Raising a venture fund in 30 days - Sahir Ali (Modi Ventures)

Sahir Ali is a the founding general partner of Modi Ventures, an early stage venture fund investing in tech-bio startups. He is also an LP in multiple funds including Draper Associates and Khosla Ventures.

He is an accomplished entrepreneur and investor. We discuss his path to launching a venture fund, founder selection, decision making and more!

Transcript

 Uh, there were deals that were coming and I'll be honest to our family office. Hey, you have three days to respond. No time for diligence, either you're in or out. There's so much money, but that is not happening. So from just macro economics perspective, it just felt like an absolute right time to come in.

And what was

your answer to those deals, sir?

Oh, no, I just, you know, fundamentally. No matter how good the team is or or perhaps the product is superficially or has captured the so called the network effects got to do your diligence.

Thanks so much for joining me today, Sair. It's amazing to have you here.

To start, if you could give us a brief introduction and then we'll get into

it. Yeah, thanks for having me. Uh, so my name is Sair Ali. I am. Uh, an entrepreneur. I've been an academic for a bit, and now I have a venture fund. Also, along with my brother, we have a single family office as well. Uh, been always sort of inclined towards, uh, life sciences, biotech, healthcare investments, but we've generally done lots of different investments as well.

Uh, we've been operators as well. Uh, in entrepreneurial capacity and, uh, my previously some academic work involved pathology and AI at the intersection of AI and oncology as well. And I've also spent significant time in the enterprise technology world where. I was very early on Salesforce architect and I've worked with Fortune 100 companies incorporating not only Salesforce, but, uh, early on, very early on, uh, incorporation of machine learning.

So we were doing some propensity modeling and also overall, uh, I've, I've seen a different things in last 10 years from the academic labs to entrepreneurial journey to, um, being in the enterprise setup of large companies. From also the investments and and sort of the family office world as well. So happy to um be here and thanks for having me Perfect.

Let's talk about why you decided to launch your own fund. Tell me how that came to fruition and What are some learnings you've had from talking to lps and raising your fund?

Yes, uh, you know, I think If, if you had asked me about, about a year and a half ago, or even sometime last year, that if I wanted to be a fund manager, my answer would have been no. But last year was very interesting from market perspective last year, by the time November ish came around, which is where I got really serious about launching a fund, there was about 88 percent decline in funding for venture, uh, venture funds.

There was even a further decline in deals that were happening. And post series a, there were, there was a 32 percent decline in a new, new funds coming up. And in fact, interestingly, there was an uptake of about 15 percent in billion dollar funds being closed. So, which was kind of a counterintuitive, actually, when I saw that report, I said, what's going on here?

Uh, they're everyone's struggling yet. There are billion dollar funds being closed last year. And the reason was everybody's toast. took a step back and said, if we are to invest in funds or venture ecosystem, we'd rather go with established players. And so that was one, uh, one sort of motivating factor that I think what this means is all the exuberant valuations and the interest rates going down and the money that was freely available has now tightened up.

This is the right time for, uh, for investors to come in because it becomes what happens is now it becomes investors market. And that's where you find best bang for your buck. Uh, there's a lot of opportunities that may not be open to, um, smaller funds. Uh, there were deals that were coming and I'll be honest to our family office.

Hey, you have three days to respond. No time for diligence, either you're in or out, there's so much money, but that is not happening. So from just macro economics perspective, it just felt like an absolute right time to come in. And what was your

answer to those deals, Sair?

Oh, no, I just, you know, fundamentally, no matter how good the team is, or, or perhaps the product is superficially or has captured the so called the network effects.

Got to do your diligence. Um, and you have to understand based on what your risk appetite is. Not every, not every investment can, has to be based on, um, how good the company is doing. Sometimes a really, really good company you may have to pass on and it depends on your risk. Tight what you're looking for.

Is it going to be a long term play? Do you see a short liquidity? I think so We we come from a background where uh, we haven't Prior to this hadn't dealt a lot with illiquidity and so that's a different mindset And so everybody has different mindset and particularly those in those investments. We were more like a Angel investments rather than, you know, the professional money in.

