B (3:07)

Yeah, I mean, many. I mean, there was a study out in June from a bunch of Chinese researchers who put. They did overexpression of Fox O3 and they packaged it up in a mesenchymal stem cell that people are doing interesting experiments that have measurable impacts. You, you can read them out or you can visually see them. And I agree with your framework that when you seeing is believing that a lot of people, I mean, biology. I think your point is correct that when I started doing this years ago, I tried to achieve the best biomarkers of anyone in the world. Like, that was like a way to say there's a healthiest person in the world, there's a richest, fastest, can there be the healthiest? And what I really learned is nobody cares about biomarkers. They just wanted to look at skin and face. And you know, and so I had been on this deeply caloric restriction diet and I'd lost so much fat in my face, I hadn't really paid attention to aesthetics and people could not see the project when my face was gaunt. And so it is remarkable that I agree that people understand longevity basically through the face. And so what they see, they can believe or hair. But that really is, I think, the entry point for most people in understanding what is aging. Do therapies work? And if you can't see it there, they don't believe it.

A (4:16)

That's right. And I think there is some sort of meta logic to that because for many people, longevity is about aesthetics and, you know, people wanting to look younger and you know, if they had the trade off between looking younger and actually, in a sense, biologically being younger, some fraction would actually even take the first over the second or what have you. Right. But I think it's almost like the entry point. It's a little bit like, you know, with, with cryptocurrency People seeing was believing because it generated transformative wealth and then they paid attention to the theory underneath. And so if with longevity we can generate transformative health, then we'll get people to pay attention to the theory underneath. So what are the top three or five? You mentioned the Foxo overexpression. Foxo three over expression. Right? Was it foxo or foxp?

B (5:05)

Foxo three.

A (5:06)

Okay, give me like three, four, five other. What are, what are your top ones that have like a visual phenotype that you think is really impressive?

B (5:13)

Yeah, I mean, sadly, I would say there's probably, honestly there's skin therapies, I'd say hyperbaric oxygen therapy, and then just the cumulative effects of good sleep, exercise and nutrition. Those are the ones that aesthetically work the best. Of course you can. Go ahead, go ahead.

A (5:30)

That's on humans. So what about on mice, what about on animals where we can do a lot more in terms of genetic manipulation?

B (5:36)

Yeah, I mean, on mice, I, I guess I've seen enough of these studies where it seems like you can generate really compelling effects on mice in many different ways. And this is seen with caloric restriction, it's seen with rapamycin, it's seen with metformin. So I, in some regards that the, the bar is lower on mice visual outcomes because there's plenty. There's a, there's a lot of them. With humans, I think it's much harder.

A (6:02)

Interesting. So why do you think that is? Do you think that the effect is there in humans but it's not as visible, or do you think it doesn't translate from mus musculus to Homo sapiens?

B (6:15)

Yeah, it's a good question. It reminds me of anyone, anyone who's selling a skin cream basically shows the same, you know, 4, 8, 12 week transformation of a light wrinkle line to a, you know, a smooth wrinkle line or something like that. It's kind of the same with, with my studies. If you, if you see enough of these, you kind of like have a general pattern of like roughly what you show in the mice, whether it's a biopsy or whether it's hair growth or something like that. So I mean, I guess I, I do, there's credibility to it, but I also take it with a bit of a grain of assault of, you know, there's translation issues from, from mice to humans. But yeah, really, I mean, I think mice, mice are compelling. Humans ultimately respond most strongly to other human faces.

A (6:55)

Interesting. Yeah. So, I mean, some of the large effect size things that have been Done in mice, obviously there's the GFP mouse where you can make a mouse glow in the dark by taking green fluorescent protein. You can, you have like a myostat and null mouse where you can make them like incredibly muscular. And there's various treatments that can do that. And there are all kinds of knockout mice and genetically altered mice that do, you know, crazy things. And my view is it should be possible. There's also like the doogie mouse, I don't know if you know this by Joe Shen More than 20 years ago, where they made the mice smarter and they seem to be able to run mazes faster. Right. So my view is that we are finally, after more than 80 years exiting the FDA era where the focus was on minimizing side effect size and finally re entering the era of large effect size. Drugs that may actually have more side effects but also have larger effects like Ozempic, which you don't need to, I mean of course they needed to do studies to determine whether Ozempic worked, but people don't need to see whether Ozempic works. Like, they don't need to like squint and be like, you know, did it work or not as a. It's just such a visible effect that it's like caffeine. Does it, does it wake you up? You can take it and it's like n of one you, you know, it works. Right. So all GLP1s, that's the new era. We're back to the era of the wonder drugs potentially. So why don't we have Ozempic for cognition, why don't we have Exempic for longevity, Ozempic for muscle mass. We should, we will, we could maybe. Right? If we take all of these things, things that we think work in humans or like mice or human adjacent mammals or perhaps like for example, Alzheimer's drugs have a neuroprotective effect, but they also have a neuro enhancing effect, many of them. Right. Like drug repurposing. And so my thought is a, we have all of these interventions on this side which have these amazing visual phenotypes and then B, we have all these jurisdictions on this side. And so what I want to do is offer a million dollar longevity prize for the people who take these visual interventions and carry it all the way through to actually have them legal in these jurisdictions. And when I say legal and functional, they should have a paper at least on archive the documents, everything, all data, open source, open state, and they should have a law there that allows them to do it. Ideally there's some government official who's even welcoming it in the area. And there's 190 countries and a lot of them are doing interesting things in crypto and now they could do interesting things in bio. Let me know your thoughts.

B (9:35)

I agree with you entirely. The thing that we get people most excited is when they see visual effects, by far. And when you do a comparison, in contrast to when we talk about machine intelligence, we talk about orders of magnitude, of scaling, of efficiency. When you talk about human effects, you're saying 10%, 15%, 20%. It's such a disparity between the, the capabilities we have in improving our machines and intelligence and what we can do with human lifespan, healthspan ability. And so the question is, and I think what you're getting at the heart of it is what is the system to create larger effect sizes?

A (10:13)

Yes.

