Dave Messina, Pioneer Fund: YC, the Human Genome Project, Health AI and Pioneer.vc

Welcome to another founder wisdom podcast wisdom podcast because we have both on the call today. Dave uh Messina is a J GP at Pioneer Fund the future of health. Uh you can find him out at health.pioneerfund.bc.

Dave's also a a founder and has a fascinating uh background work on the human genome project back then. Uh his company was called the co-actor. Uh I think he left that like a year ago.

Dave, but just a fascinating human being that I guess wants to make an impact out there, right? Uh we have a lot of health problems nowadays and [snorts] AI just came around and we're seeing the possibilities of what it can do in that domain. So, I'm just glad to have you on.

Can you tell us a bit more about yourself and what you've been up to? >> Hey, Charles, great to be here. Uh excited to talk about uh both AI and healthcare.

Uh what we're doing over the future of health fund. Um, so a little bit about me, like you said, I got my start working on the human genome project. This was at Washington University in St.

Louis, which is one of the two institutions leading the the public effort to do that. Super exciting time. Um really fascinating to see something like that, the biology happening on a large scale where when they started the project, they knew that they would have to create new technology in order to be able to finish it and they did and it was really an amazing time.

I then uh went and did a PhD in computational biology over in Stockholm, Sweden and then returned back to the US to start my company co-actor genomics with some of my human genome project buddies. There we were uh using machine learning uh sort of you know pre-ai to uh help doctors choose the right treatment for their patients. We did a national clinical trial for that which I led and we received national Medicare reimbursement for for that test which is being used now nationally in the US.

Um I've been part of Pioneer on the investing side uh pretty much since Pioneer started in 2018. I've been leading health and bio investments over there since 2019 and happy to talk more about the Pioneer journey and sort of how we got to building a health focused fund. Yeah, before we talk about that, we need to talk a bit more about the human genome project.

How complex was it? Was it a moonshot? And why was it so groundbreaking for humanity?

>> Great question. So, it's kind of crazy to see now 20 years on, but at the time it was actually somewhat controversial. So, one, because of the cost, people were like, "Hey, we're going to spend $3 billion on this.

What does it all mean? Shouldn't we be spending this on other types of medical research?" And the conviction was that prior to the human genome project, and I did some some research prior to the human genome project trying to research a disease. And prior to the human genome project, we would kind of have to guess sort of we where genes related to a disease were because we had some information.

We were able to sort of map it down sort of vaguely to a region of the chromosome. The first paper I ever was involved in in publishing on medical research, we were thrilled to get to narrow down this this disease, the the causitive gene for the disease to a million letters of DNA, which is a huge space with lots and lots of genes in it. So the idea of the human genome project is we can have the entire map.

We have the total complete set. uh we know all the we will know all the genes that make a human and having that complete set will give us far more information about how humans work and that's exactly I think what what has come to pass where now that we have sort of the complete parts list for making a human we've been able to understand more and more deeply how genetics makes uh makes a human and how that affects disease >> and what new technology was uh created did in that process. >> So there were a few new technologies.

So one was how to map uh the big pieces of DNA that were being sequenced onto the chromosome. So we've known about chromosomes for a long time, the sort of the the 23 pairs of of big pieces of DNA that that um we have in a human. Um but uh you can't just take one of those and sort of grind it up and and put it into a sequencing machine.

um we had to figure out how to sequence big chunks of that and actually take those big chunks and make them into even smaller pieces to put them on a machine and how to build software techniques that would allow us to kind of put those successive layers of jigsaw puzzles together. So that was super amazing. That's kind of an unsung piece of the story because the uh the competing sort of private effort to sequence the human genome was taking a uh a completely different approach which uh it really relied on that that heavy mapping process um to be successful.

The other piece of course is obvious uh that for was the sequencing itself. So being able to do DNA sequencing at least on a semi semi-automated fashion. Today it's crazy.

You know, the first human genome cost $3 billion to to sequence. It was a huge international effort. It took many many years.