So that's a different aspect to it. But yes, the answer was typically no, that this is too fast for us. And there's nuance

there. Say you had an opportunity to invest in Adam Newman's new company that Andreessen put in 300 million. Would you say yes, if you had a day for diligence, or would you still say no?

So let's, I think the number is important to keep in mind. If someone's putting 300 million in a company already, and

That's why I picked that one, because the checks are substantial.

I also mean someone's asymmetrically excited about something and it's very seldom you've seen that that kind of exuberance without a product and that kind of money going in yields a lot of interesting things.

It just becomes typically one of the, especially at. After hearing about a few cases, but particularly what comes to mind is Elizabeth Holmes, which kind of is how the initial, there were just extreme exuberance about someone rather than a product, let's start with that. And so when this becomes about somebody, and then the check size just goes into some.

Buddy, that's, that's usually not, not the way I think about investing. I think it needs to, I need to understand what the product is. I need to be able to see things where things are going to pan out. And I, then I'll take my risk, right? Ultimately, everything is our bets. And so. I think in the example, you just said two, I would see two problems with that.

First of all, for, for sort of smaller investors who could write, you know, um, up to a 2 million kind of check sizes, you are eventually going to be pushed out anyway, in, in next rounds and dilutions and all, unless you have the capacity to follow on those big rounds. But beyond that, I just wasn't clear what the product was going to be.

It's yeah. It's just that it was another real estate play, but really it just, it just seemed like that bets were made on an individual and I tend not to do that.

Is that something very interesting there is you, it seems like index the product more than the founder and correct me if I'm wrong there and you're a biotech investor.

The product could be very difficult to understand. I agree with that by the way, I take my time with understanding the product. A company would just invested in her and biosciences, you know, it took me. A few tens of hours to understand exactly their product, but it's necessary for me. Um, is it necessary for you to understand the product?

And then how do you look at the founding team versus the product versus the problem they're solving when you're evaluating

startups? So let me backtrack and say something that. When I meant understanding the product, that's when 300 million are dollars are going into something at that point, the better have a product, which has a right fit to the market.

So in that sense, yes, I need to fully understand it hard. Then at that point, yeah, I'm assuming when that kind of money is going in, there's already a good team and all of that, then the product becomes a big deal because now if 300 million is going in. And what valuation God knows, but to cap, to, to now go two or three X, how much momentum of the market you're going to have to capture.

That's why product becomes important. But now if you're going to ask, say, okay, well, how do I look at companies that are early stage, which are at the vision state? Of course, the, the, who the founder is, the team, that's, that's probably the most important aspect as well, because. Especially if you're doing pre, pre seed and seed, you know, I haven't, they haven't really found the product market fit or so called product market fit.

So it does become about these multifaceted things found the team is important to who the founders are. Uh, what's the, what's even the, uh, dynamics within the founding team? Because as we know, most startups early on failed because. They just can't gel together. The, the vision of the product, how big is the market?

That's a typical thing everybody looks for. And, uh, and how they're going to approach their initial in that, in that early stage, I like to understand how do you plan to go to market? What's the MVP? Yes, you have a big enough vision. That's probably the exciting part, but do you have small enough focus to get that out?

Right? So those are things I'd like to understand because a lot of times founders do tend to get. bogged down in the bigger vision and you can't get anything out, right? And then it has to be an iterative process. And it also depends, right? I'm talking mostly SAS. Now, if you come on the biotech side, it's a different ballgame completely, right?

You, uh, in biotech, uh, it's, it's a lot to do with what kind of assets you have and, and, and what's the regulatory pathways. If it's med tech, it's different. If it's going to be a small molecule, it's different. One area. Yeah, that I particularly like, which I think is a lot more investor friendly is tech bio where it's, it's sort of AI enabling, or I guess, um, AI enabling or technology enabling something, um, in bio health and medicine.

So yeah, discovery would be that. AI, radiology, pathology, uh, any, anytime there's a platform play, so you could start to apply the usual known concepts in the technology to these kind of setups. But yeah, I agree, biotech is a very different beast, but in there we have a team of folks who specialize in certain aspects of those areas as PhDs previously, and so they do the deep diligence as well.