B (10:14)

And so we see for example, like with, with the GLP1s, that is potentially the first mainstream longevity drug where now it's, it's, well, it's characterized well enough where you're seeing cognition benefits, you're seeing, I mean, you're seeing all sorts of benefits out, I mean, in addition to weight loss. And you know, most people, I mean I, I am microdosing that now myself, even though I have no need to lose weight, I can basically try to achieve a dose where I can achieve some of the health benefits. But that, that has very compelling, it's a very compelling offering. You know, it's a very simple drug. And so I think you're right is if we can train people's attention to jurisdictions where you can actually move this thing forward. Because it's very, very hard to innovate in the United States. It's burdensome, it's complicated. And of course it's not without reason. Right. Like they are trying to make things safe. But just over time the bureaucracy owns the situation. It just becomes a paralysis state. So it's the same situation we have with China where China is racing forward on energy and AI and a bunch of other things. And the US is having a difficult time because we have this large regulatory state that slows things down. So I think it's on point.

A (11:19)

Awesome. So yeah, know I might, I, I want to discuss the, whether they make people safe or not. But let's say we do a million dollar longevity prize. Will, and let's say we get it to work. Will you come with me and talk to the, the people there and, and maybe try it out if it's real?

B (11:36)

Yes, absolutely. I mean we are looking for safe, efficacious therapies that have large effect sizes. And that's just, it's really limiting right now in the field.

A (11:47)

Amazing. And the thing is, type 1, meaning just testing safety, is actually very inexpensive. It's type 2, which is testing, quote efficacy, that's much more expensive and just determining whether something is safe. You could have a totally different drug regime that just tested for safety. And there's a lot of things you can do, by the way, like potentially patient derived organoids where you've got like a proxy for the human that you're testing the drug on beforehand. You, you could do much more with pharmacogenomics, you know, PharmGKB and various kinds of resources. If everybody had a genome sequence, you could at least see whether, you know, for example, your warfarin dose is dependent on VKRC1 and CYP2A9. And so you could see what your dose was before taking the drug. You could look at relatives and what their dosage was. You could look at people who were disfamiliar with relatives if you had genomic databases and what their response was. So much we can do with technology in terms of just quickly checking whether something is safe and then willing, willing buyer, willing seller. Right. Like a, like a minimal necessary regulation. Right. And you know, this brings me, you know, when you said whether or not they care about safety, it's a little bit like saying the TSA cares about safety. Do they really? You know, because it's, I mean that's, that's the benign way of looking at it. Right. And the other way of looking at is they just care about optics. And so they're not optimizing, for example, type 1 versus type 2 errors and the sensitivity and specificity, the false positive, false negative trade off. Alex Tabarrock and others at Marginal Revolution have written about drug lag, which is if you have a drug that was approved, but it was delayed by six years by the fda, then all of the incremental morbidity and mortality over that window is directly attributable to fda because the biomedicine was the same at the time that that new drug application that NDA was submitted. Right. It's not like the dry medicine changed that. The approval time was so long that it slowed it down. It was literally just pure information in a sense that they were getting over that time period. And so that is the unseen. You know, one of the things I thought about is running a Google Ad campaign at some point to find somebody who, who had a sibling or a, or a relative or you know, parent that had passed away of some kind of condition during this period and, and say, well, did you know that six years later a drug was approved that would have saved them and then talk to them. And that's a way that you could actually make drug lag visible, you know, on the screen.

B (14:16)

Right.

A (14:16)

That's like one example of how, quote, safety often is actually unsafe because it's so risk averse that it's reward averse. Let me know your thoughts.

B (14:24)

Yeah, I mean, yes, and that statement, if you think about the, the way we do things in actually globally. I was going to say the U.S. but people, we allow people to experiment with fast food and a lot of sugar and. Right. Like you, you are, you have the freedom to kill yourself and to experiment. Does fried food cause me to age and potentially develop disease and lead to my lifetime? So you have the freedom to do that. But if you want to try to experiment to do something good for yourself, you can't. It's against the law. And so it's really backwards in that if you try to step on anyone's ability to sell anything that causes someone to die, you know, people are upset, but yet we can't get the system to move so that we can self experiment on things that could help us. Now clearly there's going to be complicated outcomes. It's not the case that people are going to make good decisions, but that's just humans. So it really is backwards that. And I think it's really a stifling factor on our ability to make any progress. Whereas like when you look at computation, it's just like anything that produces better compute or intelligence go like very little guardrails. So it's really a constraining thing for us humans.

A (15:34)

That's right. I mean the thing is, as I said before, and you've probably also observed, you can go bungee jumping, you can go skydiving. Euthanasia is even legal in many states. Right. Like you could join the military. There's many different contexts in which you're allowed to take a very serious risk of dying. Right. And with essentially no upside. When it's bungee up or skydiving, it's just pure thrills. I mean, fine, right. So why can't you take a risk of taking a new, new drug or device or something like that that could improve your life? Right. Why is, why is human self improvement, of all things, the thing which is stopped? Right. And it's always framed as protecting you from yourself? But what if you had better information? You know, for, for example, the FDA doesn't really like, it's a so called phase four, you know, like post market surveillance. It doesn't include real time reviews from hundreds of countries or like, you know, hundreds of jurisdictions. Like, whereas even like Reddit reviews do the upvotes on Reddit or whatever are drawn from around the world. So you can like crowdsource all of this data from around the world because biology remains the same and that kind of, you know, because it's locked at the level of state rather than the level of network. You know, like, like an Uber rider, they're raiding, sourcing information from New York City and when they take a ride in London and so and so forth. But drug responses aren't for people of similar biology around the world. They're not centrally pulled together and they could be.

B (16:58)

Yeah, yeah, yeah. I mean that the, some of the risks like the GLP1s, for example, like blindness. I think there's like 3,000 plus lawsuits against the GLP1 manufacturers right now. But you can really, you can take precautions by getting your genome sequenced and see if you are at risk for those kinds of conditions. So like you're saying you can do really basic inexpensive tests to see if you are a responder, non responder or a higher risk for these outcomes. And now that our testing is getting so cheap, these, this is a really great path where if, if, if the, if the regulatory structure says hey, this is too complicated and too high risk, then mean you could even offset that and say you, you, you can characterize yourself much better to lower that risk. So I mean it really is, it's very hard to justify the argument that, that we should be as limited as we are.