Um today we can sequence a genome for a 100 bucks in an afternoon. >> And is that because of Mo's law? So actually the rate of DNA sequencing has exceeded Moore's law for the last 20 years.

So the Moors law the you know is sort of the the doubling of the density of transistors on a microchip. Uh doubling every 18 months it was for quite a while. I think that's still true or maybe more given the GPU era we're in now.

Um but DNA sequencing the the amount of DNA sequencing that you can get per dollar has more than doubled. um o over the last 20 years there was sort of a plateau for a while but some new technology from Ultimate Genomics um has actually pushed that further and aluminina has also got a a more dense machine now too >> and back then you were one of the only lab you and your mentor I believe to use technology to uh improve research in health like what was that you were 5% of labs doing that >> yeah that's sort of correct so there were not that many people using computational approaches to study medicine or or DNA and RNA at that time. When I went looking for a uh uh a PhD program, I wanted to go where there were computational biologists that I could learn from.

And there was really like one place that I could find, maybe two in the entire world that had more than maybe a couple people at that time. And that was the at Washington St. Louis and they had all gathered there.

I think in part to do the the human genome project and so now it's much different. Now we have thousands and thousands of people who use computers to study biology and medicine. It's completely different era but back then it was um relatively new.

>> And I guess we already have one lesson for founders, right? Which is that you need to once you have an insight and you're sure about it, you need to grasp onto it even if you're being seen as a weirdo, right? Then even if it's controversial and everyone tells you like why are you spending your time on computers >> there are some moments in life I think that's true there are some moments in life when you can see clearly that something is true maybe before other people do and at that time I could see clearly that genomics was going to have a huge impact on the world and that computational approaches to understanding medicine was going to have a huge impact on the world.

I didn't know exactly how, but I knew that that was a direction I wanted to set off in. And and so that's what we did sort of with that general direction and that ended up being a very good decision. Um but certainly I had mentors at the time who told me I mean I remember like them sitting me down and having meetings with me and saying like this is a mistake.

You need to do conventional biological research. um this is not going to go anywhere. You won't be able to to um not only do any interesting research, but you won't be able to make a living because because no one will hire you.

And and um and I'm glad I didn't listen to them. >> The important equation is once you know that something is right, it's the when, right? Because let's say in your space that something's going to catch up eventually, but in 10 years, how can you manage to have sufficient runway to last out until that 10 years, right?

So, how do you actually compute the when when you invest in a founder? You know that they're right. Their contrarian thesis is good, but how do you actually compute when this will catch up and so that they have enough runway to survive till then and make it big?

>> Yeah. Yeah. So, so how to think about the timing in general?

I think it's almost >> Yeah, it's in general, I'd say, it's not a problem that most people have to worry about. Most of the time, if you can see something clearly and and and moving forward as fast as possible, as soon as possible is the right answer, that is of course not true. We certainly know that you have to think about that you know the sort of the funding time scale and so some of some things which are which were more like basic research that require additional new knowledge can fit into this category and so there are other mechanisms that are maybe better for this and so a good litmus test for this is to think about it from a commercial point of view so yes the thing the technology that I'm creating is interesting, but is there a version of it that someone is so valuable to somebody now or in the near future that they're willing to pay for it?

And again, there are exceptions to this, but in general, that is a good litmus or finding a version of it that you can sell along the way, even if you're not getting all the way to your farthest um most expansive vision. Generally people tend to think that um especially new medicines people tend to think of it as as sort of okay you know it's a risky bet because you've got to invest now and you won't find out for 15 years there's a heavy regulatory risk there and we tend to find that that is not true that there are good ways to signal that there is value much much sooner than that and we've actually seen some great outcomes much faster than that. So a great example is Gilgamesh pharmaceuticals that we invested in just 5 years ago.

That was a seedstage investment. They uh sold their lead asset to Abby, a maj major farmer company for over a billion in cash um just five years after we invested at seed. So that's that's a a a drug company return.

And the best part is that they have additional drugs in the pipeline that they are pursuing now. Yeah, it's a >> it's a fantastic direct example because these guys they started in the psychedelic niche, right? Like and are still in that lane, are they?

