: All right, pod friends, I am here to tell you about Send, Cut, Send. They are a manufacturing phenomenon. Whether you're working on your own car or you're a giant automotive company, a rocket maker. If you need metal cut, machine, or bent, you're going to SendCutSend. They'll do it for you right, and they will do it for you fast. Go to SendCutSend.com slash core memory for a 15 % discount on whatever you're trying to make. We love SendCutSend. Go visit them today. And you mentioned the liver pretty early on, which is an issue here, right? Because we've seen that gene therapies tend to, the liver likes to absorb things.

33:33Adrian Veres: And so it's picking these things up. And this has been a problem, right? Is you want to create something to help a condition in the central nervous system, in the eye, in the brain, and the liver is gobbling everything up and then it's creating toxicity issues. And I mean, that's part of this quest, right? Is to get this out of the liver and to the spot where you want it to go. Nailed it. Okay. Right. And this is, I see on your website, I mean, you guys really emphasize this a lot. So I, this is, because this has been the central knock on delivery is just that we come up with all these great ideas and the liver sucks everything up and kind of ruins the party.

34:10It's more than sucks everything up. In some cases, the viruses are homing there because that's where they're going to do their replication. And so if you take the delivery component from that virus, it's not surprising, it's going to keep going to the liver. That's one piece of it. The other piece of it is, is helpful to understand, you know, why are we seeing the toxicities that we're seeing? The reason for it is actually, if you take just the capsid from a natural virus, it almost looks like it should work. So it's like quite, and then when you try it, you're like, wow, this thing's actually quite safe.

34:42You can like put in a lot, like a huge number compared to almost any other virus before the immune system, you know, has even almost any reaction. And then so what we do is we like just turn this crank. Because it all, like, for example, if you're making a muscle gene therapy with an existing capsid, it goes there a little. So if you increase the dose enough, you're going to get enough delivery. And then you're, but you're faced with this trade-off of you probably, you can't go as high of a dose as you would need, but you want to go high enough to have a shot at having a strong impact. But then you're kind of, you're there, we call this the therapeutic index, like it's just very narrow.

35:22And that's why we're so focused on, well, if we can just massively detarget the liver, it's going to widen completely the therapeutic index. But it's also a place where, you know, we could do just that. That would be great. But actually, you know, we're more ambitious than that because we can detarget the liver to help with safety and at the same time just like make it go more to the place it needs to go.

35:44Adrian Veres: Okay. Okay. And then on you guys' website, you know, I noticed you were calling out promising work around muscles, around the eye, around the brain and central nervous system. They kind of get lumped together sometimes. Like, I would like to, I'd love to break down for people still how this works when it's doing what it's supposed to be doing. I don't know if we want to go through all three or should we pick one of these? Let's pick one. Let's pick the brain. Okay, so let's do the brain and let's kind of describe. So you've done all this work to make this capsid the way you want, and you're getting past the liver.

36:21Adrian Veres: So what are we doing in the brain? So you've got this protein shell with a piece of DNA in it. It's going in the blood. You inject it in the blood, and it's just moving around in the blood. The first challenge you face is the brain actually, nature's designed it quite well. one of the things it has is this thing called a blood-brain barrier. Think of it as just a bouncer. That's like literally deciding like, oh, you want to get in the brain, you have to have very specific things to let in. This is the ultimate VIP. Exactly. But, you know, we can learn about that bouncer and we can allow our protein shell to make it through that first barrier.

37:04That's actually one of the big challenges in the brain, the first one. Okay. Now you're through the blood-brain barrier, there's a neuron sitting there. So the virus just touches the cell. And this is part, this, I'm going to have to use somewhat vague words because we barely understand how this actually works in practice at like the molecular level, but we have hypotheses and it's going to start making it inside the cell. And then the virus is actually, the cell is going to start detecting, like, this is kind of weird. I should just put it in the recycling bin. And when that process starts to happen, the capsid knows, okay, it's actually time for me to, if I don't do anything, I'm going to end up in a recycling event.

37:42So it's time for me to get out. It sort of wriggles out of the state that it's in and then has and finds a way to make it to the nucleus where the DNA of the cell resides. Once it's there, the job of the delivery platform is accomplished. You've now delivered a new piece of DNA inside of a neuron. And, you know, I've described one neuron with one protein, but one capsid. Obviously, you're doing this with billions of it all over the brain. That's the goal. And now what can happen? Now it's up to you. If you are making a drug, you can now do any number of things. For example, that piece of DNA could express a CRISPR system that goes in and like fixes a specific genetic disease.

38:28Maybe you don't even need to do that? Maybe the gene is just broken and missing. So all you need to do is you can just express the gene directly. And that's it. That gives you a path to curing a specific disease. And this also highlights actually, you know, I'm not describing to you a drug. I'm not actually telling you, I'm going to go in and express that one gene to cure that one disease. That's actually because that's not what we're doing. We are not building drugs, which is different for our field. we're building this delivery technology.

38:57Adrian Veres: The mechanism, yeah. And then working with many partners who say, I need delivery technology, and I have a very, really good thesis for how I'm going to put in a payload inside of this capsid and go and intervene causally on a disease. And, you know, we're doing that together. And that type of partnership has also been, like, just really, really fun in the story of our company. Yeah, and just going back to the neuron for one second. I mean, so once you've made these changes, the neuron is doing what? It's either producing more or less of something, essentially? So it has to start by producing, you know, it has that piece of DNA.

39:33It has to start by turning that into RNA. And this is where things get, like, kind of cool. Okay. And, you know, let's find the right altitude. But that piece of RNA can do a number of things. The piece of RNA itself could go in and make changes to other pieces of RNA. So the simplest is it could go and, like, shut off another piece of RNA. So that enables you to turn a gene off, which for many different diseases, that's a great starting point. Let's do that. Actually, there's pieces of RNA which now are more complicated and can do interesting things. They can actually edit the RNA directly. And that's allowing you to now change the behavior of another gene in a way that's going to help improve health.