A (17:45)

That's right. And now so let's say there's three treatments which assume you had a friendly regulator, okay. Somebody who, and crucially by the way, people would sign something that's similar to like being a test pilot because no plane crashes. No planes, no train crashes, no trains. Right. If there weren't people who are willing to take the risk of a plane crash, you never have transatlantic flight, you never have flight at all in the first place. Right. And there were a lot of plane crashes early on because people are figuring it out. Right. And we could think of those people as heroes. So let's assume there was some jurisdiction where it had a crypto like attitude like, you know, where it's willing buyer, bulls car, you're signing all your stuff away, you recognize you're taking a Risk you're an adult and blah, blah, blah, you've been appropriately informed, compensated, whatever. The thing is, what are the three most promising treatments that we should look at that will have large visual effect size?

B (18:42)

I mean, if you look at the evidence on what has worked, it's been gene therapy, cell therapy at the very top and then the, you know, lifestyle thing. So probably gene therapy, cell therapy.

A (18:57)

What about parabiosis for example?

B (18:59)

I'm not sure it's going to have the effect size that would be big.

A (19:01)

Enough because I've seen varying studies on that. You've probably looked at it more closely.

B (19:06)

Yeah, I mean, so I did this with my father. I also did several other treatments. Now you can do the phoresis without a donor and just replacing all of your plasma with, with albumin. People are now generating the synthetic plasma for the replacement. So I think it has some promise, but I don't think it's ever going to have the effect size that a cell therapy or gene therapy would have. Okay, we'll see what it, we'll see. The data comes out. But my guess is that gene therapy cell therapies will, will be the biggest ones.

A (19:33)

Okay.

B (19:33)

And also, and also the OSK stuff.

A (19:35)

Okay. All right, so gene therapy, cell therapy, osk. All right, so let's go in order gene therapy, which genes? What looks interesting?

B (19:43)

I mean the ones that people have been playing with like telomerase or clotho foxo3 follistatin. I mean those are the ones that mostly have been top of charts, but people are working on several others. I mean now that it's really a fascinating field where it is emergent in that we're looking at much more underappreciated genes than we have been before. So yeah, this is probably endless. And it's just the case we haven't been able to characterize these things very well to line them up and say, I mean like for example, when you look at what the new limit. So Brian Armstrong and Blake Byers built stood up new limit. When you look at what they've done with the transcription factors, like basically saying like what what changes things inside the body. They turned it into a computational problem. Whereas before when you're screening like what are the combinations? It's a very hard problem because the combinatorial set is so large. And so now that they, they took a tech approach to a biological problem and they.

A (20:42)

For transcription factor binding sites.

B (20:44)

Exactly. And they've been so much progress in making it a computational problem versus what people were doing before. A much slower academic Approach.

A (20:52)

So I think, yeah, I think some combination can work. The problem is that what you do in silico, and to be clear, I like new limit and it's great. But some complement of those. Because the in silico approach will also have often false positives and false negatives. It depends on how specific that showstore factor binding site kind of thing is. But I think, I think there's definitely promise to combining both. Go ahead.

B (21:13)

Yeah, I'm imagining that it's probably just an intuition building process that even if there is some false positives and you're, you have some degree of accuracy where you can't entirely get there, still it's just this, like how do you, how do you take this insanely large problem, make it somewhat approachable and have a more sophisticated like iteration speed. So I think if you look at their, their progress over the past couple years, that's what I think is probably the potential for gene therapy. You know, people are looking at things like can we do the four hour sleep gene for example. Right. Which is like very appealing. You get more time in life. So I think there's a pretty long list of things people start trying to knock off.

A (21:48)

So let's, let's go through. So let's just talk about that for a second. For less sleep. Right. Less fat, you know, or less weight. Okay. Like oic, but gene therapy. Oic, more muscle. Right. Maybe faster healing perhaps for a lot of things that's, that's helpful. Perhaps, you know, like reversing graying of hair. I mean, I, I actually, I actually think it looks a little distinguished, but that's fine, you know. Right. You know, a lot of the skin stuff like wrinkles and, and so on and so forth. What are the other upgrades that you think are interesting?

B (22:25)

Go ahead. Cognition. Right. Intelligence.

A (22:27)

Cognition. Of course.

B (22:27)

Right. Yeah.

A (22:28)

And, and that could take several forms because there's like some spatial rotation, there's musical ability. Right. Also I think sight hearing.

B (22:36)

Yeah.

A (22:36)

You know.

B (22:37)

Yeah, Yeah. I mean I, I recently discovered, I mean over the past couple years I've got mild to moderate hearing loss from listening to music too loud as a kid and also shooting guns. And that leads to dementia.

A (22:49)

You're a gun guy. I would never have thought that, honestly.

B (22:51)

Yeah. Yeah. I mean we grew up with guns. Like it was like an omnipresent thing in our lives. Hi. We just, everyone had guns and we, we were all shooting them all the time. But yeah, so we, Yeah, I basically I have, I discovered this. I didn't realize I had any deficiencies until I got the test. And so now I'm going to get a hearing aid because there's evidence that it can lead to dementia. I'm guessing most Americans, most people in the world probably have hearing loss. Our modern day world was not built for our ears like concerts, like 120 decibels.

A (23:19)

Yeah.

B (23:19)

Damage starts happening over about 90. So we just, we have way too many loud noises, ambulances, et cetera. So. But yeah, hearing gene therapy. We've been wanting to find a gene therapy because I would love to be able to fix my hearing. There's a few groups who've tried it, who've been playing around, but there's nothing even close. So that's what I'm saying, this is so promising is once you, if we actually can figure out how to efficiently do gene therapies and do them in a state that allows rapid iteration and we can try these things and we can pass some safety threshold, then I think it really could be this, like this new era of how we think about ourselves entirely.

