Everything is a Signal: Sahir Ali on Connecting AI, Biology, and Venture Capital
Sahir Ali’s career defies conventional paths. Trained as an engineer, he worked in quantitative finance, contributed to cancer research, built software companies, and ultimately founded Modi Ventures, where he developed the BioStack investment framework for technology-driven life sciences.
In this conversation with Sean Khozin, Sahir explains how systems thinking, mathematics, and first-principles reasoning shaped his approach to investing. They discuss the evolution of artificial intelligence in medicine, the realities of drug discovery, founder psychology, venture capital, and why enduring innovation often comes from connecting ideas across disciplines rather than following trends.
Transcript
Everything in life is a signal and hence you should be looking at the world through signal processing.
Speaker A:Have this punchline where I say that my first language is Urdu, second is C and third is English.
Speaker A:We're actually thinking about 21st century problems with 20th century tools.
Speaker B:Welcome to Precision Signals.
Speaker B:I'm Sean Kozen.
Speaker B:My guest today is Sahir Ali, founder and managing partner of Modi Ventures where he's rethinking how capital flows through biotechnology by applying the principles of systems engineering, quantitative finance and artificial intelligence to the future of biomedicine.
Speaker B:But this conversation isn't really about venture capital.
Speaker B:It's about a way of thinking.
Speaker B:Long before artificial intelligence became part of everyday conversation, Sahir was moving between seemingly unrelated worldsengineering.
Speaker B:Quantitative trading, cancer research, enterprise software, entrepreneurship, and now investing.
Speaker B:What ties those experiences together is a relentless search for underlying patterns.
Speaker B:The belief that complex systems, whether financial markets, biological networks or technology platforms, reveal their deepest insights when you learn to recognize the signals beneath the noise.
Speaker B:In this episode we explore the experiences that shaped that mindset.
Speaker B:We talk about growing up as an immigrant, discovering programming before mastering English, the evolution of artificial intelligence in medicine, why biology is increasingly becoming an engineering discipline.
Speaker B:And how a framework he calls the bio stack is influencing the way he thinks about investing in the future of healthcare.
Speaker B:Along the way, we discuss what separates genuine innovation from hype, why storytelling is just as important as algorithms, and how advances in computing, biology and capital are converging to reshape the next generation of biomedical breakthroughs.
Speaker B:Let's step into the conversation and trace the signal beneath the noise.
Speaker B:Foreign.
Speaker B:Welcome to Precision Signals.
Speaker A:Thank you for having me.
Speaker A:It's a pleasure.
Speaker B:Sahir is so great to have you here.
Speaker B:You know you've had quite an unusual career trajectory from engineering to quantitative finance, cancer research, enterprise AI and now a very interesting venture firm that established relatively recently.
Speaker B:And I think there's a consistent thread that runs through everything that you've been doing which is your quantitative mindset.
Speaker B:But before we get there, I'd like to delve into the very beginnings of your journey.
Speaker B:Let's start with where you grew up.
Speaker B:If I understand correctly you grew up in Texas in the 90s.
Speaker B:Can you tell us about the environment that you grew up in and your early influences?
Speaker A:Yes.
Speaker A:First of all, thanks for having me.
Speaker A:So yes, my, I grew up in Texas but my family moved to Texas on the whim from, from Karachi, Pakistan and, and blue collar parents.
Speaker A:But one of the first sort of, I would say the, the inflection point for me was that I grew up in an area called Clear Lake.
Speaker A:Clear Lake which is known for space center.
Speaker A:And through where my dad used to work at a convenience store.
Speaker A:Through those networks.
Speaker A:I ended up started going to NASA for an after school program which was Dean Cayman's program was lego Leagues and First Robotics.
Speaker A: So anyways, as an: Speaker A:In fact, I have this punchline where I say that my first language is Urdu, second is C and third is English because I hadn't spoken English before I'd come here.
Speaker A:So anyways, that program was quite pivotal, particularly for someone like me who had come from a humble background.
Speaker A:And that allowed me to start thinking computationally.
Speaker A:And I think it was really a lot of credit goes to that program that Dean had started.
Speaker A:So I'm a proud alumni and I'm also a supporter of that program now and then.
Speaker A:Interestingly, the mentor that I was shadowing, he was quite into quant trading or the word quant.
Speaker A:And I obviously wouldn't know much about it.
Speaker A:I was just 15, 16 and I started to look deeply into that.
Speaker A:And so those were my beginnings of what we call the quantitative.
Speaker A:So it's kind of become very obsessed with quant trading and black Scholes and those algorithms and actually ultimately at the end of my freshman year in college, I ended up at a hedge fund.
Speaker A:So anyways, that's kind of how the quantitative journey starts.
Speaker A:Very early.
Speaker B:Amazing.
Speaker B:I didn't know your first language was Urdu.
Speaker B:So English is officially your second language.
Speaker A:Yes.
Speaker A:So we grew up lower middle class in, in Karachi.
Speaker A:And so English was not taught primarily.
Speaker A:Lots of the most curriculum was sort of taught in Urdu.
Speaker A:So English, when I, when we moved here when I was 11 years old, I hadn't spoken English ever.
Speaker B:Oh, interesting.
Speaker B:Amazing.
Speaker B:I didn't realize that.
Speaker B:I thought, yeah, we were actually born in Texas.
Speaker B:You know, I can relate to that because my first language is Farsi, Persian.
Speaker B:And I was drawn to math because language in general was kind of mysterious to me.
Speaker B:You know, I left before I was an articulate Farsi speaker and then I couldn't speak English that well.
Speaker B:So I was stuck in no man's land, if you will.
Speaker B:So I was drawn to math because it was rational and I could understand and I could communicate mathematically.
Speaker B:Do you think that had something to do with your interest in math?
Speaker B:Because most people at that age shy away from math.
Speaker B:You know, they find it threatening for some reason.
Speaker A:Yeah, so I wouldn't say it was just exactly math.
Speaker A:It was Something to do with algorithms and computational thinking, for example.
Speaker A:What drew me was this idea that you could was not the computer itself, although I never grew up with a computer, but once we got it to me the fascination was that you could actually use this tool to actually create something of your own.
Speaker A:And so the first time I wrote a hello World program and you hit Enter and something appeared on the screen that you wanted to appear to me as a kid, that was fascinating.
Speaker A:And so that led to this idea of computational sort of deep diving.
Speaker A:And then you increasingly realize the language that makes all of that happen is mathematics, even the programming underlying the compilers.
Speaker A:And then you look at the logic circuits, it's all driven by math.
Speaker A:So that's so, so, so, so my interest in mathematics came from sort of top down of understanding things and eventually come to a conclusion that if you wanted to become a deep tech mindset, you had to get into mathematics.
Speaker A:And, and obviously the programming came naturally as part of being very interested.
Speaker A:By the way, the other thing I should mention about my background is that my parents are Gujarati speakers and so, so we have a very interesting household where I grew up with Urdu because we're in Karachi, but my parents are from Gujarat, India.
Speaker A:They speak Gujarati.
Speaker A:So I had a very multilingual sort of upbringing.
Speaker A:And so language is always part of it.
Speaker A:So when it came to computer language, it was not anything new in the sense that it was yet another language.
Speaker B:Yeah, that's a great point.
Speaker B:Computer language is just another form of communication and language.
Speaker B:And what's interesting is that computer languages have become more human readable over time.
Speaker B:You know, I think Python was a breakthrough because it was more human readable.
Speaker B:Not entirely human readable, but it was more rational than C C. And now we're in the world of natural language programming with clock code and similar LLM.
Speaker B:So it's interesting.
Speaker A:By the way, I should mention Andre Karpathi had tweeted, I think six, seven months ago that English is the hottest computer language now because of the.
Speaker A:Yeah, the Wipe coding hub.
Speaker B:Yeah, it's in a way like a natural evolution of programming because frankly the symbolic representation was interesting, but it was really confined to those who wanted to be esoteric and you have to be a little abstract, an abstract thinker to be able to understand programming.
Speaker B:But now anyone, I guess, can exercise their imagination.
Speaker B:So it seems like you were certainly an abstract thinker and you wanted to build things and explain things in a new language.
Speaker B:And I always thought of folks that speak, for example, are bilingual or speak different languages, they manifest different parts of their personalities when they speak in that language.
Speaker B:For example, when I speak Persian, I think my personality is a little different.
Speaker B:I surface areas of my psyche that I don't have access to, frankly, when I'm speaking English.
Speaker B:So programming is probably the same thing.
Speaker B:And it landed you at Rutgers in an engineering program.
Speaker B:And what's quite interesting, and maybe you can dissect this out and help demystify it, when you were at Rutgers in undergrad, so you started to study engineering, but you were also doing quantitative finance, cancer research, and you were an entrepreneur building a company with your brother Amir.
Speaker B:Can you tell us how that come about?
Speaker A:Yes.