And then you fully understand what that is, and the risk over there would be You know, you break it down by, is this going to go to the, uh, human trials? Can it, if it's in human trials, what's the probability of going into the next one? Is there an MNA in play? Who are some of the potential customers?

Different sort of areas that you look at versus say a technology company, I guess, in that stage. Are

you focused on a specific area of biotech, say diagnostics versus therapeutics? or radiology, cardiology, oncology, or are you looking at it as a whole? And if so, let's talk a bit about portfolio construction because exits and the need for commercialization to an extent can vary depending on what area of biotech you're looking at.

Yeah, I, I, I like the tech biospace rather than the biotech for, for simply two reasons. And I think tech bio evokes a different definition. Many folks, what I like to call that is These are companies that are operating like technology companies, but disrupting something in health, bio, and medicine. And so the, the most prevalent example that anybody will give on TechBio is AI drug discovery.

Yeah, AI drug discovery has been around for, at least got a lot of money in last five or six or seven years. A couple did go to public as well, but we haven't seen a blockbuster come out of that. But I think these kinds of things take time and we're still waiting for a blockbuster success. But that doesn't mean that early investors didn't make some money.

So that's one of separate that out. I think we tend to look at success in terms of if this product came to the consumers, but for early investors, sometimes getting an acquisition or I, I, you know, I'd IPO also means an exit. So that's one aspect. The other, you know, if you look at Moderna, Moderna, yes, you can classify as a biotech company, but so it was actually a platform company.

It's a platform company for taking mRNA in different directions or different, um, capably therapeutics, uh, vaccines, uh, they can tackle infections and things like that. And so it's a platform company. This is why within just seven or eight years, it was, it IPO'd at 8 billion, I believe, which is kind of a very short path.

For typical biotech company without any approvals, without, you know, going through any, having a proper asset in the market. So that's a platform company. I like, I like to invest in platform companies that have, you know, as they call it, the multiple shots on goal, where you're, you're risk, you're, you're diversified your risk with not just not pursuing one asset, but you can create kind of multiple assets.

And then the interesting thing is, how do you make money? There's this multiple ways of making money because farm there's, you're. Pharmaceuticals are your typical customers, they will pay up front, there will be some royalties, there will be some milestone based payments. So, so different kind of, uh, area of how you evaluate that.

But if you're a typical biotech, which has single asset, it's, it's very risky, and it takes a very long time. So unless that. But on the therapeutic side, unless the, unless the drug, the molecule is in some kind of human trial, I kind of don't, I don't have the risk appetite to go in early than that.

I was listening to Vinod Khosla on Harry Stebbings and he said he's excited about single gene mutation for drug discovery in biotech.

Tell me your thoughts on single gene mutation therapeutics. What's missing in biotech? You know, we have a fairly good estimate of protein folding. Um, to an extent of genomics as well. Why haven't we seen an AI drug discovery platform really take off? Um, do we just know that don't have the knowledge of disease process or pathology or pathophysiology?

Or is it something else?

I think you're you're you're probably thinking in the right way as I would is that you look at. So if you look at the workflow of taking a drug out to market, right? There's definitely the discovery part where. just to find those right bindings and the structures can be someone's entire PhD, right?

But once you've done that, then you've got to do this, you know, the preclinical studies and I think So in some ways we're also making a lot of strides. So AI drug discovery is helping finding shortening that route, at least the discovery path quickly. We even have some setup, which helps simulate those mouse models and things like that.

But real deal starts when it goes into human trials and this is still a wild West. The toxicity, the FX season. And so those. Those are hard to predict. And so look at what happened to benevolent, right? Uh, recursion, which had 35 assets. So this is still, they still have to kind of go into the real world.

And this is where things are not progressing well. So I, and I think there are some technological advances that will help, but Ultimately, AI discovery started with a big promise that it not only would we can find these targets, but it can find these precise targets that could actually accelerate and and go through the pipeline quickly.

We haven't seen that, but remember, these are still early days. I mean, if you think about the advances in deep learning, they've just come, the core advances have just come in the last five or six years. Especially the initial kind of the innovations in deep learning. Well, of course, let's start with that neural networks have been around for decades and decades, but the deep.