A (23:53)

That's right. Because I think gene therapy, what's interesting about it is it blurs the difference between, I mean, biochemistry and genetics are very closely related, as you know. You know, because you'll have a compound and then there's like, like some small molecule and it will trigger, like a protein will bind to it and that protein will then bind to the genome and unlock mRNA, that'll do other things. So. But biochemistry and genetics are, are, are tightly related. And one of the things I've observed is that many of these things, people are fine with solving them at the biochemical level. Like, for example, caffeine does make you smarter, right? Or, you know, people are okay with chemistry to dye their hair blonde, for example. They're like, oh my God, you can't solve at the genetic level. Or actually another example is they're okay with solving at the surgical level. So, for example, height, it's okay to take hgh, which is a biochemical solution, and it's okay. People do a very painful thing of limb lengthening, which is a surgical solution, but the genetic solution, oh my God. Right? And because they just have this weird, you know, hangups, they don't actually understand the distinction or the relationship between biochemistry and genetics and how close they are. And you can do that for all kinds of traits. Surgical and biochemical are okay. And genetic is. Oh my God. Right. But gene therapy is actually something that's more like biochemistry because it's an injection or whether it's an injection or something Similar to that depends on the modality of it. But it has, it's like, it feels to people like a drug because it's on somebody who's already living, but it changes them in an upgraded way. Right. And it's actually been started to work for people who had sickle cell and other things. We can put that on screen. Right. So, okay, so there's gene therapy. Right? And then let's talk about cell therapy. What should we look at?

B (25:41)

I mean. Yeah, I mean, there's a. We've been trying to get access to cell therapies for years and it's never been possible because they're just too high risk. So we're. There's some attempts at doing this with mesenchymal stem cell therapies. You know, stem cell therapies. We have not yet seen a lot of success with these. We don't see a lot of good evidence, you know, as, as one approach, whether it's from your body or someone. Um, but there's some of the more advanced things to, to change more sophisticated biopathways in the body that just don't have good pathways. Like, it's, it's a very hard drug to get approved and there hasn't been good, good ones to like. It's. It's a very hard development path.

A (26:21)

You know, an interesting question that occurs to me is depending on what age these treatments would be effective. For example, let's say you've got a car and you, you hit the brakes and you hit them hard enough and it doesn't crash into the wall, right? There's a, there's a point of no return by which if you hit the brakes, it doesn't matter. You're still smashing into the wall. Right. But as just a thought experiment, you might actually from the ranks of older people or people who for whatever reason are pursuing euthanasia, especially the people who are pursuing euthanasia in places like Canada or whatever, right. You might say to them, why not try one of these treatments? Because if you're, if you're going to.

B (27:10)

Die, right, like, like, can't, like, can't. Like car t. Therapies for cancer.

A (27:14)

Exactly. That's right. Just like, you know, for example, with this, a novel argument that strikes me that's similar to the Act up argument that was made in the late 80s, early 90s that finally liberalized the FDA after many decades. Basically a lot of people who had HIV said, we're going to die, so why don't you approve AZT and these other things? So at Least we've got a fighting shot. You know, Dallas Buyers Club documented that.

B (27:40)

Right.

A (27:42)

So it strikes me that a potential extension of that would be given that there's millions and millions of senior citizens out there, right. Those who were like, you know what, might as well give it a shot towards the end. Now the tricky part about that is in that car analogy, they might be close enough to the wall that even if the treatment worked, it couldn't reverse it. But maybe it could. I don't know, it's case by case kind of thing to figure out. You know, maybe it's a powerful enough thing that just rewinds it. One of the things I know you know this, but like that's always struck me and many others who have written about this is a relatively older couple can have a child that is a newborn that, you know, clearly there's something regenerative in our, in our bodies that can birth again. Right. Like, and so maybe that can be triggered even much later in life than we think.

B (28:36)

Yeah, this is like Yamaneka factors. Take an adult stem cell, you go back to a pluripotent, so you make a 70 year old liver young again. And of course like now people are. Go ahead.

A (28:48)

Yeah. So, so with your biomarker thing, you could gauge someone's life expectancy and if they've got only five years to live, give them what you know, obviously they have to opt in, sign up all the stuff. They've only got a few years to live.

B (29:01)

Try it.

A (29:01)

What's the loss?

B (29:03)

Exactly. I mean that's the thing is the, the exciting thing is the technology is here. Like we actually know we can turn an adult stem adult cell into a pluripotent cell. Like legit age reversal. Of course now there's complications with, you know, can you avoid it being cancerous?

A (29:20)

Cancerous, yes, of course. Right.

B (29:21)

Can you avoid, you know, the other off target outcomes? So like there's things we need to sort through. But the, the cool thing is in 2025 we now know we can legitimately reverse age in a stunning way. And this is why I think your sentiment is correct. Biology where were too damn slow. Like, I don't, I don't know why we are not pursuing longevity like we're pursuing artificial general intelligence. You know, when I look at 2025 and I try to zoom out to the perspective of the year 2500, either you're, you're working on AGI or you're building don't die. Like everything ultimately builds up to like your existence or your shared Existence with AGI. And so I just don't know why it doesn't have a similar level of fever pitch build around it.

A (30:10)

I think a big part of it is the stuff that I think about a lot, which is the governance part, because you need risk tolerant jurisdictions since. And that's a whole change in your way of thinking. It's like, as we've talked before, the traditional financial system assumes you have some degree of inflation every year and you lose some of your health. 1, 2% inflation. Lose some of your health every year. And the traditional medical system assumes you lose some of your health every year. So you lose some of your wealth and you lose some of your health. And so it's a paradigmatic change. Even if crypto has bar charts and graphs that look like the traditional financial system, its moral premises are fundamentally different. Even if you or I will publish graphs that look like traditional biomedical papers, the moral premises are totally different because it says maybe we don't have to die. Right.

B (31:02)

Yeah. That was.

A (31:03)

Go ahead.

B (31:03)

That was our first phone call where like, I think you called me. Hey, Brian, I think we have the same philosophy. I reject inflation and you reject death. Yes, they gradual. That's like we, we're after the same concepts.

A (31:16)

That's right. Because actually it's. And the thing is that wealth in a sense is accumulated life force. Right? Because you're spending hours of your conscious, highly productive time on accumulating wealth. And so it's diluted from you. It's like you're spending. They're taking away your life. You know, there's this movie called In Time by, by Andrew Nichols that actually makes that metaphor. It's like, you know, sci fi drama, but the concept is like the currency is hours, right? Like hours of your life or whatever.