Speaker A:So actually how I got into Rutgers is, is an interesting story itself is that yes, I picked up English extremely well in a short period of time, but it wasn't good enough to crack the SATs, you know, time pressure.
Speaker A:And, and, and when I tell you that I had very, very, very low interaction with English, I really mean it.
Speaker A:It would take me, I remember early days just to formulate a sentence, would have to think about it for a period of time to actually even be able to bring it out.
Speaker A:And you know, by the time you're 15, 16, you have to take SATs.
Speaker A:In fact, I graduated high school a bit early.
Speaker A:Not, not that I skipped a grade, but it's just that I started the school, school early, when I was early.
Speaker A:So by the time I'm four or five years into this country, I had to now take SATs, so my SAT scores weren't that great.
Speaker A:But because remember the dean's program.
Speaker A:And through that I had a sort of mentorship at NASA.
Speaker A:So the mentor advised me that you're doing all these interesting projects.
Speaker A:Why don't you burn them in a CD and actually send that along your resume?
Speaker A:And actually that's what sort of got me into a top tier engineering school, not my SAT scores.
Speaker B:Wow.
Speaker A:And one of those projects was coming back to the languages.
Speaker A:So the other aspect of why I became so deep tech and programming and the computational part of all of that, it actually allowed me to find my own fitting culturally.
Speaker A:I, by the way, just the other thing I should note is we came without access to American tv and we had a black and white TV for a long period of time.
Speaker A:And so I was, culturally, there was a bit of culture shock for all of our brother, my brothers and us.
Speaker A:And so this allowed us to kind of find my own path.
Speaker A:And so I'll give you one quick example.
Speaker A:In high school English, everyone has to read Beowulf at some Point which is very archaic English sort of novel.
Speaker A:And we were asked to get into groups and come up with a very creative way of describing the journey of the hero.
Speaker A:And all the adversaries went to.
Speaker A:So I on the other hand chose to actually use Macromedia Flash to animate it.
Speaker A:And I did that project on my own.
Speaker A:And so that was one way to.
Speaker A:And I wouldn't feel comfortable going into the groups because I didn't have anything creative to contribute because I was still kind of getting into the cultural aspects of this.
Speaker A:So those sort of things, I was building unique projects.
Speaker A:I had reverse engineered a DLL of Novell client.
Speaker A:So those things, I packaged it up and I said that anyways, that advice kind of changed my life.
Speaker A:And.
Speaker A:And so yes, I got into engineering.
Speaker A:At the time I thought I was going to do computer science, but when, when I took the Engineering 101 a professor effectively made a case that.
Speaker A:Electrical engineering professor made a case that everything in life is a signal and hence you should be looking at the world through signal processing.
Speaker A:And I kind of knew that because I was interested in a quant finance world, a quant which is a shiny thing that markets are signals and you do signal processing and things like that, pricing models and all of that.
Speaker A:And at the end of my freshman year there was a small booth by City Altar Investments Group and I just went to them and I spoke about my passion and some of the things I had already done.
Speaker A:And I knew a few things about Black Scholes and all of that.
Speaker A:So they were quite impressed.
Speaker A:And one thing led to another and I became one of the highest paid interns in Rutgers history through that program.
Speaker A:A financial crisis happened during that time, but luckily I came out of that program, the summer program, with a full time job.
Speaker A:I had a BlackBerry and they said go back to school and you work will work with your schedule.
Speaker A:Actually about six months later I attended a seminar in biomedical engineering which is by a young professor at the time, Anant Madhabushi.
Speaker A:And he was making a case that we could use yet another quantitative approach quant in interrogating medical imaging and what human eye can't sort of tease out.
Speaker A:It seems like we can find signals to not just diagnostic for diagnostics, but prognosticate how do we predict the outcomes of cancer which ultimately what matters is not degrading and staging.
Speaker A:And so we can be quite precise.
Speaker A:I thought that was quite interesting, although I didn't have any background.
Speaker A:So I joined the lab.
Speaker A:That's what led to some of the PhD work.
Speaker A:I've done later at Case and around the same time, I was picking up new cloud tech Salesforce and all of those orders.
Speaker A:Kind of at the time, very quantitative in the sense that there's a new cloud language and all.
Speaker A:And my brother built an e commerce company which was also in many ways based on this sort of quantitative thinking that how do you actually optimize for zip codes and all?
Speaker A:So anyways, that was kind of the thread during my undergrad and I was looking to say, well, how do you describe this journey where you're doing so many things that a few things in parallel, which the conventional wisdom and all.
Speaker A:Every professor told me, you got to pick one thing or you'll become jack of all trades, master of none.
Speaker A:Until I heard Noubar describe what they do at flagship pioneering as parallel entrepreneurship.
Speaker A:So I said, well, maybe I'm a parallel entrepreneur.
Speaker A:That's.
Speaker A:That's how I started describing myself.
Speaker A:And so Those are the three tracks for next 10, 15 years.
Speaker A:Eventually converged into what Multi Ventures is.
Speaker A:And we'll talk about that.
Speaker B:Yeah, interesting.
Speaker B:So in college, really, that was the first manifestation of parallel entrepreneurship.
Speaker B:And I wonder how many hours of sleep you were getting.
Speaker A:Very less.
Speaker B:You were juggling four or five different obligations.
Speaker B:So just to recap, to make sure that I'm getting this correctly, so you actually started when you were in college to work in quantitative finance.
Speaker B:Right.
Speaker B:At a hedge fund through the evolution of the program that you mentioned, and then working with your brother to build a company, correct me if I'm wrong, which was very quantitative in its foundation, indexing zip codes and so forth.
Speaker B:And also you were inspired by some of the mentors that you had and were drawn into quantitative histopathology, let's call it.
Speaker A:Yes.
Speaker B:Am I missing anything?
Speaker A:Yes, exactly.
Speaker A:That.
Speaker A:It's the quantitative quantity histopathology and medical imaging.
Speaker A:I view all of these three as sort of a quant mindset.
Speaker B:Yeah, it's interesting.
Speaker B:So this was in.
Speaker B:This was a long time ago, and.
Speaker A: At the time,: Speaker B:Right at the time, I believe the biomedical universe didn't really look seriously into quantitative histopathology or AI in general.
Speaker B:I remember when I went to method school, I quickly realized that the quickest way for me to lose all credibility was to mention the word AI.
Speaker B:So I put my head down and, you know, tried to act like everyone else.
Speaker B:So how did you navigate your way through that?
Speaker B:And what was the reaction that you were getting from people when you talked about AI and histopathology?
Speaker B:For example.
Speaker A:Yeah.
Speaker A:So first of all the credit goes to my advisor at the time, an Aunt Madhabushi.
Speaker A:So he was kind of a forward thinker in that early on he had realized that digital pathology was less about pathologists but about oncologists is that we can always make tools that can do faster grading or accurate grading or staging.
Speaker A:But ultimately what matters is when an oncologist gets a patient they have to decide what dosage and which treatment and what dosage.
Speaker A:And so there are very less tools that help you figure out what's the risk profile of that patient.
Speaker A:Yes, we do staging and grading, but also there's realization that HNE stained biopsies on solid tumor were just rudimentary.
Speaker A:We've been doing this for decades and decades and we interrogate that under microscope and then we start to now understand if you could now use algorithms to quantitatively interrogate these images.
Speaker A:And that by the way includes MRI and all as well.
Speaker A:But particularly my focus was pathology is that you could now start to actually do nonlinear systems thinking, which is again with a quantitative mindset, says that it's not about a single gene or mutation and one outcome, but rather we can start to look at tumor micron environments, we can start to think about a system sort of thinking and so we can start to really approach outcomes of cancer through that.
Speaker A:And so we thought that pathology was a foundational sort of layer to that and it was getting digitized.
Speaker A:It was early days, but we certainly believe that machine learning, computer vision, intersection and convergence was going to play a big role.
Speaker A:And that again credit to Anant, he was a forward thinker in that and he started to get really good partnerships and data and especially on colleges on board who believed in that.
Speaker A:So I tagged along with that.
Speaker A:But yes, we built some interesting models.
Speaker A: In fact I wrote in: Speaker A:So I go very deep in the GPUs and actually we'll talk about it.
Speaker A:Ultimately people did see that Multiventures was an investor and did well with Groq, which was an inference chip.
Speaker A:So then the realization didn't come to me after ChatGPT realized that inference was going to be key almost very long time ago because we struggling with inference of these images and all.
Speaker A:Anyways, so that's kind of how those time periods were.
Speaker A:And to answer your question, eventually what was the reaction?
Speaker A:Actually some of the, some of we would present this at joint conferences.
Speaker A:Sometimes they were just purely computational, sometimes cap and ASCOS and those abstracts where there was a lot of clinicians.
Speaker A:I think based on some of the work that I got involved with, I think we got a lot of appreciation and obviously skepticism as well.