The proof in the pudding was shown in the computer vision space, right? The, the, the convolution neural network kind of came out and said, look, you can actually bring your networks back and make a deep, and so a lot of that was pixel and convolution based, but then the sequences, which NLP kind of redefined the whole thing with transformers was just in 2016 and 17 time period.

So these are still the early days. So I think we'll probably see a lot more progress, but a lot of money went into AI discovery, that's the only thing, right? So, and, but also market conditions are favorable. There was a lot of quote unquote, free money there, low interest rates. And so now what's, what's happening is obviously.

The money situation is tightened up. Now, investors are actually really saying, okay, well, let us evaluate what really are you working on? Yes, you can, you may be able to find these targets, but really it's going to become about how do you, is there a reliable partnership with pharma? They will want to see that up front.

Previously, they took a lot of bets that, okay, you'll go on and and get these partnerships out. But now can you. Actually, they want to see what's the tangible path they want to see. Do you have a commercialization officer? Typically, that was an afterthought in in platform companies. Do you have a drug hunter on the team?

Do you have these? They might even need some, uh, testimonials and some referrals to potential partners. So they're. They're starting to think about how, what and what things happen beyond it's been discovered and go potentially going into, uh, human trials. So I think that's probably the, the issue and there are now, um, like, for example, actually, before I continue, there are AI discovered drugs that are in phase.

Right. So until like something really makes it out, um, we will still wait for that kind of a blockbuster, but that doesn't mean the innovation is going to stop. I think the money will continue to be in this space. There will be acquisitions. There will be M and A's. But I think if you're asking the question of when are we going to see this kind of a blockbuster success?

I don't know, but I am hoping we will. But early days, though, if you look at the field overall, it's quite new.

I agree with that. And thanks for the well thought out and informed answer. Let's go back to raising a fund. And specifically, let's talk about LP relationships. When did you start developing these relationships and how important were they and your track record when going out to LPs?

Tell me about how many of your LPs you already knew, how many were warm intros and if, if any, how many were cold outreaches or maybe some were inbound. Um, so I'd love to

go deeper there. Yeah, it would be important to understand that. Let's start with it. It's this fund of 32 million was actually raised in about a month or a month and a half.

So, in for the conditions of the market, this was in some sense, uh, uh, that sounds

almost unbelievable. So, yeah,

it's almost pretty unbelievable. And there is, there are a couple of reasons for that. Is that there was, there was a strategic move on my part to look at how do we actually do this fund? And so in this time period, first of all, if anybody is to raise the fund, no matter how of a brand name they are, I'm going to start to think about and rethink their strategy a little bit, because The market conditions are bad.

And so for me, I said, if, if we're going to enter this space, the first of all, the strategy has to be solid. But beyond that, it, it needs to be, it's an investor's market. So it has to be aligned with the values of investors, which are the LPs here. And so what does that mean? It has to mean that I would perhaps have to show skin in the game, which I have quite a bit.

Beyond just the typical, uh, we're going to put 1 percent or 2 percent and, and probably towards the end takeoff fees. So, so first is without diverging details, the structure of the fund is very much aligned, uh, with, with the investors. And most of their, most of the investments that they do goes in actual investments.

So what that means is, uh, what I see as my upside is when the fund does well, not in the fees. And so that's been a very different approach and, and the other was a strategy in this time period. When I look at the ecosystem, there's, there are great startups to invest in, but at the same time, there will be really great fund managers to also invest in, which, which are iterating in, in some of the areas that I like, for example, oncology, life sciences, artificial intelligence, and so.

Some of these funds may have been kind of not available to family office oriented folks, and they would just go to their institutional investors who could write just 10, 20 million dollar check sizes. And so, with that, our strategy as a private fund is to invest in both. And this strategy is based on.

The fact that a typical VC has to continuously look for inbounds and evaluate them and has a large team have to go to events constantly. And this is how you kind of get the deal flow for us. The deal flow also comes from being LP and other funds. And, and, and, and the benefit of that is these funds are pretty tough.

Up notch and these fund managers are, are bring not only in some cases, just experience over a running couple of funds for, for last five, six years, but some of them have seen through a couple of financial cycles too, which is a very important skill. I think that's missing in a lot of GPs right now. Is that because they haven't, including myself, haven't seen.