B (31:40)

Right? Yeah, yeah. I guess as a side tangent on mythology, I just finished this book, Passion the Western Mind by Richard Tarnas. And he, he tries to go through the major epochs of ideology. And he starts with Plato and Aristotle, going through the Renaissance, going through Christianity, medieval times, Renaissance, Enlightenment, modern day scientific era. And if you look at it from that perspective of like a Ray Dalio principles perspective where you zoom out, you say, hey, there's these like these big economic cycles, reserve currency that they go through the same stuff every time. It's. We're really due for a major ideological shift right now. And I think you're on this. I'm on this. That it's not some slight tweaks to the, to the continuum. Slight. It is like a wholesale change on scale with other major ideological shifts. And I think what we're talking about today is basically a part of that that is like this unleashed want for health. Yes, right. We don't. This is not about living forever. It's not about transhumanism. It's about we all appreciate waking up in the morning feeling good. We don't like aches and pains. We don't want to be diseased. We don't want to see our loved ones die. It's about feeling great and being great and that that ideology is not part of our zeitgeist. Like we just accept this slow decay, death and decline and, and it's like almost something where you celebrate. It's like I'm living life by slowly killing myself.

A (33:03)

Yeah, well, human self improvement is, I think, the way that I think about it, where that's a continuum from simply eating right and doing better to hitting your genetic limits and then surpassing them. Right. And you know, towards that end, by the way, I think the influence of, let's call it dharmic and sinic thought on the world. So, you know, for centuries, India and China, like Marco Polo sought out India, that's. Marco Polo sought out China and Columbus sought out India. That's actually how the Americas were discovered, why Native Americans are called Indians, because India was a large enough economic power, it wasn't unified, but as it, it was worth sailing to and, and for Columbus to do it. And so for, for the last few hundred years, India and China have basically been through their equivalent of the Dark Ages. And now they're coming back and those two schools of thought have a different view on the human body and they're less. You know, the Abrahamic school of thought is, has kind of weird hangups in certain ways that the Chinese and Indian schools of thought, or more generally the Sinak and dharmic schools of thought don't. And so, like, for example, something I've thought a lot about, I may have mentioned this to you, the concept of genomic resurrection or reincarnation. Did we talk about that?

B (34:18)

We did, yeah. Yeah, yeah. But explain it.

A (34:21)

You could get your DNA sequenced and stored on, on disc, and we already can do chromosome synthesis for eukaryotes. Like, you know, we could do it for. Actually you could synthesize an entire prokaryotic chromosome from disc and actually have that thing swim around in a test tube. And as our abilities of chromosome synthesis get better, you could imagine just projecting out to be able to do it for full complex human chromosomes. And it's just it's like a technological direction. It's kind of like if sequencing got better, synthesis will probably get better. And it might take, you know, a few decades, but it'll get there. Which means that if you have a file, you could, if somebody was well behaved, you could write their genome to a blockchain or something like that. And if they had good karma, then the community would reincarnate them 50 or 100 years hence or whenever chromosome synthesis becomes feasible, because they have the data there. And you know, people argue the genome isn't everything, but it's a lot. It'll get you a lot. And then you could replay all their past life experiences for them if you recorded it. And they would just be born with all this amazing knowledge of everything that happened. You know, and that's something where if someone was going in on one of these missions where they're trying out this experimental drug, you could sequence their genome and say, if it doesn't work out, you'll be reincarnated with the knowledge that you had in life and so on and so forth. And we'll kind of, we'll look after you in the next, in the next life. Right. And so that's like a dharmic inspired way of thinking about it. And then the psych inspiration, among other things, is just extreme pragmatism in just experimenting with things. This is actually, you know, the western mindset used to also be like this bantigan Best. Do you know that story that don't so banting and best this is an important story for you. You know, like 19, if I get the dates right, I think 1921. They started experimentation on insulin supplementation because they had a hypothesis it could treat diabetes. And they tested it on first like dogs then, and it worked. And then they treated, you know, they tried self experimentation and they went for patient volunteers and those patients just like stood up in bed like this. It was like the canonical bench to bedside concept, like a bubbling beaker. You know, it's brought there and they iterated on formulation and so on. And I think, you know, right now it's so hard to go from like pill to oral to an injection to a patch. That's a whole process. Whereas in some ways it's kind of like going from web to a mobile client to like a command line client. It's kind of the same drug but in different formats or whatever. So they just did all of that. It wasn't case control studies. It wasn't. It was just iteration, quick iteration and seeing it. And it was A large. And by 1923 they had Eli Lilly doing scale production of insulin for the entire North American continent and they'd won the Nobel Prize. And so that was a time when pharma moved at the speed of software, where you went from idea to execution and via iteration in like two years. And one of the issues I think a lot about is there's no case control studies on case control studies, there's no regulatory science on the regulator themselves. Why don't we have a jurisdiction where you can just iterate your way to, you know, to something where crucially, by the way, the patients aren't simply patients are participants in their own health. They're not like a row in a table, like a very 20th century study like Framingham or something like that. That's just like a row in a table. Right. Whereas if it's a human being there and they've got a mobile app and they can send you back what their experience of the treatment is and they're in active participating in their own health, it's a totally different paradigm. It's bidirectional. And the very simplest version of that is adaptive clinical trials which show you can converge on an endpoint faster if you know you're not simply doing a really dumb open loop trial. Right. Where you're not, you're basically taking information from the trial as it's going to adjust things very shortly. But an adaptive clinical trial, but I think you can do much better than that. No software is designed with case control studies. SpaceX wasn't designed with case control studies. I'm not saying that there isn't some utility to them, but they're really for teasing apart small effect sizes rather than large effect sizes that are so manifest that they just jump out at you.

B (38:36)

Yeah, this probably goes back to the beginning of our conversation where if you have a large visual effect size, that gets everyone involved.

A (38:44)

Exactly.

B (38:44)

And if you can, if you can do so in a jurisdiction to be like, hey, you can go from idea, you know, in gene therapy, cell therapy or something else. And I guess that's why people are so interested in peptides is peptides are drugs.

A (38:56)

Right.

B (38:57)

Peptides give you drug like effect sizes. The problem is they're poorly characterized. There's questions about how they're manufactured and the quality of those things. And then also you have a lot of off target effects. So I mean, peptides are inherently a risky path when you're not doing a well characterized path. But yeah, I mean, this is if we could demonstrate some like one or two Wins of a process from a talented drug developer to a actual thing that works, that has a low safety profile, large effect. Now I guess it's tbd. Whether we can see that because I mean biology is complicated. Maybe that's too big of an ask. Maybe you know, the GLP1s, they kind of hit that sweet spot where it's the immediate visual effect of fat loss and it has all these follow on health benefits where even if you're not obese or if you're not having an eating problem, they also.