Speaker A:Appreciation was that, hey, we can imagine why these tools are going to be profoundly impactful at some point.
Speaker A:Skepticism at some times were like, well, these are small studies.
Speaker A:How is this ever going to generalize?
Speaker A:Is a.
Speaker A:Is the algorithm actually that good?
Speaker A:Because this is early days, not many people even knew about machine learning and all.
Speaker A:So there was.
Speaker A:Right, skepticism.
Speaker A:But there was, there were a lot of folks who were very excited.
Speaker A:In fact, that's where the data came from.
Speaker A:The data, the guardians of the data were these clinicians and those partnerships came in.
Speaker A:So I would say there were a lot of forward thinkers at the time, including some of the data came from Steve Hahn, whom you know him very well.
Speaker A:He eventually became the head of fda, who was at MD Anderson.
Speaker A:We were getting data from very large institutions.
Speaker A:So anyways, it was an interesting time period for me as a young student to learn all of that, but also one thing that led me to realize that it was not always about the algorithms.
Speaker A:It was about how well you were projecting, what the utility of these things were.
Speaker A:So it made me a better communicator of science, better communicator of complexities, how to unravel the black box, how do we tie it to these things, to biology, which is ultimately what clinicians wanted.
Speaker A:And so those were very important tool sets or skill sets that I learned while I was doing that research.
Speaker B:Interesting.
Speaker B:And it seems like Sahir, you carried that forth to Case Western where you started your PhD.
Speaker B:And what's remarkable to me is that you published 25 papers, which is something that is almost unheard of for somebody pursuing a PhD.
Speaker B:And then you also had a very interesting TEDx talk on digital histopathology.
Speaker B:Walk us through that experience.
Speaker B:And what do you think was behind your academic productivity at the time?
Speaker B:What were you trying to pursue?
Speaker A:Yeah, just to be clear, the 25 papers were first and mix of co authors and things.
Speaker A:But yes, I did have a high, high number of publications.
Speaker A:I think it was kind of an era in the time period was that computer vision had started to show some amazing breakthroughs.
Speaker A:So you had imagenet, you had sort of convolutional neural networks breakthroughs that had happened.
Speaker A:So really overall, broadly in the academic settings, not just medical imaging, but overall, people start to realize that computer vision is here to stand.
Speaker A:It's finally had its moment.
Speaker A:And so if you think about the work we were doing was in medical imaging and so it started to gather a lot of attention.
Speaker A:In fact, medical journals were open to accepting publications that were based on quantitative approaches.
Speaker A:Machine learning, which I thought was seeing a bigger inflection point.
Speaker A:So I'm not saying it was easy to publish, but I think it was the right time to start publishing quantitative approach across many different channels rather than just computer science.
Speaker A:My first paper, journal paper, was actually IEEE Transactions of Medical Imaging Heavy Mathematics.
Speaker A:But after that there was a paper in surgical pathology about oropharyngeal cancers and it was written with a clinician as a co author and I was the first author on that.
Speaker A:So you can see that was a whole evolution from just purely computer science to actually in a journal paper.
Speaker A:So I guess it was the time period that allowed us to find the right journal papers and avenues to publish and people were a lot more receptive because at the same time computer vision was having its big moment, the breakthrough moment as we say it.
Speaker B:Yeah, interesting.
Speaker B:You know what's quite intriguing to me is that you had a very successful research career while you were pursuing your Ph.D. and then you went to KPMG as a consultant and then later on became the cto at in TiVo Health.
Speaker B:So you became an operator.
Speaker B:Walk us through that chain of decisions that led you to pursue consulting and then be being an operator at.
Speaker A:So in that parallel entrepreneurship, I was always a consultant actually to.
Speaker A:So one thing I forgot to mention that while I was at Wall street due to my financial crisis and market crash, what I. I actually picked up Salesforce.
Speaker A:It just happened to be the right time, right period.
Speaker A: esforce is very early days in: Speaker A:And it allowed me an opportunity to learn about a growing field of what it means to be in cloud.
Speaker A:For example, Salesforce, a pioneer in actually creating a platform where even your programming language they had was a unique programming language they called Apex was built to write code on the cloud.
Speaker A:What does that even mean?
Speaker A:It means that when you're on the cloud there's something called governor limits.
Speaker A:You're a multi tenant system, so you can't just write code that used up all the resources.
Speaker A:That was a different thinking versus say you wrote code for a desktop software and all.
Speaker A:So that was a bit of mindset shift.
Speaker A:I came in with no baggage of that enterprise.
Speaker A:So I picked it up, I guess as a young guy, picked it up really well, which is one reason why I stayed on much longer.
Speaker A:But three or four years went by by the time I was senior in college, I was considered a senior sort of resource.
Speaker A:So a lot of the headhunters would reach out and said, hey, can you come on to this project and advise?
Speaker A:And people will be shocked when a 22 or 23 year old walked in and advising on an enterprise architecture.
Speaker A:So that part continued.
Speaker A:Yes, I did go on as a, as a director at kpm, just quite young, but that part continued on.
Speaker A:I'd done 50, 60 implementations of Salesforce actually for Fortune 500 companies.
Speaker A:So that was one part.
Speaker A:Just continued on during my undergrad master's and when I was doing my research at Case Western as well.
Speaker A:But the other part was that at some point I did say that what does the intersection of blockchain technology mean in medical stuff?
Speaker A:And so Intiva Health was that they had, they had brought me on to say is there a way we can actually solve the medical credentialing problem?
Speaker A:Everyone has to maintain the credentials.
Speaker A: t least it did at the time in: Speaker A:But once you get verified, is there a way we could actually create that provenance structure on a, on a sort of a blockchain where it's just immutable and once you're verified it could be shared across as a, as, as your vault and you can just give it to other.
Speaker A:And the idea was that can you sort of onboard folks who are licensed medical professionals within hours rather than weeks.
Speaker A:And so that actually went live, it was in thousand clinics and all.
Speaker A:So that was actually a pretty.
Speaker A:So I took any opportunity that made me cross thinker and ability to think across different modalities.
Speaker A:And so I was not afraid to kind of get into those things anyway.
Speaker A:So that's, that's, that's where that's what you see come up in my background sometimes.
Speaker B:Yeah, interesting.
Speaker B:You know, the ability to cross pollinate and see the common threads across various disciplines I think is so valuable today because the discipline boundaries are breaking because of our computational capabilities mostly.
Speaker B:And I think in healthcare and biomedicine, the more we are able to measure our diagnostic capabilities are increasing exponentially, the more we can compute.
Speaker B:So in a way the life sciences is approaching the physical sciences.
Speaker B:Is that a fair statement to make and is that how you thought about biology versus finance and engineering?
Speaker A:I.
Speaker A:Yes, actually, if maybe it's the right time to think about, you know, we did really well from our entrepreneurial experiences, the companies we, you know, the companies did extremely well financially which allowed us to.
Speaker B:With your brother.
Speaker B:You're talking about the companies you were building with your brother or.
Speaker A:Yes, that's also couple of setup that I had in the cloud space.
Speaker A:Obviously the hedge fund allowed us initial financial freedom, the research work.
Speaker A:But I think at some point sort of all of these sort of distinct areas were there had a common thread to it and the common thread to it was well, how do you think about systems approach to many things?
Speaker A:Is there a first principles thinking?
Speaker A:So let me walk you through a couple of things.
Speaker A:Right.
Speaker A:Salesforce.
Speaker A:I know it has nothing to do with cancer, but Salesforce Marc Banioft had a very first principle thought is that you know, what if we actually take an entire schema and take it away from enterprise to manage and what if we put it on the cloud and all you have to do is just log in.
Speaker A:We handle everything in the back end and by the way we can actually track everything you do from lead to cash.
Speaker A:Nobody was doing that because in that entire journey at the time in enterprise everyone had separate databases, separate softwares to do the entire customer relation management.
Speaker A:And so this was a first principle thinking is that that should all be just one database and by the way that should be put up on the cloud.
Speaker A:So you all you have to do is log in and maintain your data and everything is that led to what we call the poster child of a SaaS company, Salesforce Systems thinking first principles again.
Speaker A:In the lab that I spent close to eight years, seven to eight years doing research was the first principle thinking.
Speaker A:You know, for a hundred years we diagnosticate ends of high I guess prognosticate solid workflow cancer using microscopes which haven't evolved for long period of time, still do the staining.
Speaker A:We do the same thing.
Speaker A:And so what are we leaving behind?
Speaker A:We have.
Speaker A:We're actually thinking about 21st century problems with 20th century tools.
Speaker A:Actually the staining was 100 years ago, the microscope was 100 years ago.
Speaker A:We look at the same sort of structural changes.
Speaker A:Visually it's a highly subjective field.
Speaker A:So first principle thinking again what can systems level thinking think about?
Speaker A:We should start thinking about human microenvironments and those digital biomarkers of sorts.