40:14That's very direct. So that's one way. And this is even just RNA. It's actually, these are maybe the more, you know, sophisticated for the connoisseur approaches. The simplest one to think about is, you know, let's go from DNA, you go from RNA, then you make a protein. And now that protein, maybe the natural protein is missing in some way. So you just have to replace it. And there's many gene therapies that have that approach. In some cases, that protein itself is then going to go do something more complicated. So, for example, if that protein is a gene editor itself, then that protein is going to go and find the genome, make that change.

40:51And you can now kind of see like what the potential of genetic medicine is, right? Like you have all of these different ways to intervene on this like central repository of information that is the DNA and RNA of the cell. And ultimately, it's such a point of high leverage to go after disease.

41:10Adrian Veres: And like you mentioned, I mean, so if we're talking muscle, eye, brain, you guys, normally I hop on some biotech company's website and they have their pipeline of their drugs and which stage of a trial they're at. When I was looking at you guys, like you mentioned, you have this list of pharma partners that you're with. So they are licensing or creating some deal with you around this delivery mechanism, and then they're testing whatever they want on the eye, the brain, their particular – whatever they've discovered, their gene edit that they want to push in there. Exactly right. And it's not just pharma.

41:47It's not just the largest companies, although they are very interested in genetic medicines, and that's an amazing moment there. It's also biotechs at every stage. It's also increasingly nonprofits or families who are trying to make a drug in an end of one setting. And that's, you know, really speaking to the moment that the field is having through this convergence of technologies of if you can bring in, solve one component really well, it's going to unlock everything else. I also want to tag that ultimately if, you know, if in your intro you're hinting at, man, progress in genetic medicine has been kind of slow.

42:26Adrian Veres: Yeah. If you had to blame it all on one piece, it's probably delivery. Well, okay. I was just about to ask you, okay. I feel like we've set a framework, you know, to have everybody up to speed on the basics here, what you guys do. And so, yeah, this is where I wanted to open up is I just have a bunch of questions. I mean, I had Jake B. Craft from Strand Therapeutics on my podcast. I had Jennifer Doudna on, you know, both of whom were talking about this delivery problem. And so, okay, like 20 something years ago, 20 what? Yeah, 26-ish years ago, we sequenced genome. Almost immediately, everyone's like, oh, we're going to cure everything.

43:01Adrian Veres: This is going to be pretty easy. It turns out to be really hard. Some years pass, we get CRISPR, and it's like, okay, well, now we have this great tool to make all these edits. And everybody gets all excited again. And then it turns out, okay, this is hard. Sometimes when we make these edits, we're really ripping up the DNA in aggressive ways. We can't do this on as big a chunks of DNA as we want. and we need to be more precise with how we go about this. And so, I mean, this has been, what happened? Like, was this over enthusiasm or like were people just not paying enough attention to this delivery issue or this is just the nature of how technology is developing?

43:42Adrian Veres: You got to do all these basic steps and then focus on this end game. I think it's - Like, do we fuck up or - I don't, no, no, not at all. I think all of these things, it's very hard sometimes to know, especially in biology, where these systems are so complicated. It's sometimes hard to know what is like really hard versus, actually, this is easy. You just need the right solution. Like once you actually put the right solution in place, like bam, you just sort of solve it. So it's sometimes hard to know like which of these combinations of things and where do you want to go. For a very long time, actually like changing dna in a programmable fashion was this like impossibly complicated problem right it was it felt like how could we do this there's very few machines like there's actually very few ways to copy from nature that's that's our first trick is like can we copy from nature if it's doing this somewhere else and and actually programmable changes the dna not something If you like look at most of nature, you don't see that.

44:50And it took some, you know, incredible scientists to find and be actually, wait, this, this actually has the features, like some very strange system that actually is doing this for a reason to protect itself from viruses. Funny enough, we can, you know, we can borrow this. And then suddenly that problem went from top of the list of, wow, we really need to sort this out to like, we have, we have traction now we need to engineer it. delivery maybe part of it is we felt like we had more more of a handle on it from the get-go because for many years actually we'd picked up these viruses from nature we're going somewhere it was going pretty well it's pretty safe and then we also upped the ante of like where do we actually need to go right part of the challenge is once you realize like oh no to have impact it's not sufficient if if all disease that we needed to treat was in the liver I mean we'd be almost done right but that's the challenge is actually we want to go in many different places we're going to go there really effectively so we just got to put these pieces together

45:49Adrian Veres: okay and then on the when you guys so when you guys are calling out muscle central nervous system I is there a reason it's in those buckets like if you figure this out why can't we be delivering this to all the parts of the body that we want to get everywhere We absolutely want to get everywhere. The reason we started there is it felt strategically to us and really to the field that this is a place where you have a ton of unmet need and probably no other path to get there. For example, on our list, we used to have liver. All NMPs just got pretty good. but actually to getting to cns or to like very effectively getting to every muscle we think using av capsids is a you know ultimately the by far the front runner and with the engineering approaches that we have it's it's you know starting to look like we have the right solutions in hand and we have momentum to make them even better so it's an exciting time for that reason okay and i don't want to put uh words in his mouth because i don't understand this field well enough but i feel like when i talk to jake bcraft you know and other people they always tell me the virus stuff is just really expensive it's just it's never going to be the path because it's when you see how much like these these one-off gene therapies cost it's millions and millions and millions of dollars and they say it doesn't get like that much better um even for broader conditions and and i guess strand has committed to lipid nanoparticles like you're talking about mrna being delivered you know as a much cheaper mechanism?