>> Yes. So they were some of the first to recognize the value of naturally occurring known psychedelic compounds for medical purposes. And so what they've been able to do is adapt psychedelic compounds removing the psychotropic component but preserving the therapeutic component.

And their first drug is remarkable. They showed in a phase 2 study. So these are advanced human uh studies that people who had been unresponsive to traditional treatments for depression or PTSD were able to see a significant benefit from from their drug where all other drugs had failed.

So that is incredible. >> Yeah. My question for that isn't it the actual psychedelic trip why there's so much neuroplasticity on the back end.

How could they like scientifically objectively prove that both are actually separate? For example, if we take psilocybin or [snorts] iawaska for example. >> Yeah.

So how are they separating the psychotropic piece from the therapeutic piece? That's a great question. Um that's a that's a little uh deeper in the into the sort of the the science of the the neurochem neurochemistry then I think it makes sense to go here but but they are in fact sort of connected but related mechanisms and it is possible to separate the two >> right >> so you so you don't have a you don't have a trip in other words when you take their their new medicine >> yeah I know but how did they actually discover that it seems like a contrarian thesis right because in my mind as a guy that have a ritual on a a yearly basis to do psychedelics.

The one good part that I attribute my neuroplasticity and my betterment of my mental health is actually uh because of one allocenic uh aspect of it. Um, you're saying there's other compounds, for example, MDMA, right? You don't trip, but you do get major neuroplasticity and all kinds of chemicals, serotonin, dopamine.

Um, no, it's just did they do like clinical study is my question. >> They have. Yeah.

So, these are actual they've done, you know, the phase one >> tripping and no tripping is sort of my question, right? [laughter] Like, >> oh, that that piece I'm not sure about. But I think the way that it works is that that they decouple the therapeutic piece from the duration or intensity of the psychedelic experience.

So it's the same core biology, but it's engineered in a way so that the subjective effects are shorter, milder, and more clinically manageable. >> Right. Right.

And I'm saying that this is the the perfect case to illustrate what I was asking because the the key uh cases in other niches we have Yuber, right? that hyper push with a lot of money with full force Travis as a wartime CEO uh to create market right like that that's the thing they don't have to wait until product market fit they are creating the demand same thing for self-driving cars and uh Tesla for example right like they have they gave such a push that they didn't have to to wait for for the proper time to launch actually created that which is very hard to do um and Gilgamesh for example we know that in the US there's um the war on drugs, right? That is certainly directly conflicting uh with that.

So why do you attribute their success? Is it the founder that was just like a a maverick and that could push it to market? Is it the funding that you participated in?

Like why do you uh think they were so successful so quickly? >> Great question. So the war on drugs uh actually set back psychedelic research for decades.

And there there around the time that Gilgamesh got started, there actually had been a reduction in that uh ban on psychedelic research and and in fact it had become sort of full circle and and was possible to receive uh federal research funding for psychedelic medicine again. So the Gilgamesh team were experienced um uh drug developers already and had were just early into sort of this new opening up. Um but you're right.

I think there's part of it is them seeing the the possibility of of making mainstream uh FDA approved medicines was something that most people would not have uh agreed with and I and I think part of that is their own conviction and also their skill um as scientists in being able to put those together. So, I I while I'm thrilled that we were able to support them and still support them and we're huge fans of theirs, I I can't can't claim that uh our funding or was instrumental, but I think the venture capital process worked exactly how it was supposed to in this in this way. So, it's >> great.

I suspect it's part the YCDNA, right? The YCDNA is very much like run through walls and and don't stop. Um and it also if you can tell me also about the history of YC investing in health right because it's not super obvious to choose YC as an incubator when you start a health company but I do think it creates like a very special breed of founders especially in that niche right so can you comment on that >> sure so YC has been investing in health and bio since almost the beginning I know that uh there was a a successful EHR company um uh that was started Dr.