Speaker A:I think come to building the companies was the same idea is that we have a convergence of online payments on the cloud.
Speaker A:You have convergence of finding consumers by the way through Google Ads so you could just have a laptop.
Speaker A:And by the way you had rise of platforms like Alibaba and those so you could technically just have a laptop.
Speaker A:You can find millions of customers optimize for that and you could build a three to four people company and you could scale that to in totality 150, $200 million in revenues.
Speaker A:Again, it involves to think about first principles thinking.
Speaker A:If you got bogged down and saying, oh, we're going to need a warehouse, how are we going to find the customers?
Speaker A:You're going to need a large ads, budgets.
Speaker A:I think the thread was all of that is that what computational tools you can think about systems level thinking and first principle questions.
Speaker A:And so that's what led to what, what what for last three years, Modi Ventures is it is kind of taking a step back and thinking about how do we fund in private market, some of the therapies and some of the tech bio space that it's quite prominent and all.
Speaker A:So anyways that's.
Speaker A:So that was the common theme that led to what I'm doing now, which I'm passionate about, which I think I'll be doing for next couple of decades, hopefully.
Speaker B:Oh, interesting.
Speaker B:So let's talk about Moody Ventures.
Speaker B:So you started Moody Ventures about three years ago, doing a market environment that was unfriendly to raising capital by any means.
Speaker B:You had never managed a fund.
Speaker B:You were quite successful as an entrepreneur, you were a quant, you had an academic legacy, which was very respectable.
Speaker B:What was your pitch and how were you able to convince your limited partners to trust you as a first time fund manager?
Speaker B:Because you, if I, if I'm correct, you managed to raise quite a large sum very quickly in a matter of months.
Speaker A:Right.
Speaker A: about it because in December: Speaker A:I mean I knew we were, I was quite convinced in the strategy.
Speaker A:It was newer way of thinking.
Speaker A: y who was tracking the market: Speaker A:And so it was hard to raise a money fund.
Speaker A:Well, I would say in many ways we continue to see the funding challenges today, but the journey to launching Multi Ventures didn't start just on the whim.
Speaker A:Actually during the COVID I was starting to think about what is it that I still feel passionate about and what I wanted to do next which was still required a first principles, thinking we could take a systems Approach again, all those threads that I was talking about.
Speaker A:And it's worthwhile entering the space as an outsider.
Speaker A:And if you think about all the threads that I pulled, including coming to this company, I just tend to do well if I'm an outsider because it allows you to have a naive mindset, but also approach to ask the first principal question.
Speaker A:So the very first question I asked was in venture capital world, where does the money come in and where does the money go out?
Speaker A:And just like everybody else, it was an opaque world to me and it made me realize that 90% of the money that comes into venture asset class comes from pension endowments and those folks, but it doesn't go into directly into the underlying assets, the startups.
Speaker A:It actually goes through these vehicles we call venture capital funds.
Speaker A:And in fact it goes through funds that are quite well established.
Speaker A:What I mean by that is typically fund five, six, they've been around for a decade plus and obviously it makes sense because pensions managing public money, they want a lower volatile vehicle to park the money and then the money then goes to startups.
Speaker A:So my first principle question was if there are two ways of investing in venture asset class, which is going through the established funds and directs.
Speaker A:Well, perhaps there's a way we could actually think about modern portfolio theory here because we have seen this kind of problem before where Markowitz won a Nobel Prize and his fundamental question was if you take the base case of a stock and a bond, a stock was a bit highly volatile compared to a stock and gave you a higher return and stock was lower volatility, lower returns, but you, but they tend to be uncorrelated.
Speaker A:So I thought what if we modeled a venture asset class as sort of a bond and a stock, not, not exactly as that.
Speaker A:But imagine your LP positions were kind of like the, the bond structure where, you know, low risk, low return, but you put direct money.
Speaker A:And is there a way we could apply the Markowitz sort of thinking and what he came up with was the idea that you can bend the curve of the efficient frontier that most of the the folks listening might recognize that's where it comes from.
Speaker A:So my punchline was can we build an efficient frontier in the venture asset class at a high level?
Speaker A:And second was it turns out when you look at the bio spectrum, it actually allows you that volatility to be able to do that.
Speaker A:What I mean by that is that an idea spins out from a research lab, someone's PhD or some IP or university, and then it enters IND ready for human testing phase one, phase two, phase three and post market you'll draw this sort of bias spectrum.
Speaker A:At each of those stages you have different volatility.
Speaker A:And so for example phase one you have 80% volatility versus say and phase three is still 50% where you could still fail.
Speaker A:But in all of those you still have deterministic outcomes.
Speaker A:For example phase two, a lot of the companies end up exiting to get bought out IPOs, but we know the peak sales and all of that.
Speaker A:So it allows you to now say if you can build an uncorrelated portfolio of these, you can bend the curve.
Speaker A:And by the way, you can also invest in funds that are also on the same spectrum.
Speaker A:And so hence if you look at our strategy, it is quite Markowitz in its thinking, but it applies to sort of bio where you're actually really precisely indexing the market.
Speaker A:So if you from.
Speaker A:So for example we have a fund called Abingworth, they do extremely well in building portfolios with assets that enter human trials and do well.
Speaker A:But they also have a clinical co development fund that does phase three and upwards we have a fund that does royalties after phase three.
Speaker A:So think about it like from a fixed income perspective.
Speaker A:We're LPs in Flagship pioneering, they own the whole stat of bio spectrum where they build the companies from 0 to 1, right?
Speaker A:So if you start thinking about and if I overlay their assets on the biospectrum, what you've done is really effectively precisely index the market.
Speaker A:We also have direct investments which are more high risk, high rewards.
Speaker A:And so to do all of that I thought we needed so that's a horizontal way of sort of indexing the market.
Speaker A:But we needed something vertical.
Speaker A:Meaning what is that framework ultimately where all of this sort of converges?
Speaker A:And that's what I, what I call the bio stack, right?
Speaker A:From DNA, rna, protein cells, tissue, organs and organism.
Speaker A:On that stack we'd like to make uncorrelated bets.
Speaker A:So for example, protein layer, we have a company that recently went to IPO generate biomedicines that drove protein designs.
Speaker A:But we also have another portfolio company that does macrocytic peptides unnatural product.
Speaker A:So you can see they're kind of uncorrelated but on the same layer.
Speaker A:Anyways, I'll take a pause, but that's kind of multi ventures.
Speaker A:It's not a traditional venture capital fund where you just construct a portfolio of certain indications or certain therapies you go after or a traditional VC where they say we're going to do frontier technologies and then you invest 50% and then keep Reserves.
Speaker A:There's a, I would say a first principle thinking, a systems level approach to it.
Speaker A:And also it is quantitative in nature.
Speaker B:Right.
Speaker B:That's fascinating.
Speaker B:So there's a lot to unpack here.
Speaker B:So going back to when you raised your first bolus of money, was the financial engineering framework that you just laid out part of your pitch or did that.
Speaker B:Okay, Interesting.
Speaker B:Yes.
Speaker B:And what was the reaction at the time?
Speaker B:It seems like, well, you had very good positive feedback because you raised again, a good sum of money very quickly.
Speaker B:So you build on the foundation and you were able to communicate that to folks that perhaps have never thought about portfolio theory in the context of venture investing.
Speaker B:That's very interesting.
Speaker B:So what do you think was behind your success in those early days?
Speaker A:You know, I will say, and I've alluded to this fact earlier, is that while I was at Wall street, while I was in that lab with an aunt and all, I learned how to communicate complex black box ideas to folks who were the guardians of things.
Speaker A:For example, communicating what an algorithm is doing to a clinician who's providing data to us.
Speaker A:It was less about algorithm, but about a story that the data was telling.
Speaker A:So storytelling.
Speaker A:In fact, one other thing is that Salesforce actually, again, Salesforce is a constant mix of my career.
Speaker A:And I think a lot of people get surprised.
Speaker A:As someone who was in cancer research and doing deep quantitative, deep tech stuff, the Salesforce comes up a lot.
Speaker A: Salesforce in: Speaker A:So if you click the button, it would actually take your opportunities, data in the pipeline and it would create a story about it.
Speaker A:It would take the regression numbers and tell you why certain things were happening.
Speaker A:And so if you're a salesperson, you.
Speaker A:You don't want to see the regression scores.
Speaker A:And what does it even mean 96% chance is going to close.
Speaker A:It actually constructed a story, it charts.
Speaker A:And.
Speaker A:And that is how I approach the fundraising as well, is that at the end of the day, if someone's investing with you, they want to know who they're investing with.
Speaker A:They want to understand the story really well.
Speaker A:And third is they're investing because they want to make returns on it.
Speaker A:And so you could have all the fancy algorithms or fancy theories and all of that, but if you can't communicate it.