47:25Adrian Veres: Is there something, first of all, like, you know, I don't know, I guess address that, but is there something you guys have done that makes yours somehow cheaper than what other people have done with viruses? Absolutely. So as you, there's two challenges really. One is, well, first of all, I just want to say on that perspective, I think it's always a tough bet to bet against engineering technology. I think there's a lot of things that a lot of people say, it's like, that's not going to happen. And, you know, I'd say, like, let's try it out and let's see what plays out. I think part of it is just developing the right technology.

48:02There's some incredible companies that are saying, for example, you know, manufacturing even natural AVs. Before we engineer them to make them easy to manufacture, manufacturing natural AV is a pretty expensive process. If you break down the process of, like, why it's hard, there's a bunch of different reasons. There's a very creative startup in San Diego that's saying, what if we just made this in plants? Because it turns out it's just expensive to grow like human cells and giant containers that you have to stir and prevent from being infected. But you know one thing we're like amazing at is just growing.

48:34Have you ever seen a tobacco field? Like that's a lot of biomass. And can we move the engineering there?

48:39Adrian Veres: It's making that AAV sequence in a plant? Yeah, that's exactly what they're working on. And it sounds as crazy as that. Who is that? Circium. Circium, okay. And that type of approach is exactly the type of technology on LOX that we can bring. It also comes down to the community aspect of there's a few different pieces of technology, and in each one, including on lowering cost, folks are working on doing this. Maybe you're saying, actually, that sounds, making avian plant, wow, that sounds too sci-fi. Can you just automate better? make better bioreactors and like use it like just use ai to guide you through the manufacturing process so it becomes cheaper so i think we're already seeing huge unlocks on the cost in the manufacturing side of things let's actually talk about the capsid for a second one thing is imminently engineerable is if you make the capsid 10 times better you can lower the dose 10 times so that's it you've got a direct cost unlock there by just making the capsid itself more efficient Like if you make it go to the brain more than the liver, the ones that were previously going to the liver were wasted.

49:48So you had to make up for it by adding more. Now you can just make cost savings right there. And then the next step is the actual process of manufacturing, the way this nanomachine assembles is also engineerable. And we can focus on that too to drive down cost. So I think ultimately for cost, the perspective here is it's kind of like solar panels. used to be super expensive and people were saying it's not going to scale. And we started doing more and more and the cost went down and down. I think that same trajectory is going to happen to genetic medicines, including delivered by AD.

50:23Adrian Veres: So do you think are we, you know, that solar process was a multi-decade journey? Like, and I, maybe it's impossible for you to speak on behalf of your partners in all this, but you know, if I was getting your therapy, a therapy based on your underlying mechanism, your platform today? And what, I don't know, let's pick the brain again. Um, you know, I, one of your partners has come up with this great, this great drug for, for the brain and it's going to be delivered with your mechanism. Would this be today, like a multi million dollar type thing or no, far from it. Okay. I think, I think we're looking at costs that are, you know, closer to the thousands or low tens of thousands, depending on, on the dose in the approach and even that's like going down.

51:07Okay. So, so already we've dialed in, I mean, in the past, less than the past few years, dialed down a 10, 10 X drop and it's going to continue. So I think the reason for it is kind of obvious is people looked at it and saw it's going to be a huge problem. Yeah. If it costs a million dollars per, to make this drug. Okay. The impact of this drug is going to be tiny. And, um, you know, we've put our technological and engineering capabilities to, to lower that cost and it's worked. So it's advancing.

51:37Adrian Veres: All right. This, this will be my nastiest question. You know, you guys, in some ways you were so far ahead of the game and where we are now. I just wonder if you were almost too specific because if I talk to a company like say Chai today, they'll tell me we're not just doing proteins and amino acids designs around viruses. We can do anything, You know, and you look at, I don't know, just some of the work Google DeepMind puts out. I mean, it seems quite broad in what they're attacking. If I open my inbox today, I guarantee you I would be being pitched by a biotech company that tells me they're using AI to design all kinds of stuff.

52:17Adrian Veres: And so, you know, George Church, who is the professor that you worked under the most? So I worked with Doug Melton. Doug, right. Making beta cells in vitro. You know, different thing. But I mean, you know, both pioneers in their field at the cutting edge, and then it feels like the world has kind of like caught up in some ways to where you guys were thinking. Yeah, so did you, are you, is this a curse or a blessing? Were you too narrow, maybe? You're observing something great. And actually, there's two important things here. One is, I saw this elsewhere. I'll borrow their phrase, but say it, you know, show it in vivo or it didn't happen in biology is like really important.

52:59Adrian Veres: Like I think actual proof in the body, not just in simulation is really key. So that's one piece where I'd say once you do that test, you actually see much more clearly what works and what doesn't. The other piece is we noticed this too. Meaning we noticed that as we got better and better at engineering capsids, we realized that everything we learned on the machine learning side, well, could be applied to proteins more generally. And it's also a place where we're starting to share that technology. So we're presenting our foundation scale models for broad protein engineering that have come from our experience of building capsids.

53:36The other thing that's cool with capsids is these are really hard to make. So in a way, it's like if you're going to the gym and doing the hardest possible workout, out, that's going to help you for the general problem. I see. Sure. And that's the experience we've brought in. And we're also excited to share our technology, both on the models that do the protein engineering, as well as the approaches to use those models effectively, not just have a thing that only machine learning wizard scientists can use, but actually make it very easy for any scientist to deploy this technology at scale confidently.

54:16We've benefited from many, many reps of doing this ourselves where we don't just ship the model, but ship the model, use it, put the sequences, test it in vivo, see what works, be disappointed about all the stuff that didn't work, learn from that lesson and repeat that. So I think being early is never a curse. But you got to keep your mind open to how can you have impact from what you've learned. How will this play out when we see the work that Google DeepMind has done in science?