The financial cycles, right? Of dot com than the financial crisis of 2008. And it's a very different experience. While a lot of funds have done well in last 10 years, not many GPs can say they've gone through the ups and downs of, of the cycles. One of the reasons why we decided to become LPs in coastal ventures is that, you know, Vinod has been around and seen these cycles of exuberance, not only that, but financial ups and downs.

And so last was also track record. You know, um, In this fund, my brother's involved as well. You know, we've been involved with building companies. We've been successful ourselves as entrepreneurs. We've been on. Operators. We've also done investments. We have our own kind of track record in investment world.

And so it was easy to see and follow that for folks who decided to come in. And, and the question about, did we do any cold outreaches? No, this was all just kind of in our network. Um, folks that have either known us and followed my trajectory, um, since I, I guess, left for college, or these are friends of friends who initially started believing in this fund and brought other friends,

Amoeba your decision making framework.

How do you decide what projects to take on in your life? And let's go back to college. Um, let's go back to post graduation and what are you thinking? What opportunities did you have or you were chasing and how did you decide what to do? Yeah,

I, you know, I think it's a, it's a, it's a pretty good question and something that not often has been asked.

So I'll say that the term chasing I was on. I was always chasing through two things and that's since high school. I don't think it was a strategy or some brilliant move. It's just something that's how I was always inclined towards. And that's based on the economic, socioeconomic status that I come from and the, the background, which we don't need to get into.

But what eventually that made me is that I was very inclined towards getting asymmetrical knowledge and asymmetrical access. And what I mean by that is how do I gain some knowledge that will put me on top? Then. My peers, and so I was always in pursuit of that. And so in high school, as you know, as an immigrant who barely could speak English to me, the way to stand out was had to have some, something else that, you know, while I was trying to have a cultural fit, it was technology.

It was, it was programming. I became in fact, a very solid programmer by age 14. I was just full C Programmer did some work at NASA as well, because it was right next to us in Houston. And so that was a symmetrical knowledge while everybody else didn't kind of have that and asymmetrical access would lead to that.

So sometimes these are kind of hand in hand. Sometimes if you have a symmetrical knowledge will get you any access and sometimes having that is medical access will get you that. So I guess that was always a pursuit when I came to college. That mentality and being able to like, look for that got me into very early on computer vision machine, computer vision, machine learning and quantitative trading, which led me to become an intern at a high profile fund, um, in Citi, which was at the time reporting to Vikram Pandit that the fund was.

And so that was a pretty phenomenal, uh, access. In fact, um, my internship was very high profile, highly paid. And that at that time, I also picked up on Salesforce. Remember in my earlier introduction, I mentioned how I was the enterprise world. I became one of the very youngest architects of Salesforce very early while I was just an undergrad because Salesforce was new at the time.

And so again, these things I've always pursued that I was my, my passion. And, and I guess golden at the time was not a job or a certain level of income. I've always pursued these two things and what serendipitously has happened is that I did get the right jobs, I did get a lot of financial success by pursuing these two things, so I don't know if that helps answer the question.

In fact, when I decided to realize that in the middle of my undergrad, by the way, I had a full time job at Wall Street and so I was on campus. Perhaps I need to look at something else beyond just a financial world and being a quant. And so this was after the financial crisis kind of got disheartened. I was only, you know, 19 years old.

So I said, what could be some other field that. Maybe quant or quantitative, but a little bit ahead of his time. Again, I need to be asymmetrical in terms of knowledge. And so this is how I landed on working in the pathology and AI world. If you look at digital pathology in 2009 ish, 10 ish period was still very new, very niche area.

And now add some AI to it. There was very few people working on that. So, so that's my hope right now is the similar strategy. Now, if you go back to the fund, I guess a good venture cap or VC or investment. comes at the nexus of these two things, asymmetrical knowledge and asymmetrical access.

Let's talk about portfolio construction.

I'd love to get deeper later on into how you look at wealth and money and power and status, but let's, let's talk about portfolio construction first. So you have a 32 million fund. How many checks, what size are you targeting in ownership? How many are you falling on? Um, and how are you thinking about risk and reward in terms of the stage of the company?