Speaker A:So that was one I would say, if you ask me, what made me successful was be able to tell the story on where the world was going to go and why we need this sort of approach and why it's differentiated.
Speaker A:And third, how we were going to do our risk adjusted returns here.
Speaker A:Well, if you think about the strategy, even more bare bones is that we were investing in funds with pensions and endowments, with public money.
Speaker A:So you had a basket of sort of low risk, low return to protect the underlying investment and then obviously direct investments, which are startups and high risk, high rewards.
Speaker A:So the volatility of the fund itself was very different than traditional VCs that most people thought about was a very simple story around that and particularly high interest environments.
Speaker A:It worked out well.
Speaker A:And yeah, I would really think that storytelling was the number one thing.
Speaker A:And really keeping in mind on what kind of investors you wanted to go after.
Speaker A:And it also helped that I had a skin in the game.
Speaker A:I was also an LP myself.
Speaker B:Yep, that certainly makes a huge difference.
Speaker B:So Biostack is essentially the heartbeat of Moody Ventures investment philosophy.
Speaker B:Yes, it's quite intriguing drawing from financial engineering and portfolio theory.
Speaker B:What's quite interesting, Sahir, is that you've applied these concepts to a fund that manages diversification in a, in a unique way.
Speaker B:Most people think about diversification in the context of a single portfolio of assets, whereas you broaden, I think that definition because you are taking positions, LP positions in other funds, you're making royalty based bets and investing directly into companies.
Speaker B:So in totality of those maneuvers, I guess.
Speaker B:Would that be part of your diversification strategy?
Speaker B:Is that a fair way of thinking about it versus a drug portfolio with an uncorrelated, uncorrelated assets of just drugs.
Speaker B:But you are, you have expanded that definition.
Speaker B:Is that a fair statement?
Speaker A:Yes, I think it's fair.
Speaker A:But also it depends on, because this is complicated strategy, it depends on which angle you look at.
Speaker A:So one could have a.
Speaker A:One could say this is just boiling the ocean, really.
Speaker A:That would be a fair criticism.
Speaker A:But I would also say that the reason why Biostack came about, actually I was inspired by a clip I had seen from Steve Jobs where he was talking about that, I think it may have been 80s where he said for last 40 years what we have done is really created something called a tech stack.
Speaker A:And what we've taken is a substrate like an electron and we created silicon wafers to suspend it, suspend the electron.
Speaker A:If you can suspend it, then you can sort of control it through these transistors and gating technologies that we developed, which was a semiconductor sort of revolution.
Speaker A:And then we, we created microchips that allowed us to kind of then create a system sort of layer and we created something abstract notion of programming and ultimately a personal computer.
Speaker A:And he said, well look, if you look at what we've done is taken a whole Tech stack, compressed it, gave it to somebody on their kitchen counter.
Speaker A:But he also said something that was quite interesting to me was that everyone forgets about the most top layer, which is the user interface and user experience.
Speaker A:Without that, that tech stack has a very limited use cases to our limited population.
Speaker A:Which is why, you know, Steve Jobs was very heavy on thinking about how do we actually simplify all of that.
Speaker A:So that's Tech Stack.
Speaker A:So I, you know, as Deepak Srivastava Gladstone and, and who was the author of the Emperor of All Melodies, Siddh said, Siddhartha Mukherjee.
Speaker A:Right.
Speaker A:I mean, so when, you know, when listening to these guys, it comes out that the framework of thinking about disease, human disease particularly can be quite simplified.
Speaker A:Kind of like what we did with electron is that at a basic level, most human disease is either a function of cells being lost or cells malfunctioning.
Speaker A:And so we have started to develop tools to do both of those.
Speaker A:So for example, if cells are malfunctioning, we can hack through CRISPR gene therapies and few other things.
Speaker A:And if the cells are being lost, we have IPS stem cells and be able to replace those things.
Speaker A:So we're starting to see those things.
Speaker A:Ultimately we can start to tackle human disease and human life in a very engineering, curative ways that we obviously never seen before.
Speaker A:So if you think about that framework and we look at what's going on in private markets, you could just take a chunk of it or I'm saying I'm going to take an asset class view and actually capture as much of innovation as I can.
Speaker A:And hence this biostack framework comes up, is that just like Tech Stack, from electron to everything we have today, ultimately we gotta be able to manipulate the cells and so on.
Speaker A:That stack, what we've seen over the last two decades, profound breakthroughs on read, write and reprogram.
Speaker A:What do I mean by that?
Speaker A:Is if you look at DNA, rna, protein cells, tissue, organs and organism.
Speaker A:So even if you start at the DNA level, we call genomics through nest gen sequencing, we've had profound read breakthroughs that leads to RNA seq proteomics.
Speaker A:In fact we have a portfolio company, Glyphic Bio, that's built on nanopore, that's trying to do a lot proteomics in a, in a very interesting way.
Speaker A:And then we call it the human level, the organism level, the health.
Speaker A:Right where you read full Body scans, mri.
Speaker A:So you can see, you can think about it from a read perspective.
Speaker A:Like we think about in the tech.
Speaker A:If you think of the same, you take the same stack and now you think about the right technologies.
Speaker A:We've have CRISPR at the DNA level, we have MRNA at the RNA level, we have, you know, we can think about the cells and how to edit that in an interesting way.
Speaker A:In fact, at the tissue level we have regenerative medicines and at the human level, organism level, we call it precision medicine, personalized and precision medicine, that's been the goal anyway.
Speaker A:So that's how I think about that framework is allows us to kind of go after not just drugs, but really that nonlinear thinking about how do we engineer life itself itself.
Speaker A:Right.
Speaker A:So we can go into prognostic digital biomarkers, we can go after AI scientists, we can go after designing drugs through AI platforms, we can go after new ways of T cell engagement platforms.
Speaker A:That's why it's broad and so you can't do it all yourself.
Speaker A:So that's why this idea of Markowitz where you can go after sort of these assets that are low risk, low return, but these funds established to do certain things extremely well, but you can also make your own high risk bets.
Speaker A:And turns out that strategy was attractive to a lot of people.
Speaker B:Right, right, interesting.
Speaker B:And you know that nonlinear thinking has taken you in very interesting directions.
Speaker B:For example, you were one of the early investors in Grok, which was recently acquired by Nvidia for 3 billion plus 20 billion.
Speaker A:Oh, it was non exclusive licensing deal in cap 20.
Speaker B:Okay, well I should have known that 3 versus 20, that's a huge alpha.
Speaker A:And interestingly a lot of people were surprised when they saw multiventures in the Axios article that came out and said what is the life sciences or bioengineering or tech bio investment thesis doing in Grog, which is an AI chip company.
Speaker A:And so my question back was why not?
Speaker A:Because if you think about the BioStack framework, the topmost stack, which I didn't mention after organism is enabling technologies.
Speaker A:Anything that powers that biostack should be part of our uncorrelated thesis.
Speaker A:For example, about three years ago when I launched the fund, I told LPs AI inference, not AI training is going to be the focus.
Speaker A:I would like to find some opportunities all the way down to infrastructures that can enable AI inference, which is going to be bigger market than AI training at the time was.
Speaker A:And hence Grok made a lot of sense because if you think about the BioStack, we're going to need a lot of inference all the way from medical imaging down to looking at generating new molecules and such.
Speaker A:So that's why Grok made part of his thesis.
Speaker A:And the other company that we actually had some exposure through through one of our funds was oklo, which was doing nuclear fission, modular nuclear fission energy, alternative energy for data centers.
Speaker A:And again, if you think about the biostack in near future there from not only just medical practices all the way down to biotech companies and all, everyone has major uses of data and data centers, alternate management plays a role and the AI chips.
Speaker A:And so that was that topmost layer that powers the biotech.
Speaker A:Hence we were in Groq and obviously we did well with that.
Speaker B:Interesting.
Speaker B:So what's interesting also is that certain layers of your bio stack framework obviously go beyond just biomedicine, which in a way is a form of diversification because you're not pegged to a single sector, which you know by biomedicine there's quite a bit of group thing that drives a lot of investment and research decisions, unfortunately.
Speaker B:So Sahir, maybe it could be useful if you could walk us through the various layers and categories of the biostack.
Speaker B:If you were to categorize the main layers of the stack, what would they be?
Speaker A:Right, so, so one part of the stack follows.
Speaker A:So the way, again the way I think about just like tech stack, again, I'm going back to it.
Speaker A:The goal was how do we actually manipulate and control electrons?
Speaker A:If we did that, that's the basis of electronics.
Speaker A:If you think about it, even today, what we call AI is really at the core basis manipulation of electrons.
Speaker A:That's what it allows.
Speaker A:So you put in a prompt, it goes over the wire, it hits some data center, it hits a GPU chip, it generates a token.
Speaker A:At the end of the day it's all just electron sort of manipulation.