54:46Adrian Veres: I mean, they've won like a Nobel Prize, you know, and then we see OpenAI, Anthropic, these giant companies with billions of dollars behind them and not infinite compute because it's being dragged in all sorts of places. But the world's biggest computers and they've gotten interested in biotech, obviously, sometimes through a partnership and then sometimes trying to do their own internal work. if I think about fields like brain research and things like that I'll talk to brain researchers who are doing connectome stuff trying to build these maps and you know they have to like to do the work that they want to do they have to fight for time on these Google supercomputers and it's like have you have you buttered up the person at Google enough to sort of be put at the at the front of the line for a startup I mean you know I obviously I can see a world where a focused startup is going to make more progress than some broad company that this is not their sole business and area of attention.

55:44Adrian Veres: On the other hand, it just seems like, man, we get to some point where the foundational model companies have so much resource in their model and their compute that if they, yeah, I mean, do they become competition here? Maybe. The way I'd frame it is, There's two things to say here. One is access to compute is key. Yeah. And, you know, we've mentioned our pharma collaborations. We've also had the privilege and the honor to be partnered with NVIDIA as we've been doing this work. And that's been, you know, a fantastic collaboration to like push us and like work with, collaborate at the very leading edge.

56:26So, you know, we absolutely want to stay there because it's critical. I think there's three things. To make AI work, you really need a combination of three things. You need compute. You need algorithms. You need data. Data is a place where if you don't have the right data, it's just really hard to intuit the shape of the problem, especially when you're working with completely new problems or new questions. And delivery in humans or in primates is a problem where you just don't have the data. You have to go make the data to apply it. And it's also not data that's publicly available. So we've created that own data set.

57:00The other part that to me is exciting about all this progress and the scale of people wanting to contribute to this is, you know, we're not, we're doing this for our mission, right? Like we're doing this to make genetic medicines happen. We're calling this concept genetic agency. Like we want to be in a place where you can make decisions. There's not one drug available for some diseases. There's several drugs for every disease. And you can make choices about which one is right for you and when. And that's a really meaningful problem. It's going to take a lot of effort. And applying the combined effort of all of these individuals, organizations, now increasingly AIs, we're going to need all of that.

57:43It's just not a problem that we're going to look around and be like, wow, that was easy to solve. And so the more we increase the scale of the resources that are deployed to this, I think that's all good news.

57:54Adrian Veres: Okay. And a bit on that point, I mean, I asked you on the cost elements. And again, maybe it's hard because you can't talk about what your partners are doing or you don't want to speak on their behalf. But this initial batch of work that's being done with your technology, is that going to be on – will the therapies that result from that in the near term be for limited, more on the line of the rare disease type populations? Or are some of your partners going after quite broad diseases in the general population? Like you guessed, I can't comment on any specific, but I can tell you that the answer is yes.

58:33It applies to each one. It's also part of it is we have several partners that have different strategies. Some partners for, you know, if you're a foundation focused on one disease, they're going to be focused on that one disease. And it's going to be as rare or common as they want it to be because that's their mission. Then there's organizations that are working across several diseases and, you know, we're happy to support that. And that comes down to why we made the company the way we wanted to do it. We wanted to, as we created these capsids that unlocked possibilities, we wanted to make sure that the impact wasn't limited by the drugs we could make using that one capsid.

59:12But instead, the impact could be maximized by everyone being able to, who wanted the technology, could access it and use it and build their programs. And so, you know, it's kind of this full court press on, on, on disease using genetic medicines with, you know, highly engineered delivery capabilities.

59:30Adrian Veres: Okay. Okay. All right. We're going to broaden out for a second. Cause I have, I have like more general questions for you. I mean, one we've been dancing around or we've, we've addressed it somewhat, but just to put a fine point on it. For 20 years, I have been told that biology, we've gotten to the point where biology is like engineering. And we've just gone through many reasons why it does not seem exactly like computer programming. I mean, is this ultimately like a false problem? I felt like 20 years ago, synthetic biology became this term that started to enter the world. And for the reasons I outlined, we sequenced the whole genome.

1:00:11Adrian Veres: It's like, okay, this is just, this is, this is code. This is math. This is all stuff we know well. I think your background, you're a physicist and a chemistry undergrad at Harvard. I mean, so you're kind of approaching this from the brain of somebody who's wants to engineer the world and deal with these concrete things that we can manipulate. I just, and I don't feel totally let down. I love this stuff. I cover it all the time. I know things take a long time to develop. But like, have we been sold a false bill of goods? Will biology always be so much more complicated than code in a software program?

1:00:52That's funny because I want to say yes, but I worried you'd be disappointed. To me, that's a good thing.

1:00:56Adrian Veres: That's a good thing. Okay. But let me explain that. I think unpacking part of it is, I think the challenge with biology is this multi-scale computational process. and part of it is we have this desire to be to be like cool okay we understand one of the layers that's it we're good yeah because if we had if you write code or if you actually engineer a system you want to have like if you're going to have multiple scales you actually want to have like very elegant abstractions that separate them so that you can you know operate on one independent from the other. Biology just doesn't play with those rules.

1:01:36Like it's just squishy. And it's allowed to just make shortcuts between the scales and how things impact. And that's the main challenge is it's actually very hard for us to keep track of a lot of that. That said, every time we make improvements on the ability to measure the biology, the ability to make sense of what we measure, the ability to interrogate it in new ways, we actually unlock more of that bigger picture, there's probably always going to be surprises. Like even in, you know, the far future, we're going to look at something, we're going to find some bug somewhere that does something or like, wow, we'd never thought of this.