Chrono which was started I think that was in the 2011 batch so very early. Um but in earnest YC turned its attention to bio health companies I would say around 2014. Genko was one of the first and they're now a public company.

Uh my company Kofactor Genomics went through in 2015. Notable labs was around the same time. Billion to one which just had a nearly $4 billion IPO.

uh my friend Ogjan and his company uh were I believe a 2017 uh YC company. So you're right, YC is not known as well for bio and health, but they have been steadily investing a fairly significant portion of their portfolio in in bio and health um for at least a decade I would say and today the YC bio and health community is quite strong and that's part of how pioneers health and bio community is is so strong with over 100 bio and health founders today. We can talk more about that.

But to the the key insight I think that YC has had which has been so instrumental in in uh influencing the companies that have gone through the program and even uh more broadly is that yes there are specialized pieces of knowledge that are very helpful in pursuing highly technical companies like uh biotech or or health companies or hardware companies And it's possible to provide help and get get support from other founders and peers who have that same background. But also the fundamentals of what makes a company successful and some of the key key things that any founder needs to think about are universal. And so we touched on one of these earlier.

You need to build something that people want. the even in biology and medicine, you still have a consideration of who the end user is or how it fits into say the the types of drugs that already exist or if you're on the health care side, how it helps in in the provision of care. All those the questions around, you know, who's going to buy, what's is it a really a big problem you're solving still matter.

I think also questions around the the strategy and sort of the the um initiative that it requires to build a company of universal. You know, one of the things that people think a lot about when they're starting a company is competition. Oh, I'm not sure I'm going to start a company in that space because there's so many other companies doing it.

We see this a lot now with AI certainly as as a technological u barrier to entry is lower but actually competition is typically not the thing that kills companies. Yes, you need to be aware of competition but mostly companies die for other reasons and so and and often the landscape can change too. Companies that seem formidable competition at one moment in time tend to fade away.

We saw this several times with with co-actor where we had competition um in some companies and I remember I was thinking like do we need to pay attention to those and and a few years later um those companies had faded away. So >> 100%. And yeah the moat is getting thinner with uh AI.

Hey do you have uh 15 more minutes by the way? [snorts] >> Yes. Yes I do.

>> Okay. Okay. I want to ask you about the FDA.

Um, have their stance changed as of late? Aren't they like stopping innovation mostly? Is there like some development in that field?

Because I just feel it's pretty ridiculous to spend all that time to get a drug approved that would potentially save lives. Um, so give me the news on that and also if you would be chief of the FDA, what would you actually do to make sure that consumers receive uh safe drugs? >> Sure.

So the FDA plays an important role for sure in evaluating drugs that come to market and every every company excuse me every country in the world has something like that. I think there are certainly some aspects of the FDA which are not as receptive to innovation as we would like and I can talk about those but there are other areas where actually the FDA has been more recently quite forward thinking. So for example, there's been an initiative to allow more uh computational or even AI approaches to evaluating drugs as opposed to using only animal models.

So yes, animal models using mice and so forth to test drugs before they are tested in humans can be a helpful step. But we also know that mice and humans differ enough in their biology that you can have success in mice that doesn't necessarily translate into humans. And this is one of the the big challenges in in drug development anyway.

So the FDA has actually been quite receptive to alternative approaches. One of which actually is um in the greater use of what are called human organoids. So when you can grow either uh grow human cells and test drugs in in what are sort of sim sort of organs they're they're actually living in and you know biological clumps of cells but these can be more accurate or more useful for predicting things like safety and toxicity in drugs and even to some degree um effectiveness of drugs.

So we've backed some great companies in that space like Rosebud Biosciences for example that has built a platform where they can simulate sort of the the three key uh organs of of uh the let's see if I get this right the the liver uh the um the kidney and um I want to say the heart. Yeah. So the they're they have their heart organoids actually beat just like a real heart and they're great for testing toxicity.