A (39:46)

It gives you power, gives you discipline.

B (39:49)

Yeah, exactly. Yeah, yeah, yeah. That reminds me of like in the future like we talk about like you and I have imagined what gene therapy is. We say hair, skin, you know, muscles, et cetera. But it might be fun to a thought experiment of like what would gene therapy be like in 30 years? Like what kind of nuanced, expansive concepts would we think about? Like we think about like designer babies of like we're going to choose height and eye color. But like how would you think about human experience?

A (40:16)

Mini circle is interesting because they have the concept of a plasmid that's reversible. You can turn it on, turn it off.

B (40:21)

Right.

A (40:21)

So almost like you know, I have some caffeine and I don't. Right. You could or you change your, you know, put on your contact lenses or not. You know, biology is complicated. There's all kinds of off target stuff. It's not trivial to get it right. People's biology is different. All that's true. But there's something cool to that. You have a very specific target that can flip on and off a protein that can add or delete a function. Right. And Patrick Friedman I think had that therapy and he showed his. This is a mini circle, you know which is doing the plasmid. So Patrick Friedman and Farb Nevi both had the therapy and Patry said that he was just crushing it with his VO2 max. Farb also said the same thing. Right. And that's something where it's n of 1 but they know their own body better than anybody else. It's N of 1 but it's over. It's. It's a lot of variables which n of 1. You know Mike Snyder's thing on this from many years ago the talk about Mike Snyder is a profit Stanford and he. Yeah, the integrom. Right. He just took every possible thing from expression to the rest and just ran on himself. And he's able to for example see himself getting sick in the gene expression Data like a few days before he actually got sick.

B (41:37)

Right, exactly.

A (41:38)

And so we could, with better metrics. That's another piece of this actually. Diagnostics are non invasive, right?

B (41:46)

Yeah.

A (41:47)

So if you had constant like whole, you know, like, like all 30,000 odd or 22,500, whatever, you know, a human gene expression panel and you track that and maybe you know, metabolomics and so on, you did something like the Mike Snider integrom. While you're giving people this gene therapy or cell therapy, you just got a much richer readout, you know, from them. Right now you're not just relying on self report, you know, it is, it's like your biomarker stuff obviously, but it's just like, you know, we're getting tens of thousands of them. Yeah, that seems like, you know, something that is very feasible because costs are coming down of all that sequencing and so on. That'd be a very good thing to do. Like an integral like thing that's like a time series integral. First you get whatever data points on them to determine healthy, then you impose a stimulus and then you look in state space and you see is it actually changing anything or not? And what's it changing?

B (42:36)

Yeah, yeah, that's exactly what I'm doing at Blueprint. So we are going to try to do for. I guess we've done that for myself because like when you build an asic, you know, if you ever haven't, if you have not built an asic, it's really cool that if you like get a picture and zoom in, zoom in, zoom in, zoom in, zoom in, you find that it's like down to the nano level of granularity. It's beautiful. It's amazing. We can design these structures.

A (43:02)

It's crazy.

B (43:02)

Yes, it's crazy. And the thing is you can characterize this at every level, every circuit design you can characterize, you can say, what happens if I do this? And so we can characterize our machines and our intelligent machines so well from the circuit all the way up through the application layer, we can characterize the entire stack, the surface they run on the energy. Whereas in human biology we can't characterize. You can get this really high level of what does my blood draw say, my cholesterol or my liver enzymes. But you can't get down and characterize these smaller molecular interactions which like you're saying they read out really important data. Am I getting sick? Am I showing signs of disease? Am I responsive to this drug or not? And in what ways? And so like if we can pair the characterization with these jurisdictions with good drug developers. Like that's your path. I mean, and then we can open up and once people get like a, a few successes we'll say huh, like this is a thing we can, we can actually do it. I was talking to Jason Kelly the other day. He's founder of Ginkgo Bioworks. Yeah, I was, I was first. Yeah, exactly. I was first money in and they, they really led the charge for biotech in industrial application, industrial applications. It was a really great moment. They've run into some roadblocks and doing it which like, you know, breaks my heart because there's nothing I want to see more than biotech in the industrial world, like really smart biotech designs. And we were saying the other day like we, because we're talking about doing something together and we're both just like why can't we build human with the same fever pitch we build tech, like why, like why can't we chase this thing? There's just, it's, it's so frustrating that.

A (44:45)

I think it boils down to risk tolerance. And the reason is with tech we can have crashes because when it crashes and the computer crashes and it crashes all the time, it's like acceptable to fail. Right. With humans people get really mad if there's a failure. Right? They are, you know, but, but they didn't used to. With banting and best in that era people were more reward seeking and less risk averse.

B (45:11)

Right.

A (45:12)

And I think we're kind of coming back to that era gradually where risk tolerance has radically increased over society over the last few decades. I think. And you know, one thing on this is an intermediate between though maybe we can just go to the full thing. I think just the diagnostic zone, right. Where you can just get, get genome sequence without a prescription from your doctor. It's so stupid. The entire fda, you know, CDRH thing where it's like requiring a prescription to look at the mirror. Like why should you need a why, why should that be bottlenecked through the medical system to get, to get your genome sequence?

B (45:54)

When, when the AI models are smarter than my doctor, like why do I have to go ask them?

A (45:58)

Exactly. That's right. So I think there is definitely room for something where you get your genome sequenced by for example Nucleus or some at home kind of device, you know, like as Oxford nanopore or something like that gets smaller and smaller. Eventually you'll get there or you just do it in a lab and you just get the file back and you have it interpreted by Something like an open source version of Prometheus and then. But. But it's an AI enabled thing and it's all done locally. And so I've got some companies that I'm looking at that do some combination of these things. But you know, I think the second generation of truly personal genomics could be another big thing that we do. And maybe that's the intermediate step that then people have enough data and enough time series on themselves that then that opens up basically. The reason that's interesting is now you're not flying blind. Every treatment, you have a sense of your own dashboard and your own metrics. Like you're one of the very first on this. Actually. There's that guy, the measured man. There's the article in the Atlantic many years ago on this guy. He's a Prof. I forget his name. You should maybe meet him if you haven't. This is back when the Atlantic was good. But the. So the, this guy also was measuring all kinds of time points on himself and he, and he found he was like an early quantified self kind of person. Right. So if you have that dashboard now, you have a sense of kind of what's the speed bump, what's a real major thing. And then you're kind of better able to process what an intervention is doing. Is it messing you up? Is it just moving these genes, that kind of thing, like these expression levels, et cetera? Do you actually, in your, all your biomarker stuff, have you done gene expression time series on yourself?