1:02:12But, you know, I think that's what is framing the problem that's hard. That's also why it's beautiful. Like that's, there's this complexity that's just such a puzzle and it is so satisfying as a scientist. And, you know, for a while we were starting dyno as a company it was just so fun pitching machine learning people to on this problem being like hey 700 letters goes everywhere in the body yeah like that's it there and now there's like a question mark in between help us puzzle the question mark and you can see people be like oh wow yeah so you know it's coming um will it ever be completely computable i i don't

1:02:50Adrian Veres: think so um but i think that's great yeah i just suddenly you know i just get i get hopeful that I get frustrated. It's like if I cover IPSCs, I cover mRNA, what you guys are doing, I get so excited. And then if I talk to each person in these various fields, you know, the IPSC people will tell me, no, you know, like 10 years, we got this thing working for sure. Going to reprogram our organs back to their state. And then I go talk to my other very smart, rational friend, and they're like, no, whole field's bullshit. You know, it's not going to work. I always feel like there's, there's a rug that gets pulled out from under us on, on each one of these areas.

1:03:29Adrian Veres: And that, um, I mean, I got so synthetic biology when I first started to hear about it, just, I was like, this makes sense. This is blowing my mind. How amazing. And then we have done amazing things, but yeah, yeah, it just, it feels just like a brutal, brutal slog. And I think everyone's, I think the way it gets presented to the public sometimes by people like me is like, you know, somebody is going to unlock this universal key that cures all illness and it just it just seems like um because it is it does get presented as dna gets presented as software that we can just fix and it just seems so difficult i i think it's i think what's helpful there is i i understand the disappointment and i think it kind of comes down to what's different between biology and you know other forms of technology and the difference in my mind is biology is underpinning medicine and with medicine comes just very real stuff disease of yourself disease of loved ones and with that comes a desire for like you know you have this like iphone in your pocket that's this like incredible piece of technology and then you might have a family member with a disease that we've been studying for 50 years and there's nothing you can do and it's like hard to believe what's going on there and I think that's one of the elements where you know if if that makes people want feel impatient I think that's actually that makes sense that's normal like I'm we feel that that's a that's a totally reasonable feeling to have in that position so I think we're part of it is we're holding biology to like a different standard than other technologies in that sense.

1:05:14And I think maybe we should. But I also think that if someone's saying like, oh, all of this stuff is just like vaporware, wet vaporware, that's also wrong. That's also just an overly skeptical view. And the nice thing there is we're seeing the proof in the progress in the clinical trials. Those are an extremely high test. You can't fake it. You actually do the trial and you see, does it work or does it not? And the picture is going to be really clear, but we're seeing, you know, the work I used to work on making beta cells in vitro to cure type one. And it took years to get that into people. But then once it wasn't a person and there's just, you know, people going around who've received these cells that have like all been made in the lab, but they just don't have to inject insulin every day anymore.

1:06:03You know, it's not perfect yet. They have to suppress their immune system for the privilege of not injecting insulin. But for a bunch of people, that trade-off makes perfect sense. You know, that's incredible. And that progress is, you know, absolutely real.

1:06:14Adrian Veres: Okay. You grew up in Romania? I grew up in Canada. In Canada. I was born in Romania. Grew up in Montreal. Great city. So your parents are Romanian? Yes. Are they academics? Mix. Dad's a physicist. Mom's a programmer. Okay. Okay. Yeah. I mean, your academic credentials are quite extraordinary. all these honors and everything at Harvard. I mean, you did undergrad and grad at Harvard. And so what you were, did you assume you did quite well in your teenage years at school? I did well in school. I think the thing that was really special for me before college was, and I got very lucky about this, is I was a very serious nerd participating in science fair competitions.

1:06:59And that was, like, really fun. And that was just such a permission to play.

1:07:03Adrian Veres: Science fair kids always do well, man. You know, you just get that bug at that age and it's hard to let go. No, but like through my career reporting, it's one of the, it's a high signal, you know, somebody who does well in these international, well, even national science fairs, they tend to do quite well in life. I mean, for me, it's almost stronger than that is like for me, going back to the first science fair I participated in, that's where I like bumped into biology. Okay. And that's where like it started. That's what I was going to ask. Yeah. Is it like if you were, if you were starting over again, given everything you know about AI, many of the things we talked about, you know, if you had to pick physics, chemistry, biology, energy.

1:07:48Adrian Veres: So, I mean, would you still pick biology now? Is there any element to this where you see what the AIs are doing and you're like, oh, they're going to do a lot of the work from here on out. then maybe this will be less exciting for me. I think my feeling about what's going to happen there is as the power of AIs grows and they can help us do more and more things, it's going to become, like purpose is going to become increasingly more important. Like if you can just apply this thing to like solve anything, applying it on the things that matter most. And that's a bit of a personal decision of like, Like, you know, what do you feel matters, right?

1:08:31Is it, you know, something beautiful? Like, is it creating beauty, which is really what art is about? Is it understanding how the world works, which is what science is about? You know, doing this with more and more AIs just sounds better. And I think for everyone in the field, we're going to get to continue doing it with increasingly better collaborators, increasingly focused on the things that actually only we can do or we have the strongest leverage, that sounds fun in the years to come.

1:09:03Adrian Veres: So if we sit around, and you're going to sleep at night, and this is the most optimistic case scenario, because you have this window into all this technology and where it's heading. Do you sit around and you're like, fuck in 10 years we've just like cured every disease i can see that very much on the horizon or are you still more like oh this is this is this is a 40 50 60 70 100 year journey we're still on and i mean every disease is a lot you don't have to like take that verbatim but i mean you know like do you sit back and you're like man the world in 10 years people have no idea what's coming for them is going to be a huge change in health.

1:09:47You know, a bunch of my SF friends keep selling me on like, this is how you feel. And it's just kind of not how I view it. I think, I mean, I do hope that in 10 years, like we are ripping through diseases in terms of places where we can have impact and we can just put together pieces of technology to achieve the impact that we want to have.