Speaker A:So the basis, what I talked about is that if human disease and life engineering revolves around manipulating cells or understanding that the system's architecture that involves cells, then the biostack starts with the central dogma within the cell, which is the core basic is the DNA, the rna, the protein, the cells, the tissues, the organs and the organism.
Speaker A:That's the layer of the stack.
Speaker A:And what allows us to do for each of that layer is to make sort of bets that are spanning the entire ecosystem from technologies that enable better read, write or reprogram and so say for example at the protein layer.
Speaker A:Let me again reiterate the example.
Speaker A:Generate biomedicine, which is now a publicly traded Company went to IPO about a month ago.
Speaker A:I was honored and privileged that they asked me to come and ring the bell with them again.
Speaker A:The idea was how do you optimize AI optimize use sort of these diffusion models and AI models to, to create sort of drugs that you know, took, took four to five years and now you can do that in a matter of weeks and months and take it into clinic.
Speaker A:Right.
Speaker A:But then we have unnatural products that uses macrocytic peptides as a platform using AI and really proprietary chemistry based data to say what kind of drugs we can create that have a, that fit into a different mathematical function between small molecule and large molecule, but that can have an interesting mix of those properties.
Speaker A:So you can turn say a biologics into oral, but then we have another company called Glyphic Bio which is using on top of nanopore to actually accurately sequence amino acids so that you could have proteomics.
Speaker A:So you can start to see that these sort of individual.
Speaker A:Traditionally a life sciences fund would probably only do say a drug that targeted certain protein that would be purely therapeutics or a tech bio fund would only look at AI drug discovery.
Speaker A:So to me that layer sort of enables the full story of cells being lost and cells being malfunctioning.
Speaker A:Ultimately we want to be part of capturing where this convergence of tech, biomedicine and health is heading towards and that allows this approach of precision indexing is, is what I has been part of the Markowitz mindset anyway.
Speaker A:Now you could be say flagship or you can be third rock or you can who have been around for 20 years and they'll, they'll take a view worldview and build companies and they'll always do well.
Speaker A:While I think about the asset class itself, how do I play a role and have an investment vehicle that actually captures it like in index funds did in hedge funds back in 28s.
Speaker A:So anyways, so that's where.
Speaker A:And it's not just an asset management but it's driven by underlying understanding of how the userman disease is evolving, what tool sets we need and how do we actually engineer life itself.
Speaker A:So it's driven by scientific thinking, it's driven by obviously financial engineering.
Speaker A:I don't want to keep thinking about financial engineering but, but it is driven by a deep thinking and understanding.
Speaker A:We have the right team to continuously think about what indications, what are mechanism of actions, what kind of tech bio setups, what sort of frontier models we want to be in and those are driving a lot of the decisions on which funds we invest in strategically and which companies we go in.
Speaker B:Right, interesting.
Speaker B:So the biostack is really the focus is at the atomic level in a literal sense.
Speaker B:Which means that you have to go really deep into the science.
Speaker B:So tell me about how the investment decisions are made.
Speaker B:Is there an investment committee or.
Speaker B:Because you would need really quantitative thinkers that can go that deep across various disciplines in a way that are part of the biostack.
Speaker B:Can you tell us a little more about how the investment decision.
Speaker A:So one of the advantage is that when you're thinking quantitative, you're thinking data from day one.
Speaker A:And if you think about the kind of funds and folks can go to our website and you will get a sense of kind of funds we're in and the quarterly reports that come from them, the annual gathering, the amount of data that we get, although it could be lagged by two or three months in private markets perhaps makes us the most market aware fund in biospace as well.
Speaker A:While say a single venture capital fund do what they do best.
Speaker A:We're aware of some of these top tier funds and emerging funds that we are partnered with.
Speaker A:We're starting to see the data and where things are, what's shaping up.
Speaker A:That is our single most advantage as well.
Speaker A:Remember, we get to see all of that across now.
Speaker A:Typical.
Speaker A:Even if you leave aside our direct investments, a fund of funds is really about asset management and just partnering with the right managers, not underlying view of the technologies and that deep.
Speaker A:But we're very different in that sense.
Speaker A:So that's number one.
Speaker A:Number two, this mindset of how we're going to invest and all this is honed into anybody that's involved in the fund, from folks who are working at the fund, from advisors.
Speaker A:And one of the things that you have to realize that you don't know everything.
Speaker A:So you need the right set of folks around you.
Speaker A:Although you could bring that mindset.
Speaker A:But at the end of the day you still need the right folks to help you sort of evaluate things, bring in new sort of deal structures and all of that.
Speaker A:So yes, I have got a tremendous layer of advisors, venture partners who are involved at various stages of we think about investments.
Speaker A:So I'm quite privileged at that.
Speaker A:These are some world class folks.
Speaker A:And the third is every investment we make, it plugs into a model that actually has correlation index.
Speaker A:It has what potential return like anybody else.
Speaker A:And third, at what part it actually fits into biostack as well.
Speaker B:Interesting.
Speaker B:So you're actually crunching the numbers and doing modeling and simulation for every investment.
Speaker A:Yeah.
Speaker A:I'll leave you and folks who are listening with what one example of quantitative number that come up.
Speaker A:So one of the common algorithmic thread has been in my career has been graph theory and graph networks.
Speaker A:Even in my digital pathology and AI work I had an algorithm we called a cell cluster graph.
Speaker A:It was actually a graph local graph network that was built on digital pathology where each cell became the node and depending on if it was an epithelial stroma, we created a graph interaction.
Speaker A:And then once you construct a graph, you can have the quantitative metrics outcome from that and you can put it into a machine learning and you can characterize that tumor signature and then you can start to compare against high risk, low risk.
Speaker A:Similarly, in financial world the graph network allowed you to actually start to think about different signals and different things.
Speaker A:In the E commerce space you can think about zip codes as cancer cells and the US map and you can create graph networks and interactions with them.
Speaker A:However, same thing in multi ventures when we invest in a fund.
Speaker A:What I actually did, and I spent a year doing that before launching Modi Ventures, was created a graph network of the venture capital world by mining bunch of data from different places.
Speaker A:And it allowed us to see which funds tend to cluster together.
Speaker A:And so you don't want to go after all of them, you want to maybe unlock that node and find uncorrelated funds.
Speaker A:And so in fact any fund that pitches us today, we actually put that fund and obviously with Pitchbook and everything else, you can actually find the profile of that fund and the features.
Speaker A:It gets plugged into that and it out comes a correlation index of not only to the other funds but also to our existing portfolio.
Speaker A:So that's one way you can actually say this is quite quantitative driven.
Speaker A:Every fund we invested we have a correlation index not only to our own portfolio, but across.
Speaker A:And in fact there's a visual graph to this which at some point I'm going to publish an article on.
Speaker A:So there's a visual graph that actually shows where the correlations are and all.
Speaker A:And again, as thinking Markowitz mindset, you want to be in uncorrelated assets that are, which gives you part of that is diversification.
Speaker A:But diversification you could still be correlated.
Speaker B:So you have one of the most valuable assets, which is the data in the sector.
Speaker B:This is fascinating, Sahil.
Speaker B:What happens when the simulated technicals tell you one thing, but your intuition is pointing in the opposite direction?
Speaker B:Has that happened before?
Speaker A:Yeah, it has, yeah.
Speaker A:That's an excellent, excellent question.
Speaker A:The end of the day you can be, you can try to be as quantitative as you can Venture world is about the few things.
Speaker A:What I say there's a gut check that's involved and that could be that materialize many things.
Speaker A:For example, if you're just investing in a very early stage company, what you end up doing is backing the right team.
Speaker A:The evolution this is not public markets where the fundamentals are available.
Speaker A:So that's why I would say that the quantitative approach to public markets and privates is very different.
Speaker A:Obviously I've not talked about all the proprietary way, you know, you balance it out.
Speaker A:But in early stage you are taking a lot of risk, but you're trying to back the right team.
Speaker A:You gotta meet with them, you gotta get a feel for who they are.
Speaker A:There's a lot of psychological sort of attributes that are involved.
Speaker A:You do have to take a look at what the syndicate looks like.
Speaker A:There's.
Speaker A:But what allows us to be a little bit different is that we don't always have to think about the power law.
Speaker A:That's what I think a lot of funds in venture capital won't have to think about.
Speaker A:For example, Power law says that you invest in 30 companies.
Speaker A:Maybe three or two or three will drive the entire sort of fund profile because most of these companies either will amount to Nothing or maybe 1X.
Speaker A:But if you think about the Met and that by the way, the power law, if you were to plot it, it looks like a spike.
Speaker A:In electrical engineering, we call it impulse response.
Speaker A:It's just a spike and there's a long tail to it.
Speaker A:Private equity buyouts, they don't work on power law.
Speaker A:They work on this sort of non normal distribution, kind of like a two bumped of Gaussian curve which is sort of shifted.
Speaker A:I think about our fund's mathematical distribution in the middle of these two.