So I uh know that uh it's very difficult to run an organization as as big and complex as the FDA. But one thing that I'm also excited to see glimmers of in in folks who are who are in the FDA is uh a change in uh the the some of the fasttrack pathways. So the the breakthrough pathways that allow certain types of drugs to move faster on a faster timeline through we're seeing more and more of that and likewise coupling that to reimbursement better.

So you know traditionally you get FDA approval and then you you get improve uh approval for the payments for reimbursement from Medicare and private payers. And there are renewed efforts to to say, well, if you've got FDA approval, then you can get um get sort of automatic or much faster uh reimbursement approval. Um but there's still a lot of progress to be made there.

>> Are you still bullish on Medicare? Do you feel people should not pay for their health? Uh well, if we zoom out in the broadest sense, I think that one of the things I'm most excited about with the advances in technology that we see is being able to move to a scenario where uh basic health care is um is less fraught in in this country and and globally than it is.

It's still far too challenging to to get some types of basic care. Um but the reality is that it is difficult to think in terms of significant change to these very large programs um anytime soon. And so um even though Medicare is imperfect and I think it will be difficult to change and what we find is more useful is is finding ways to either work within the system or work around the system as companies are developing new technologies.

And so the in the best case they can really form good open dialogues and and partnerships with the FDA and with CMS which administers Medicare. Um in addition for early stage trials it's possible to do those in other countries where there's the same uh where where the the the clinical trial data is acceptable to the FDA and things like that. So it is certainly um not as easy as we would like always but I think focusing on the risk is really only half the story.

I think there are there are definitely ways to move drugs faster and to develop faster. So I'll just give one quick example. Um a company we funded called Andromeda Surgical which is uh doing autonomous surgical robotics an amazing company built by a repeat founders where they're taking the same technology that we see in Ubers and excuse me in Whimos you know Whimos which have been shown to be safer per mile driven than human drivers.

taking the same concept and applying it to complex surgical procedures that uh robots can be more safe and effective at doing than than humans. Um they've been able to go from starting the company two years ago to treating 30 patients with their technology successfully for difficult procedure in just two years and um and that type of pace is possible today despite sort of the complex regulatory environment. abrupt question, but uh do you feel you'll uh die in this lifetime?

Do you think you'll live forever granted all the technology and the uh longevity escape velocity we're about to hit? >> I do think uh that we will still die. I do hope that we can extend the maximum lifespan and certainly we have uh we we invest in companies that are working on not only those long-term solutions but intermediate solutions that are again like still viable.

So for example, one company that we we backed which is doing amazing work has a long-term uh goal for uh for impacting maximum human lifespan but in the in the near term they their same technology can be used to improve um or the organ transplant system and so making there you know probably I think it's 30 or 40% of all um transplantable organs actually don't make it into a a recipient and and so that's a problem that they're they're working on in the nearer term. But it I think maximum lifespan is only part of the story. So we now today already are seeing significant impact on maximizing our existing lifespan in a way that wasn't possible even just a few years ago.

And to me, that is sort of the unsung remarkable story of of longevity where we can make it much easier for people to make the most of of the natural lifespan that's available to us. And so, you know, for for for many people, I think that's going to be getting closer to 100 years, which whereas today we're we're much closer to 80 on average. And that to me would be a huge win completely separate from you know the the larger breakthroughs that we need to be able to get to say a lift to 150 >> because thinking you'll die in the valley is a contrarian statement when most companies are like we're going to fix all diseases right um what about cancer 5 to 10 years do you think will fix it that gigantic problem that tumor >> so I think it's probably not going to be possible to completely fix cancer or cure cancer in 5 to 10 years.

But I think we will make way more progress in the next 5 to 10 years than people realize. There's a couple of reasons for my optimism here. So one which I think maybe people understand on some level but maybe not consciously uh discussed as much as it could be is the significant progress we've already made over the last 20 years.

So, everyone, nearly everyone has friends or family members that have been touched by cancer at some point. And it is uh now getting to be where some types of cancer that you know 15, 20 years ago or even 10 years ago would have been a death sentence is now more of a chronic disease. It could be managed like a chronic disease.