1:10:06Adrian Veres: Well, you're saying hope, but like you think that like, is there, what's, what's the low-hanging fruit list of the diseases that fall into that 10 years i think if we start listing monogenic diseases yeah um you know things like huntington's disease als um you know sma is already on the list like we've already what's spinal muscular okay so this was one of the first gene therapy with one of the first av gene therapies olgensma was approved for um for sma and some of the kids that received that therapy are now entering their teenage years soon. And there's maybe a, actually we're going to get to learn like is the function of that gene even necessary once you're done growing?

1:10:49And we will learn that information soon. So that list is going to grow, but also diseases like Parkinson's, which are not anywhere near a rare disease. It's one of the most common diseases, but it has single gene handles that we can pull on. And there's quite a few approaches that are being, the trials are starting soon is how I would describe it. So in 10 years, we'll know the answers, which of those work. And at that point, we'll be in a deployment phase for this low hanging fruit. We're in a scale and deploy and also looking to bring in more pieces, bring more diseases, more areas of opportunity to the application of these technologies.

1:11:41I think the other part here is as we grow, you know, there's like aging broadly. There's a place where a combination of different genetic interventions should help, like will help. Question is just let's find the right ones. And, you know, that's, we're absolutely going to be in that direction.

1:11:57Adrian Veres: Like how into the aging stuff are you? Is there anything in particular that excites you? I think I'm a bit more myself as a pessimist there, just honestly, partially because of the time I trained as a physician. I never practiced. But that was a very grounding experience on, like, what can and can't happen. So it's maybe bringing a certain pessimism to me in terms of, you know, what are the, you know, if you look at the max age or like how healthy you are until you die, right? Like understanding these perspectives for me kind of is very grounded in the reality of just meeting many patients through medical training.

1:12:43I am quite hopeful that like increasing health span by quite a bit is absolutely possible as we understand like the drivers of this health span. And I think it's amazing the focus we're bringing and also the will we have to test both hypotheses. I think to me that's like amazing. And I'm really glad we're doing this investment in terms of what it's going to take to increase the like sort of maximal age of a human. And that's going to be tough. Part of me feels like I just don't see as much the like proof of concept unlock that I'm more in that camp of I don't see the reason to know that we have the pieces that like creates this unlock.

1:13:26But at the same time, if there's many, many different diseases we can stack and fix, that's going to have a direct impact on healthspan as well.

1:13:34Adrian Veres: We can. I've got Brian Johnson's olive oil up there. We can have a shot. Amazing. Well, you were making fun of the San Francisco, your San Francisco friends. I go back and forth between Cambridge and San Francisco, and I obviously see what you're talking about. I mean, there are two distinct flavors to these communities. You know, one, when I go to Boston and Cambridge, just absurd. Everyone I meet is an MD, PhD or a PhD, PhD. and just at the cutting edge of incredible breakthroughs. I mean, the Church Lab, Boyden Lab, I feel like they're just knocking out all these people that end up making fantastic startups and everything.

1:14:21Adrian Veres: But there is this more conservative, traditional, slower-moving sort of vibe there out here. It's a bit reckless at times. I just keep running into 20-year-olds who have a bunch of mice in their lab and they'll just come out and be like, I've cured insomnia, you know? And it's like, have you? But so, yeah, I mean, do you take pleasure in both communities or do you find San Francisco just over the top? I was, you know, rubbing my friends though, but I think the trick is find the best of our worlds. And that's really fun as an approach. I think even in the work that we do, like at a practical level, combining bio and AI, you want to try to play to the strengths of both rather than be constrained by the relative weaknesses of both.

1:15:13That's essential. But, yeah, it's fun to see the different vibes in the different communities.

1:15:19Adrian Veres: You don't yearn to move the company to the Bay Area. No, we're not. No? No near-term plans. We're very happy where we are. We have an amazing space. And look to us also, the focus has been, we're just so blessed by this incredible team of really, really creative, really committed scientists. And the same way that Silicon Valley is this for tech in the world, this is what Boston is for biotech. And so we want to continue to benefit from that and continue tapping that. But, you know, we do have some remote folks across the country, including NSF and, you know, playing to win. Yeah. Arne Marogi, who writes for Core Memory, has even hosted a Core Memory podcast, is studying to be a doctor like you.

1:16:08Adrian Veres: Was he a co-founder of China? He was one of our early employees. Early employees. Yes. Should we shame or praise Arne? All praise. All praise for Araneh. I think we're going to start like a Ask the Doctor Q &A series where people can like send in questions to Araneh to be answered. Do you have a favorite? I'm trying to humiliate him here. Trying to think. You know, what can I do? Can I share right here? no but erin i mean i think part of what um what actually probably makes him effective for for core memory yeah uh is i think he sees what things are like sees them as they are and also is willing to speak up yeah which when you're making a company is you know sometimes it's like annoying right like you're like oh this person speaks up in this way like just like just you know just like yeah but then you actually sit down and you're like you're like oh yeah that's you know there's something here.

1:17:10And then I think actually getting good at that is key. So, you know, Aaron, having taught us this loop and having set us this loop of, you know, should have a culture where if something's not how it should be, you should speak up so we can change it, make it better. You know, that's been a great impact beyond the scientific contributions more directly.

1:17:29Adrian Veres: He was a coward. He had to go get his medical degree. I had a lot of fun myself being like, hey, just finish my medical degree. Like, here's some thoughts I have as, you know, someone who, you know, realized I'd have a bigger impact, you know, not practicing myself. Okay. He did send me a couple questions to ask you, one of which is low-hanging fruit, but let's do it anyway. Peptides. All the rage. Should people be playing with their own peptide stack? Let's put it this way. I'm not playing with my own. peptide stack. I think what people should do is up to... There is a framing where as long as it's understood that it's safe, let people do how they please.