Um, when you think about portfolio construction,

yeah, I think so. The portfolio construction for me is a little bit different than a typical venture fund. Like I said, we're more of a private fund in that one part of the portfolio is constructing a portfolio of other funds. And so I'll first talk about that and then I'll come on all the direct investments

we do.

And if you could kind of. Give a rough estimate or percent on how much in each. And how do you think about the tension? And if there isn't any, just correct me of competing with the funds you're investing in, um, for deals and maybe at seed stage is more collaborative. And as your fund grows. In later stages becomes more competitive, but just I'd love your thoughts on that.

Yeah. No, I think that's a good question I don't see as a competition fact is quite collaborative because my focus as a fund that is kind of in the fund one stage is Is again, I operate on those two principles access and knowledge and and so when you think about constructing a portfolio of funds, there are these areas that I would like further exposure and And alpha ultimately what I care about are returns.

We're in financial world, right? I mean, yes, we can say everything about impactful technologies and everything else, but we are in the business of making money eventually returned. And so from that perspective, partnering with fund managers, particularly in this time period, is very beneficial because the fees have come down and, and everyone's kind of taken a step back.

So everyone's kind of back to the drawing board and really thinking hard about their strategies and such. So for me, I have, I think of that portfolio in three buckets. One, our established fund managers. Who have been who are on their 8th fund or 9th or 10th or beyond, then you have the sort of the mid to emerging kind of managers who are on their 3rd, 4th fund.

Um, and then the last small bucket is 1st time or very niche focused managers. Right? And so in the established ones need to understand how they're thinking about. Deploying large amounts of capital there because they're these are folks who are raising half a billion dollars plus, right? Uh, or even more.

And so, so there's some strategies there. And for me, what benefits do they give us? I think with those funds, it's just about really alpha. Um, and sometimes if because the LP base there is quite similar, it's mostly institutional. Someone like, yeah. Us, um, tend to have some advantage in terms of moving very quickly if they have co investments and things like that.

Or sometimes, and I'll give you already an example, we are going to be co investing directly on the cap table with, because these deals were brought to us by some of these managers. So to answer your question though, in fact, what I can tell you is that as an institutional investor, uh, at a pension fund, endowment funds can't answer co investments directly.

First, they don't have the capability of diligence. Second, everything runs through an investment committee and it takes time. It's an, it's a corporate thing. It's, it takes a while to move while someone like us, we have the capability diligence, especially fits in the area of what we like and we can move quickly.

And then this time period, again, it's a very powerful thing to bring down table. And so we become very strategic. And so that's, it's a very collaborative thing. And then, so in the middle bucket, what I said, which is the kind of mid to emerging manager, these are the folks that. Are focused on some interesting areas, like, for example, tech bio, uh, there's a fund that's.

Completely focused on oncology. Um, you know, and then I just won't name names just yet, but you can read about, um, that, um, in some of the PRS, but these are funds that just bring a lot of capabilities from prior, um, iterations and they're really focused on certain areas. So that check marks certain, uh, areas that I'm very excited about.

And I, I believe in the next 10 years, because remember, we think about 10 year cycle could bring them. Good returns and have impactful technologies. And the last one is funds that may be operating below 30 million fund managers. And as you know, some of you are some of the folks that are listening to this may have seen cameras reports.

These are the funds if they're successful, can return 678 X. Right? And so that's how I've constructed that. So one part of that is You play it safe with the established fund managers. One part kind of get a diversified, uh, portfolio in different areas you want exposure to. And the last one is little risky.

I mean, actually quite risky with first time, but the upside is too high. And so that's how I think about that. Now, if you come on direct investments, I have, we have our own deal flow for sure, but the deal flow that also comes from this kind of being an LP, it's actually very, it's. It's incredible because first of all, these are deals that have been, you can piggyback on some level of diligence.

You're also understanding that some of these funds are going to make sure these companies succeed in having next capitals and things like that. And so that's one, there's a deal flow where I like to lead. And so that means I'm on the hook for making sure that the founder is successful, make sure that, um, that not only from the.