Speaker A:But it's not really a power law.
Speaker A:But it's not really non normal where it's private equity.
Speaker A:It's in the mix because we have LP positions, we have directs, we're kind of bending the curve.
Speaker A:So that's another kind of even mathematical distribution of how we think about risk and return is a bit slightly different than a traditional venture capital fund.
Speaker A:But it's not exactly like a buyout.
Speaker A:So just giving.
Speaker A:I'm putting a lot of things out there, but this is all part of how we think about the return profile at the risk level.
Speaker A:In fact, I can make a claim that unlike most venture funds, we can actually put a quantitative number.
Speaker A:Of course these are all projections on what the risk level of the fund is compared to the return.
Speaker A:It's Going to produce.
Speaker A:And it's very difficult for most funds to do because ultimately you make bets and two or three of those will turn out to return the fund and more.
Speaker A:Right.
Speaker A:In our case, we have LP positions, we have royalty funds, we have directs.
Speaker A:Some are late stage, some are early stage.
Speaker A:So, yeah, that's.
Speaker B:So you're calculating Sharpe ratios a lot.
Speaker A:Yeah, yeah.
Speaker A:So there's.
Speaker A:Yeah.
Speaker A:So I think an equivalent of that in the public markets.
Speaker A:Yeah, the Sharpe ratios.
Speaker A:But what we're really going after is that do we have the right mix?
Speaker A:Actually, the best way to think about that is modern portfolio theory allowed wealth managers to always tune your portfolio by saying, let's do 20% in stocks and 80% in bonds.
Speaker A:And that.
Speaker A:That's how we think about in some ways.
Speaker B:So how do you reconcile, since your biomedical investments are highly scientifically nuanced and you go quite deep, how do you reconcile scientific validity versus real world deployment?
Speaker B:Sometimes there's a gap that cannot be bridged.
Speaker B:In some cases, how do you reconcile that?
Speaker A:So let me see if I understand the question correctly.
Speaker B:So, for example, an investment can be scientifically sound and valid.
Speaker B:However, the real world may not be ready for that tool or technology yet.
Speaker B:The first cancer vaccine that was approved, Provenge, several years ago, it was scientifically quite interesting, but operationally was so complex, you know, they used to actually use helicopters to pick up the samples.
Speaker B:And it did not scale, unfortunately, because the real world implementation was very nuanced and not practical.
Speaker B:How do you reconcile that?
Speaker B:And we have technologies today.
Speaker B:For example, there are a lot of cell therapy innovations that work beautifully in the lab, but they don't necessarily translate well into the real world because of manufacturing and a lot of other factors that go beyond the science.
Speaker B:How do you reconcile the two?
Speaker B:Do you think about the real world deployment implementation when you evaluate the science?
Speaker A:Oh, absolutely.
Speaker A:So again, this sort of precision indexing approach allows you to make some bets that typically may not be comfortable to do.
Speaker A:You talked about cell therapy.
Speaker A:We do have an investment, say in a, in a next generation of capset platform.
Speaker A:It's lattice bio, which traditional thinking says, well, you know, cell and gene therapy caps, you know, the viral, viral vectors are out of, out of style.
Speaker A:But this is a new platform way of thinking about that approach.
Speaker A:So it latches onto me.
Speaker A:But at the same time, we can also make a bet on a new way of sort of doing manufacturing in the cell.
Speaker A:So again, a traditional life sciences fund would not go into a tech bio setup that actually is improving how to manufacture these faster and also I think about these as known correlative bets.
Speaker A:So you can make a bet that has early science probably ahead of its time, but also make a bet that is just going to improve the space in the short term period of time.
Speaker A:So that's what allows you to take that layer approach, the biostack layer and so what kind of bets you want to make on that layer.
Speaker A:So it just gives you a kind of a world view of where this convergence is happening.
Speaker A:So that's one way to think about that.
Speaker A:We actually do and we have done that.
Speaker A:The other thing is that remember, unlike a tech company, there's pros and cons.
Speaker A:Where the evolution of a biotech company is that even if you look at the financial rounds we call series A and B and Cs and all in tech, Series A was really a term that was coined to say you got some kind of product market fit in SaaS world for last 12, 15 years, show at least 1 or 3 million ARR and you've reached series A and then everything is arbitrary.
Speaker A:Right.
Speaker A:While we have taken the same sort of funding around structure and fit it into biotech, which doesn't really apply as much.
Speaker A:What does a series A biotech mean?
Speaker A:What it really means in the bio spectrum is you're either filing for an ind, you have certain number of programs.
Speaker A:If it's a platform company, which ones you're advancing, things like that.
Speaker A:I think inherently biotech companies have portfolio and quantitative thinking that needs to happen.
Speaker A:Andrew Lowe has talked about this for a number of years.
Speaker A:Is that the peak sales and the volatility risk of each program that you go after, how are you going to get a buy of bucks, deals?
Speaker A:Is there royalty plays later, non dilutive grants?
Speaker A:Especially if you're in Texas, you have separate all of these things a typical tech company doesn't have to think about.
Speaker A:A tech company is, you know, you build something deep tech, hardly.
Speaker A:You get non dilutive money unless you're in deep tech.
Speaker A:You get some government money, but you're really just talking about distribution channels.
Speaker A:You go out to market and different way of thinking.
Speaker A:Yet the funding rounds we label and we follow the same thing.
Speaker A:So there's a different way of thinking about at some point, hopefully.
Speaker A:So I think that in biotech particularly you could still bet in early science, but part of your as a fund manager or any investor, you have to imagine that maybe it doesn't reach the bedside, maybe it doesn't go production live in any other platform.
Speaker A:But is there an intermediate market where a larger player May need something to build something on top of that.
Speaker A:And that's what we call M&As and things like that.
Speaker A:So you don't necessarily have to rely for, think about that technology to go all the way through.
Speaker A:It may not happen for another 15 years, but you could clearly do that.
Speaker A:One example Solexa, right?
Speaker A:When they had that sort of multiplexing technology on their own.
Speaker A:We were not ready for multiplexing, but they needed Illumina and it was a billion dollar outcome, I think if I'm not mistaken.
Speaker A:And so you do have to make those bets to say is there an intermediate outcome?
Speaker A:Similarly to Grog, Grog had a chip was just quite ahead of its time thinking about language processing units.
Speaker A:But they needed Nvidia to come and say, okay, we'll plug this in now we're going to turn this into $100 billion stream.
Speaker A:So same thing applies I think in biomedicine's tech.
Speaker A:Bio is, can you imagine at some point, at the end of the day you invest to make returns, right?
Speaker A:So that's where it's not quantitative.
Speaker A:To answer your earlier question, it has become thought experiments.
Speaker A:A lot of, I think fund managers would admit that they have to do a lot of thought experiments of what happens with this investment, which is not quantitative.
Speaker B:Right?
Speaker B:Yeah, it makes sense.
Speaker B:I think a lot of the real world implementation nuances probably are driven by intuition plus data because there's so many different factors involved and there are emergent behaviors in an ecosystem that is not always rational, that is the healthcare system.
Speaker A:And Sean, if I may add one something.
Speaker A:So what I've actually thought about how do we do this thought experiment part better?
Speaker A:And also after investing, we're not a big fund, we're approaching 150 million AUM.
Speaker A:But in the grand scheme we're not that big.
Speaker A:But how do we play a bigger role in some of the companies we do invest in?
Speaker A:So I started to think like a drug developer.
Speaker A:So whether you're a drug development company yourself, you have a business development, BDS and partnership capabilities.
Speaker A:So I actually ended up getting some folks on board who actually specialize in BD of biotech.
Speaker A:And so you think about what is a VC firm doing hiring some of the BD guys.
Speaker A:So we can actually not only when we're investing, so we can think about that as an outcome and they can help do back end diligence with the former relationship with Biopharma.
Speaker A:But once we invest, we're actually starting to work with these companies to help them commercialize, get those partnerships with Biopharma get those calls and so we can play a role even though we're small.
Speaker A:So I think about our, even internally as a fund a layer of folks who are just BD guys.
Speaker B:Interesting.
Speaker B:There's a lot of hype, Sayuri, about AI right now, especially when it comes to biomedicine.
Speaker B:What is the threat of truth that's running through the hype today?
Speaker B:What do you think we're doing right today?
Speaker B:What types of tools and technologies and companies are making you excited?
Speaker B:And where do you think the field is going to be in five years?
Speaker B:Specifically the application of AI in advancing both drug discovery and development?
Speaker A:Right.
Speaker A:I mean, I think at this point no one doubts that AI is here to stay in life sciences pretty much anything.
Speaker A:I mean, I can't imagine a single endeavor of human life that AI is not going to play any role.
Speaker A:In some places it's going to be slow.
Speaker A:Some will.
Speaker A:Obviously in programming we've seen, you know, who's coding.
Speaker A:Actually it's through AI assisted.