1:18:20But I think the question is for depending on the peptide, is it safe or not? It is maybe a question.

1:18:27Adrian Veres: We seem to be tilting to quite a large, ever-growing number of people experimenting with things they're buying off the internet, which I guess like the weightlifting community has been doing for decades and this is just kind of becoming mainstream in some ways i think there are some some you know what's giving me pause for myself anyways is the there are some places where you know there's interesting surprises when you study something for a long time that you don't understand yeah and i think the trick is in when you're making a drug the trade-offs is often very immediate and worth it so you you are willing to tolerate some side effect because there's a very clear like benefit yeah so that you're willing to make that trade i think the risk here is that there might be times when we realize over time like oh this had a side effect and we might also realize like actually we don't understand the benefit or the benefit's not nearly as clear and then people might end up kind of disappointed because they're they're gonna have certainty on the side effect but not on did it do anything and And they might, you know, would have done things differently if they knew that information.

1:19:33At the same time, I think, you know, if we're looking at the peptides, like the GLP-1s, right? Like the fact that they're blurring this line of medicine into consumer product.

1:19:45Adrian Veres: Yeah. That's great. I think that's amazing from the perspective of not tying into this thing of like medicine is this, you know, thing where a small cast of appointed few can tell you what you should and shouldn't do in some conditions. But, you know, you as a person can look at information and understand what something is and decide if it's right for you. And that's that transition. But, you know, it's helpful to point out GLP-1s have been studied for a very long time. And again, the impact is so clear. So it makes the tradeoffs easy. Okay. I ask this next question. I always go to China at some point in the podcast.

1:20:28Adrian Veres: China. And I hope people understand that I love China. I love going there. I am not like some jingoistic, crazy American, you know. Anyway, these questions do not come from that place. I just generally think people are not paying enough attention to the rate at which China is making progress in science and all sorts of fields. I just feel like most Americans are kind of sitting around expecting we're going to dominate all these things for a long time. And, and, and I'm not sure that that's the case. What's, what's like the, so China can run trials faster, cheaper than the United States. It is catching up if not, um, on par with us in many areas of cutting edge science at this point.

1:21:20Um, you know, I, I see a world where these technologies are getting powerful.

1:21:25Adrian Veres: and if we are creating, I don't want to call it like super humans, but if humans are starting to become healthier, gaining functions we didn't expect, I kind of fully expect most of these things to maybe turn up in China first. What's the pro case for American science? What's the best thing we have going for us? Or answer that question in any U.S. or China sort of frame. you prefer yeah i think to me the i agree the rise of like chinese biomedical science has been amazing and it's really interesting to see i think i have friends you know who have very good jobs like professorships at the university of michigan um who are not unlike tech in some ways are not like being actively recruited home they're just choosing to just like go move back to china yeah you know yeah i can see why and and i feel previously if you looked at 20 years ago it's obvious that if you wanted to be cutting edge research you should be in america for biotech now you know there's parity or in depending on the field uh there's just key differences i think the the and i think you're right to point out there's a real risk that we've been at the forefront for so long that we sort of assume that this is just our place to be rather than a place where we need to keep innovating.

1:23:02The hope for American science here is if you look at other historic times, like in rocket science, of times when there was like great power competition where the point was the technological race, we know we can do it. I think we're finding the will to do it, which is exciting. and ultimately I think if competition makes us accelerate that's awesome that's again these are such hard problems which is a clear value if we hit them we should do it I think the risk is there is a risk which is that different countries all do have different somewhat different diseases like what's most common can be related to genetics can be related to you know the the food you eat the society you have.

1:23:48And so there's a real risk that if we, right now as Americans, we more or less get to prioritize the list of diseases based on where we want to apply this focus, or we do this jointly with Europe. We have a responsibility to do this for the whole world, but there's a risk that we need to want to continue being able to allocate and know that we're applying the best technology in the right places. I do think the changes that are coming, for example, from a regulatory perspective are key, right? It's like a wake-up call to be, you know, this is just done faster and simpler here, so why are we doing it in this, you know, cumbersome way?

1:24:25But at the same time, if you look at, for example, the story of maybe KJ, right? Like, we still have the goods. We can run the race. And, you know, the wake-up call from that perspective, I think, is good news. And, you know, let's find the will to speed up.

1:24:46Adrian Veres: Yeah, I'm long American, ingenuity, creativity, work ethic. I'm short on, I see some will. I don't see massive, I don't think people are paying enough attention to sort of the scale of what we're competing against. I think in biotech, it's probably a place where people see it, feel it, get it. Most people do not really. I mean, again, God, I hope this doesn't come off as really crass and nationalistic or something. Most of the, at least 50%, if not more, of the people who are postdocs working in a lab are Chinese immigrants, right? Working at U.S. labs. Is that fair? I don't know if it's up to 50%.

1:25:37It's a lot, right? It's a big number. But I think more generally, right, like myself as a Canadian, I love Montreal. Amazing city. Would have loved to stay in Montreal if I could work at the absolute cutting edge of biology, which is what I wanted to do. But it was clear that coming to America is the right thing to do. I think that's just been the case for a long time. And I think the fact that that is shifting is a bit new. and will want to respond to that and continue to benefit from this massive influx of the most committed and talented people.

1:26:15Adrian Veres: That was sort of my point, is like at this particular moment in time, maybe we don't want to scare off like a bunch of brilliant Chinese researchers and make it really hard for them to come here and work in these labs, you know? Yeah, I mean, at the same time, there's, you know, America's got so many good things going for it. Yeah. down to its like principles, um, you know, got to, got to play to those advantages that that's going to remain true. So. All right. Making me feel a little better, maybe. I don't know. Um, the cool thing about biology is usually whoever's solving these puzzles, this, this, this will be dispersed throughout the world.