From just being able to, uh, work with the founder and getting to make sure they meet the milestones and all, but more importantly, getting an ecosystem excited about them. And so guess where we are sitting, all these funds they're invested in can event, we can also bring them on to deals that we find. And that's, that's a very important, uh, strategic sort of advantage than saying, knowing somebody at a fund versus being an LP at a fund.

That's a very interesting portfolio construction. One I haven't come across before.

Also remember, in, to construct a portfolio like that, you have to rethink fees. I'll just leave it at that. I think if a typical 220 model, uh, And I don't think the 20 percent or whatever that carried interest is is the issue because that only comes into play if the fund is going to be Start to return money, but really the typical fee structure starts to become a problem is that it's a two percent let's just keep it for Um argument sake the two percent if someone decides to invest a hundred thousand dollars It's not a hundred thousand that gets invested.

It's really eighty thousand after the you know Everything fees set in on eighty or eighty five thousand depending on how it tapers off. So right off the bill But you start with minus 15, 000. And so the, that's why you have the J curves and things like that. And that what now what's, what has happened in the venture world in last 10 years with the Cambridge report and all is that there hasn't been enough actual cash being returned.

Yes, there's a lot of these paper values of, um, you know, the fund is at this X multiple and all, but from LP, one of the biggest complaints that LPs have is that where is it? If I put in a dollar, am I getting a dollar back in my pocket?

And so fees has one role to play because a lot of times the distributions there's liquidations and distributions sometimes GPs do hold on like to hold on to that and so there's a lot of things that are reshuffled I think in this market conditions a lot of folks are asking for DPIs and so this is also a good time.

At the day my strategy is that on the fund of funds kind of Mindset, you're looking for a non normal kind of distribution with, with, with, um, with direct investments. Yes, you're going to need to construct a portfolio of many companies. And so that's where it kind of. Uh, the power law starts to come in obviously some most of them may fail, but there will be some quote unquote fund returners or or Big bets in the portfolio.

So that's where Once that I have high convictions, of course, no no company you will say I have 100 conviction But extremely high level convictions We like to keep ownership and that will depend on initial stage of the company if a seed, you know Start with it at least having some ownership that we can maintain in And, um, over the course of that company, uh, if, if, if we're leading, then of course, this is going to be very important for us to take our pro ratas and things like that.

So the typical, uh, venture sort of, uh, math and operations come into play there.

Okay. I'll ask a very selfish question. I'm not raising a fund right now, but it's something I'm thinking about in the future. What is the ideal emerging manager to you? What characteristics do they have? What's their background?

Um, what's their deal flow? What's their value add? What's their diligence process? And what's their track record? You know, an ideal answer is I've invested in 20 companies. I have a 10 extra in grade, but I think There's a lot of emerging managers who don't have that track record. How can they get your attention and how will you get to conviction in them?

What do they need to show you without, uh, amazing track record maybe because they're just new in this field? Um,

well, I would, I would, what will get my attention is I've been repeating both of these things, asymmetrical knowledge, asymmetrical access. What I see with a lot of times folks who are raising new fund or first fund. And I've seen at this point about 80 or 90 funds in the last six months, you know, in terms of obviously I get pitched, um, from the fund side a lot as well, right.

Cause we were right in the middle of the ecosystem. And if, if there isn't a track record, say being a principal at other fund and you've done that ones, you can show, look, I led these deals. That's always easy to kind of imagine that, okay, I can take a leap of faith that. Being at a different fund now, you're ready to kind of spin out, but someone with that doesn't have that kind of background.

So then you have to bring something in these two areas. What is that asymmetrical or some deep knowledge that no one else has? And what does that access look like? And an example of that could be, Hey, I was at, um, this, uh, prime base editing gene editing lab while there's only two labs like that in the world.

Well, I'm one of the postdocs or was, and based on that, one of the companies that is now out there and doing very well, and I was somehow involved with that. I think in that area, and then you kind of start to have a thesis that's formed around apes, um, sell. Is the new pill, I suppose. I'm just making things up now that to me, would, would attract my attention because first, not many people know about that space.