Speaker A:So in life sciences we're starting to see obviously in documentation, obviously even FDA has talked about, you had a fantastic call yesterday that they're thinking about what role just in the reviewing the process and giving these tools to reviewers.
Speaker A:Those are obviously the first layer when you think about in drug discovery, obviously discovery process and finding the hits and all of that.
Speaker A:Over the last 12 years we've started to see some evolution of that.
Speaker A:And now of course, one of our own companies and so tremendous amounts of data has been generated in wet labs and imaging the cells and all of that.
Speaker A:When it comes to medicine, we have tremendous amounts of data, although siloed.
Speaker A:And so that's a different part of the discussion is that we have collected disease information.
Speaker A:We have collected some interesting data about the human journey in this process.
Speaker A:I know we've all primarily always relied on claims data, but we have now biomedical data available from medical imaging to sequencing data, particularly in oncology.
Speaker A:We have now rise of liquid biopsies, mrds, longitudinal.
Speaker A:These are nonlinear sort of approaches to monitoring detection, prognostication.
Speaker A:So I'm very excited about that.
Speaker A:Digital biomarkers have roles to play.
Speaker A:FDA has started to think about how to prove this.
Speaker A:You know, Arterra AI, it's not a portfolio company, but it's worth mentioning, you know, in prostate AI, it's now part of standard of care.
Speaker A:It's the first AI algorithm that's part of it.
Speaker A:So obviously they were the trailblazers.
Speaker A:You know, another example is Oncotype.
Speaker A:DX has been a great tool for 25 years, but it requires a tissue sample to be shipped to a lab.
Speaker A:Takes two weeks of turnaround.
Speaker A:Used to be 5,000, but I think with now reimbursements it's maybe thousand or two thousand.
Speaker A:And yet the accuracy is it's not as we want for a prognostication in breast cancer.
Speaker A:Why can't we imagine a world where a digital pathology which is now digitized, can run on a cloud.
Speaker A:It can do that in matter of minutes and by the way, you preserve the tissue and it can go beyond all of that.
Speaker A:Right.
Speaker A:So I am investing in those sort of platform technologies that can imagine the future, next five to six years, yet they have to work with the market today and things like that.
Speaker A:So just giving you some view that I think that's where the real opportunity for AI is.
Speaker A:The other opportunity is that, by the way, look at what's happening in healthcare as well.
Speaker A:One of the, I would say second or third most killer applications of AI is knowledge generation and copilot.
Speaker A:So like Open Evidence is one example where 50, 60% of all physicians trust OpenAI and they use it on some basis to validate and all of that.
Speaker A:But also ambient AI.
Speaker A:For the first time, the interaction between a physician and a patient was not just mumbled notes that a doctor decided to put at the end of the meeting just so that they can reimburse.
Speaker A:This is now a proper structured interaction.
Speaker A:Imagine the kind of data we're going to actually learn from that interaction and what usefulness is going to come out of that.
Speaker A:I'm not excited about ambient just for revenue ops, by the way.
Speaker A:Oh, this coding, this coding I'm more excited about now for the first time, you actually have interactions captured between a physician.
Speaker A:And I mean of course we need to think about HIPAA and all of that, but that dataset never existed before.
Speaker A:Never.
Speaker B:Absolutely.
Speaker B:You know, one of the ambient scribe companies, I was talking to the CEO recently and he said something very fascinating, that there's a lot of data that's surfacing in their database.
Speaker B:Patients telling the physician really important things about their health that never makes it into the ehr, because the way that physicians think is very structured and orderly.
Speaker B:You know, we have the history of present illness, literally a blueprint for how to take notes and capture information.
Speaker B:Or the SOAP note.
Speaker A:Yeah, yeah, the SOAP note.
Speaker B:Yes, right.
Speaker B:For the subsequent visits.
Speaker B:So things that fall outside of that, even if the patient is mentioning it, may never make it in the electronic health record.
Speaker B:So he's been thinking about doing something with that data which never, it's basically just percolating in the air and now he's capturing it through ambient recording.
Speaker B:So I completely agree.
Speaker B:I think ambient recording is going to do two things.
Speaker B:First, it's going to make our data in electronic health records a lot better.
Speaker B:So finally, I think there's going to be a new generation of EHR vendors and companies, aggregators that will fulfill the promise of the learning health system that the first generation of companies some argue did not fulfill and failed for the most part.
Speaker B:And also it's going to surface these new insights that never make it to the EHR in the first place.
Speaker B:But you really talk and focus on what is actually going on with the patient at the intrinsic level.
Speaker B:Very interesting.
Speaker B:Well, Sahir, we're almost unfortunately out of time.
Speaker B:We could talk for another hour I think.
Speaker B:But I have a couple of final questions.
Speaker B:One is so Modi Ventures, you mentioned that you have about 150 million AUM.
Speaker B:Where do you want Modi ventures to be in 10 to 15 years?
Speaker B:And where do you think we are going to be in the context of your bio stack in 10 and 15 years?
Speaker B:What will the work look like?
Speaker A:Yeah, so the first thing I would imagine that this sort of approach in private markets, I'm hoping it would do what index funds did to public markets.
Speaker A:It allows an opportunity for us to really capture there's great innovation that's happening and allow us to fund innovation that takes a systems level approach and an asset class approach rather than just disease based, indication based or modality based.
Speaker A:So it's allowed us to think about it at a systems level.
Speaker A:So hopefully we can, as we become bigger, we can fund technologies and therapies and things like that that traditionally wouldn't get funded.
Speaker A:And also because of the ecosystem we're starting to build with different LP positions, the royalty funds and all, we will play hopefully in next coming years a deeper role in enabling an ecosystem of companies and all of that in a different way where we will be stakeholders in so many different areas that we would be able to connect the dots that say a fund itself may not be able to do so.
Speaker A:That's kind of my view of where Modi Ventures will be.
Speaker A:So hopefully this is a new way of thinking about everyone is doing the right thing and making the right impact.
Speaker A:Particularly if a biotech investor you have, obviously you want to make money but there's a big impact part you are enabling, saving lives.
Speaker A:At the end of the day, some of these technologies will go on to save their lives.
Speaker A:So I could say that in the bio World and biomedicine's investment, there's a huge impact aspect to it.
Speaker A:So obviously one of the personal goals of mine is that how do we enable saving lives and improve the human conditions and while also make money.
Speaker A:Right.
Speaker A:That's the overall punchline.
Speaker A:So new ways of thinking and the biostack.
Speaker A:I'm hoping that in the realm of what I was talking about earlier is that if we can manipulate the cells and cells are lost and cells malfunctioning, we can address some other diseases that we just don't have cures for and we're starting to think about curative stuff.
Speaker A:But the cures don't just happen even if we have some tools.
Speaker A:But it's all built through the stack.
Speaker A:Just like AI didn't just happen overnight.
Speaker A:It needed the data collection mechanism, it needed software evolution, it needed consumer technologies, it needed a bunch of things to happen to converge through that stack.
Speaker A:It's the same way I think about where the stack starts to go towards the read, write and reprogram.
Speaker A:We're starting to actually reverse diseases.
Speaker A:We're starting to come up with new economical models of treating prevention.
Speaker A:Prevention is a big one because remember prevention can't happen if you don't measure well, if you don't know the outcomes, if you don't have consumer mindset of how do we take some of these technologies and consumer.
Speaker A:By the way I forgot to mention we at the organism level, unlike other life sciences funds, we also invest in consumer tech.
Speaker A:For example, one of our really well performing company is a sleep buds called Oslo.
Speaker A:They have now sold 150,000 plus devices.
Speaker A:It lets you sleep better.
Speaker A:It's next generation is going to have a PPG in there.
Speaker A:You know, they can, they can turn into a medical way of treating tinitis.
Speaker A:This is a new platform of thinking about sleep.
Speaker A:To me that is part of engineering the human life and the condition at the organism level.
Speaker A:At that biostack.
Speaker A:You see traditional life sciences would have nothing to do with this company.
Speaker A:Right?
Speaker B:Yep.
Speaker B:It's perfect.
Speaker A:Yeah.
Speaker A:So ultimately what I imagine next 10, 15 years, we're starting to engineer life, engineer life through this biostack concept and we're able to fund that and we're able to have a bigger impact to I guess human life.
Speaker A:That's the ultimate goal.
Speaker B:Yeah.
Speaker B:And it seems like you're using finance and financial engineering remarkably well to have system wide impact.
Speaker B:Reminds me of what our mutual colleague Andrew Lowe always says that don't fight cancer but put a price on its head.
Speaker B:You might, you might be more successful makes perfect sense.
Speaker B:So here.
Speaker B:This was a fantastic conversation.
Speaker B:Thank you for your time and thank you for joining.
Speaker A:No, no, thank you for having me.
Speaker A:Me on this was.
Speaker A:I really enjoyed this.