B (47:38)

By the way, his name is Larry Smarr.

A (47:41)

That's him. Yeah, that's the guy.

B (47:42)

I've never met him. When was this done? Do you know him?

A (47:48)

I don't know him. I just know him from the. I think Larry Smarr, Mike Snyder. And you should do, if you want, we should all do a podcast together or something because, you know.

B (47:57)

But that's a good idea.

A (47:58)

Those, those two people are very aligned with our way of thinking.

B (48:02)

Yeah. So he's a founder of the national center for supercomputing applications at NASA, born in 1948. So he's just.

A (48:09)

Yeah, so he's 80 or something now. Maybe. I'm not sure what health he's in, but him and Mike Snyder are both pioneers in this space. Mike Snyder with the Integrom. Larry Smar. Right. And I think it's worth, I bet, by the way, if you put that into AI. AI is very good at expanding suggestion lists. So if you put both of those there and say, who else is like that? You'll come up, but maybe others will come up too.

B (48:30)

Yeah. He found early Crohn's disease in himself before his clinically diagnosed. I mean, this is like. So I have colonel on my desk here. This is, this is the brain interface I built. I measure my brain every day now. Ah, I'll show you. Yeah, so this is like characterization the brain. Like, you can start. You can see the inside here maybe. Yeah. So this is like we did. We, we spent seven years building this technology. It's basically wearable FMRI using light. Using light. And this is the problem is we've never had the ability to characterize our brains. It's like if you want a brain scan, you do FMRI or you can do mri. FMRI is a functional side. MRI is a structural side. But just to go to facility, sit in there in an environment that's claustrophobic. But now I can basically characterize my brain every day looking at the functional networks.

A (49:16)

You know, I have a, my intuition, by the way, does that need to be. Can you walk around with that or does it have to be plugged into the wall?

B (49:24)

Yeah, yeah, it's stationary.

A (49:26)

It's stationary. You know, a fun test would be what does it look like when someone is just scrolling versus what does it look like when they're on the treadmill?

B (49:37)

Yes, exactly. And like, I didn't do this, but my, my colleague did. He drank alcohol for science and he saw an immediate aging of his brain. I think it was two to three years the following day. We. It's been cool to just see basic correlations. What happens when you're sleep deprived, what happens when you exercise, what happens. You eat well, like. But you can run all these experiments now where I couldn't before. But this is like the characterization thing if this, the characterization thing is. What we've been missing is people run to do therapies, and even when they do therapies, it's poor care, poorly characterized. It's very characterized, let alone before, like having some, like you're saying, some time series longitudinal data looking at these things, which is, I've, you know, I've become the most characterized person human in history. There's no human that has. Is more characterized than myself. And so we have. It's enabled us to find so many insights that we just haven't been able to find in the literature or through other observations through doctors. It's just like it's, it's all about measurement.

A (50:35)

And what is, what is the cost per day of your characterization, by the way? Like what, what assays are you running?

B (50:41)

Honestly, it's not that expensive. These are like, I mean, blood draws, saliva, stool, mitochondria, fitness tests, imaging, methylation, proteomics. You're like, I don't, she's like, not honestly, probably, I'm guessing without thinking about this more robustly, but something like fifty thousand, a hundred thousand a year, like not, it's not that much.

A (51:08)

It's like a thousand a day. And so if we could bring that down to 10 bucks a day or even a thousand a month.

B (51:19)

Yes, exactly.

A (51:19)

Right. Actually it's less than a thousand a day. If you're saying 100. If you're saying 100k a year.

B (51:24)

Right.

A (51:25)

Then it's somewhere between two. It's like 200, 300 a day. Right. It's already not too bad for what it is. Right. So it's like, you know, the thousand dollar genome. Yeah, there's, there's obviously, I mean, I've run a clinical lab. There's definitely returns on scale with clinical lab for all kinds of stuff with instrumentation to reagents to everything. So I've thought about a residential clinical lab where we just build it into network school.

B (51:53)

Yeah.

A (51:53)

And because it's a pain to go and you know you want is a phlebotomist to come to you, right. You want mobile phlebotomy, you want saliva, you want ideally even in your apartment or something like that, you can just leave your toothbrush or something like that and then it can just, you spit in a cup or something. And then there's various ways of probably streamlining if not fully automating, certainly making it easier to get samples.

B (52:22)

I have a sensor in my toilet.

A (52:24)

Yeah, exactly.

B (52:25)

That kind of automatic automatically. Yeah.

A (52:27)

Bill Gates was interested in this 20 years ago. And you could, you know, like urine colorimetry and so on and so forth. There's a lot you can, you can do with that and you just have it there and maybe, you know, you just image it and so on and so forth. I think the Chinese could do a lot here because they can innovate on hardware and there's something to be done here because they're building a lot of cities so they could actually install these smart toilets, some kind of Chinese, Japanese combination. Because the Japanese also do a lot of toilets.

B (52:57)

Yeah, I just watched that south park episode where they got the Japanese toilet. Anyways, yeah, the characterization thing is really cool because when I started Blueprint, I was trying to solve my own problem, like how do I actually find nutritious food? That's third party tested and low on toxins. And then friends and family are like, hey, can I have access? I'm like, sure, here you go. And it just like accidentally became a business. But the problem with that is like even though I'm trying to solve a real problem in life, people have such a negative connotation towards therapies, like towards things you actually like. It's complicated and a characterization is much more friendly. Right. If you're helping someone measure themselves, there's much less blowback. And so I'm moving very hard in that direction where I'm going to help people characterize themselves not drugs and devices. Yeah, like just characterize yourself like me. Like let's just like get this baseline measurement. It's a much safer path to build because then people somehow they don't have suspicions like they do. I mean now when I'm like really? I'm like, I don't, I don't want to be doing this. Like there's so many, like if I were to out to make money, there's like a thousand other ways I want to do, to make money. So anyway, yeah, so the thing is.