1:26:53Adrian Veres: You know, I just, um, so, so that part seems that good. Um, I asked you all of my biggest questions and I've taken up quite a bit of your time. Is there anything like in particular you want to talk about that I didn't hit? To me, it's this, we've talked about it, but maybe it's helpful to like put the pieces together of this like moment. Yeah, yeah, yeah. That's coming in genetic medicine. To us, we've, when a moment is happening, we found it's helpful to label a thing because then you can notice it. To us, we've, we found the best label is this concept of genetic agency. and to me I think that's tying together all of this story so let me just define what is genetic agency it's to us it's the ability of an individual to act at the genetic level to live a healthier life right at the beginning we talked about genetic medicines as being the thing the operant that does the acting thinking about genetic agency as this like now you're like someone's ability to do this and I think part of it is there's this is really like a movement that's building in the field between the ask for people who want to have this, who look at the technology and say, seems like we can put the pieces together and can you do it faster?

1:28:11And like creating the pressure for that. And, and the system, and this various systems also saying like, wait, this is true. We actually need to rise to the occasion.

1:28:18Adrian Veres: You're talking about somebody who's ill. Yes. Or a family member. Okay. Yes. And you're saying, you're saying, okay, well, okay. Going back to where we started. I mean, obviously the U S government is addressing this some way for these very rare cases but yeah i mean what i feel like you're maybe saying is something even broader which is is like um the tools are getting quite good and and and and this applies to a broader group of people who who if they're willing to kind of you know be on the cutting edge could be helped is that yes and the proof of the principle is there too like people are looking at the baby KJ story and saying, can we do it again?

1:28:58And of course the people who did that are, but more people are looking at it and saying, what needs to be true for it to happen to me? How can I drive that? And it's creating this community that previously was like a drug maker that's like making one drug. Now it's this like distributed community of like technology builders ai has a role to play here it's probably one of the like purest applications of you know obviously unalloyed good that ai can play is like furthering this future so the the folks who are building i want to enable that want to apply it in that way so i think that that moment is happening and it's um really exciting to to see for me personally to being a part of but also to invite like others who want to be a part of this.

1:29:47There's just many angles by which folks can contribute to this, you know, mission and movement. So we're excited for that.

1:29:55Adrian Veres: Okay. I'm going to get one final thought from you on that. If anyone, the story of baby KJ is fascinating. It's this child who had a rare genetic disease and we, in very short order, were able to create a therapy which needed to be very fast because his development as a kid was going to be stunted lifelong. And then this was also a debilitating illness. And it was miraculous. A bunch of scientists all got together. It kind of proved out how far we've come with these gene therapy technologies. If you want to learn more, Aaron A. did write a story for us on Core Memory. He also did a podcast, and we also talked about it, Jennifer Doudna, if you want to check it out.

1:30:34Adrian Veres: But so like what you're saying, I mean, I'm just trying to wrap my head around it fully is, is because those are such isolated, um, uh, specific cases. And, and you seem to be saying just that, um, what, that that's just exciting that we've been able to do that. And we've shown that this stuff works and their speed and now the government's adjusting and that maybe, um, even though we're starting with rare disease, uh, just that, that people should be they can contribute to this but also be cognizant um of of technology that's like available yeah i'm saying it's a herald of a time to come okay so you know notice like pay attention to the movement get curious that's you know part of the beauty of the movement is it's also just like really interesting for many different reasons so so approaching it with curiosity is probably the everybody can do that.

1:31:28And then also, um, in some cases approaching it with, you know, morning curiosity with agency in terms of, you know, if I want to get involved in this mission, there's many different ways to do it. And, um, you know, we, we, again, just the problem is so hard that anybody, someone raises their hand and says, yeah, I want to participate. How can I, there's, there's probably a way to find now it's just a matter of, you know, find it, finding that point of leverage and, you know, coming, coming into the genetic agency family.

1:32:01Adrian Veres: Okay. Okay. Last question. And then I'll let you go. And it's along those lines. If you were 18 years old today, would you become, given what you just said, would you become a MD? Would you become a computational biologist? Would you become a biologist? Would you become a computer scientist? I guess we need all of them, but what's, what would be, if knowing what you know now. Biologist. Yeah. Remove the computation, like the computational part is probably going to be more direct, like you'll be able to collaborate with amazing computational AI biologists as well. So I would, but I would focus on the, on the biology just cause it's, it's so interesting and so powerful.

1:32:44Once you understand the interesting bits, um, it's going to be, uh, you know, still, still, I'm confident that for the rest of our lives, it's going to be a thing we talk about, a thing we marvel at and a thing we just get a ton of value from.

1:32:57Adrian Veres: Okay. That was specifically for my son, uh, my oldest son. I want to become a biologist. and he may or may not listen to this podcast. He should come see Dyno. Happy to give him a car. I'm trying, man. I'm trying. He told me he wanted to be a computational biologist when he was like 12 and I haven't been able to let this go even though I think his interest may have shifted. But that's what I would do if I was his age. So Adrian, thank you so much for coming on. Really appreciate it. Everyone can go to Dyno. Was it Dyno Bio? It's dynotx.com. Okay. And there's all kinds of cool information and get a better sense of some of the things we're talking about.

1:33:38Adrian Veres: And thank you for trekking all the way out of here. Thank you. The Core Memory Podcast is hosted by me, Ashley Vance, and or Kylie Robison, or both of us together. It is produced by me and David Nicholson. Our theme song is by James Mercer and John Sortland. and the show is edited always by the John Sortland. Thank you so much to Brex and SendCutSend for all your support. And thank you most of all to everybody for listening or watching. We love you. Please leave us a like, a review, a subscribe, all those tremendous things. Thank you and we'll see you again.