So, our success rate in mortality against cancer is is significantly better than it used to be. and and likewise the um the impact on people's lives is we getting much better medicine. So amunotherapies for example can when they are effective in a in a patient allow a a much longer um uh period of remission than than than and and much fewer side effects.

So those advances are important already. But now with AI, with our ability to understand complex biological systems much more effectively with use of computers and um move a little bit faster in getting through some of the um the research studies that we need to do, I think it is highly likely that we're going to see an increasing rate of success against cancer in the next 5 to 10 years. I would put that at at 90 plus% likelihood.

>> Nice. Good news. Um, what's the next holy alpha fold moment in your space?

And yeah, what would that moment look like and what company would be behind? Is it OpenAI? Is it Retro?

Uh, is it Dennis at Alphaold at Yeah, Deep Mind again? >> Deep Mind. Yeah.

>> Yeah. What uh what are you foreseeing here? So, it's a little bit too hard to tell because exactly who's going to do it because uh things are moving so fast.

So, humans are generally pretty bad at at predicting things that are happening at an exponential rate. We're definitely in an exponential part of the curve. Now I think what gives me a lot of optimism is that we have so many people working in so many different areas.

You know in in the US alone healthcare is a5 trillion dollar market. That is five times the size of the software market. And yet we normally thinking about AI in terms of its impact on software.

And I believe that AI's greatest impact is going to be in health. And the reason for this is one like we talked at the about at the beginning there's way more people uh using computational approaches in studying biology and medicine than there were in the past. We now have way more data available for them to work on and that is growing.

The amount of data we have is growing exponentially. Not just biological data but also medical records which has been up until recently pretty hard to get good medical data. And then these transformer models where we think generally in terms of the large models that are built by the um the foundation company uh excuse me the research companies like open AAI or or anthropic um but people are also building specialized models and we are also funding companies that are building specialized models which allow us to understand a particular process much much better and train and think about uh genetic and biological mechanism.

in the same way that the the large language models are trained to think about language. So there are lots of different approaches that people are taking to that and at the same time we're actually seeing the LLMs do kind of way better than than we might have expected even a year ago on some of the complex biology and both both all of those large labs have announced major efforts into health and medicine. So we are in a golden age I think of AI for healthcare.

That is part of why we started this fund is because we are seeing the ability for founders to move faster with less capital and make a larger impact on health and the way that care is delivered than would have been possible even a couple of years ago. >> And can't breakthrough be mostly predicted by the greatest talent density in companies? If we check open AI, right?

They just reunite a bunch of genius in a room uh and yeah gave them unlimited capital. Isn't that the main equation to predict breakthroughs in your space? [snorts] >> I think that is not necessarily the only thing that matters but it's certainly a great model for how to do it.

So I think of it back to sort of the Bell Labs days where it was this same kind of model. Um and where we have seen the research uh institute that I worked at in Stockholm. Part of why I was attracted to it was the same thing.

They were bringing together a lot of people from different disciplines to work together. I think the human genome project was a similar approach. Bringing technologists and biologists together and anytime and YC actually YC and Pioneer both follow the same approach.

One might even argue that that part of why Silicon Valley and the Bay Area has been so successful as a breeding ground for innovation is because of the density of talent and the optimism and the interchange of ideas. So >> I think the model of putting people together who are motivated to solve a problem and who have the um the background and skill skill set to contribute to a problem is likely to get to significant advancements faster than you would otherwise simply because you can iterate faster. And this is sort of one of the key elements of any enter of any enterprise particularly in a startup.

It's being able to iterate fast to to have a high velocity of discovery. You you know you it's not that you're going to get from point A to point B in a straight line. There's always going to be twists and turns, but the faster you can get through there, that's a compounding effect.

>> I love that. Dave, thanks for your time. Where can people find out more about you?

I'm Dave Msina. I'm the heading up the future of health fund at Pioneer. Our website is health.pionerfund.vc.

Uh and you can reach out to me there or on Twitter. Um looking forward to uh connecting with your audience. Thanks, Charles.

>> Pleasure. But I was Dave Msina