A (54:06)

Actually, you know, years ago, that's why, you know, my first company is in diagnostics. Because also diagnostics has a lot of advantages. First is it's half computer science because after you do the assay then you can analyze everything on the computer.

B (54:19)

Right.

A (54:20)

And you can make all the dashboards and so on. So it's like at least half within our bailiwick. Yeah, it can't hurt the person usually. So the risk is way, way, way lower. So you can move much faster. You know, the worst thing you can do is you can give them a wrong result, but you can move close to speed of computer science. And then most interestingly it means that the intervention is informed. So arguably diagnostics creates the room for therapeutics. So perhaps. Yeah, we start with the quote, you know, the universal longitudinal diagnostic network state. First longitudinal longitudinal network state, then longevity network state.

B (54:58)

Okay, cool. Baldi, I like this. So basically we get a cohort of people together, we say we are going to invest the capital to do high fidelity characterization.

A (55:09)

I love this.

B (55:10)

We don't know why. Right. We don't know to what end.

A (55:13)

Alexander Integrom for 100 or a thousand people.

B (55:16)

Yeah. Now it, and it may not be useful to be this characterized for 5 years, 10 years. Like we don't know. But like we want this longitudinal characterization so that when we find things to, to trial we have this time series longitudinal data that we can, yeah, that gives us a better shot identifying what therapies are going to, what we will be responsive to, what we won't be and then see the off target effects and so probably a much better way. Whereas right now even when you look at the clinical trials, they are so poorly characterized.

A (55:47)

It's like I read this one scalar variable on a very complicated human.

B (55:52)

I read these studies, I'm just like this is not reliable data. Like when you're telling me it does blank and blank and it has these side effects. Like sure. Like nothing close to what I would think it'd be. So even this, like that's why like the GLP1s are a good one because like they have done good characterization better than, than, than like you know, a, a non characterized peptide. But that said like we don't know a lot which is why like all this stuff is coming out like blindness and stuff like that. So yeah, we really, I guess like this conversation has landed us on characterization as the key input. But then we need to go to these jurisdictions and figure out with that how we'd leverage fast track safe high effect size therapies. Because we need a moment, we need a moment in for longevity. We need something.

A (56:36)

Yeah, go, go say what you're saying.

B (56:38)

Yeah, no, just like it could be the case that it's like, you know, the history book says, you know, from like 2025 to the 2035 these new therapies came out and I had this like starting this cumulative effect or there's like a moment where something happens. Maybe the GLP1s or that story of like people familiar with doing these.

A (56:57)

They're important. They are. But, but, but, but we could have the Ozempic for aging would be that moment.

B (57:02)

Yeah, like something, something pretty. But the effect size is big enough where everybody turns their head and they're like I understand reality differently now because of that.

A (57:12)

Yes, that's right. And I think, I do think that's possible simply because we're seeing it in mice and so on. We just got to figure out what buttons to press within humans. But I really like the idea of the longitudinal network state and then the longevity network state or the longitudinal special economic zone, longitudinal diagnostics and then, and we can cost that out and we could do something with it as a DCI thing, you know? Dci.

B (57:35)

Yeah, yeah, we could.

A (57:37)

Or we could do it at a traditional raise. But it's just a very, very scoped thing which is these end people for this period of time and frankly they could even Pay for it where it's something where it's not that expensive. It's just expensive enough to be within like a, you know, sort of pro zoomer budget. Like a crypto person. Right?

B (57:59)

Yeah.

A (57:59)

And, and they could maybe live at network school and be part of the study, you know. Right.

B (58:04)

I like this.

A (58:05)

So you have the list of all the instruments that you do for your measurements. Do you have that in a public spreadsheet or do you have that internally?

B (58:12)

I need to put it together. Yeah, basically. Yeah. What would a high fidelity characterization look like? What tests would you need? What machines would you need? What labs would you need? Because some of the assays we want done are just not available off the shelf labs. So even if we had some people who could spit up some nice. So it could accommodate some of the unique things we want to measure, that'd be great.

A (58:35)

Yeah, I think. And we, we do it in participation with NUS or somebody like a, you know, some, some sort of local university or hospital. And I think we could do something here. I think there's something very, very cool.

B (58:48)

So if anything, it's like a new model of even. It's a new model of healthcare. Really.

A (58:52)

Like the key is, you know, like at Stanford there was this concept called back again when Sanford, back when San Francis was good, residential education. And the idea was that the education didn't stop when you were. It wasn't one hour in class, you were 23 hours outside of class. And so education happened. The dorm rooms and sort of, sort of. So it was like, you know, immersive in that sense. So this is residential health?

B (59:18)

Yeah, yeah. I mean, this is kind of like we, we, you and I have been spinning around this idea, I mean, since our, since we first met, of like how we bridge the worlds of ideology, health, governance, nation, state. Like what is this moment? And I like this as a path because it does. It brings together the things people care about.

A (59:40)

Yeah. And there won't be that much resistance because we could start with the longitudinal measurements. No one gets hurt, Right, Exactly. Then you have enough data to show what normal looks like, what interventions look like, what fitness looks like, what bad foods look like. You've got all their nutrigenomics, you got their pharmacogenomics. It's a whole thing. There's a whole stack of actually open source software that you develop in doing this. Yeah, I like this a lot.

B (60:05)

And then we can also, if we. Okay, this is probably too much. But then we, we add some organelles. Right. So now you start doing. Try therapy trials against the organelles instead of. So then you've got organoids. Exactly. Yeah, exactly. So now you start doing the. The experimentation. High throughput experimentation on your organelles. Yep.

A (60:24)

And they're proxies.

B (60:26)

Exactly. Yeah. That'd be bad. I mean, that'd be amazing if we actually set this up. All right. Yeah. Okay.

A (60:31)

All right.

B (60:32)

Good outcome.

A (60:32)

All right, so we got a plan for the longitudinal network state towards the longevity network state. Making progress. Okay, well, thank you very much, Brian.

B (60:40)

Thanks for having me. I really enjoyed this.

A (60:42)

Awesome. Talk soon. Okay, bye.