Not people may even have, have access to that. And, and so then I would have to ask like, what is the amount they're trying to raise? And, and, uh, and then they'll have to convince that their, their focus area thesis has a very promising next five to six or seven years in terms of going after making these bets.

I think that's, that, that would be at my attention, but I would be honest, I think it would be very difficult for a first time fund manager to come in and say, look, without a big. Sort of some sort of access or a very deep knowledge or understanding of a particular field will be very difficult. Those do get raised, but they do come from investment banking world.

And their claim typically is, look, we've, uh, I come from the banking world. I have relationships with all these funds previously where because my clients were and so I have that access again in that sense, they're bringing asymmetrical access. Typically how fund funds operate. They're not typically deep, um, into the operations or or understanding of the field, but they have previously been in the financial world and, and they have amazing relationships.

And so they formed these fund of funds and say, look, we will expose you to N number of funds through us instead of just taking positions in one or two. So that's, I would say it always boils down to those, these two things for me. Um, if it's not a traditional route, say, some emerging managers taking.

Thank you for clarifying that. And I'll repeat asymmetrical access and asymmetrical knowledge for emerging managers. Um, and. Really try and quantify that it seems like the advice is what I'm getting.

Yeah, that's also a very important thing. I should also mention is being an operator is a very important aspect as well, if, if particularly you're going to take bets on startups, it's one thing to be working for somebody, even though you may be an executive.

And it's one thing to start something from scratch, to take a vision from, you know, napkin, so to speak, and to bring that alive, no matter how big or small that is, is a different experience that you just can't get because the risk, what that shows you have the risk taking abilities. You can be an executive at a company, then you're, you're taking risks just with your career.

Say, okay, I'm just going to spend some time. You always have a fallout fallback option, but when you're executing your own vision, you have stakeholders. It's not easy to just unplug from that. You're, you're on the hook. There's a lot of risks. You're managing risk. You're taking risks. You're going to track.

So I think that also would attract my attention where someone's started something and failed. As well, it doesn't have to be a super successful thing, but is there something, you know, there's some fail fast, fail forward sort of thing. In fact, I came across a LinkedIn profile the other day and, and honestly, the top, I don't caught my attention said serial entrepreneur.

Forex failed or something of that line, right? That's should be a bad, I mean, well in the Valley, it kind of is, but the point is the experience matters a lot in, in doing that. And so I think when you're looking at a founder, especially pre seed, if you're going to be a first time manager. And you have a small fund, you're, you're most likely to be looking at very early stage, which is the most riskiest.

And so how are you going to identify, first of all, what challenges of products, uh, are this vision is going to have, is there a right trait of founders exhibiting? And if you're claiming that you're going to be able to help them without operational background or being an operator, it's very difficult.

That's, uh, when I send emails, my second line is founder of one field startup, um, after my initial intro.

That's, yeah, it's good to mention that. I think it's, it's, it's, it's important to identify that because there's, it's, it's all about experience. You know, like I said, we tend to think about resulting. We were our previous conversation just, just to kind of close the loop on that because we're, we're so tuned to think about our outcomes and decisions in the same kind of way.

Correlation, but they're not really. I mean, I think outcomes sometimes become a function also of right timing and luck. Let's just start with that. But decisions are what are you, what data you have at the time to think that you're making the right decision. Sometimes a brilliant decision will turn up on the pretty wrong side of the equation when it comes to outcomes and a really bad decision could go wrong.

the other way. And so I think that's what it shows that you have the risk appetite. You have the, the sort of the aptitude to go after and make those decisions and outcomes sometimes can, won't go in your favor. And there's a famous, um, you know, popular term resulting in poker.

Yeah. Uh, I think Adam Grant talks about this as well.

And his book originals were reward the process, not the outcome. Don't reward luck. Cause bad outcome could, could come from a good process. Let's go deeper in that and let's get into building a startup and operational experience. And let's specifically go into incentivizing your team and your employees and let's transition to sales.

Right? So oftentimes you will incentivize the outcome in sales. It seems fairly, um, industry, uh, standard. Do you think that's Uh, incorrect thing to incentivize. And how do you think about incentivizing, say your own team, your own principles or analysts? And really focusing on the process, um, there

it's a difficult question.

I'll be honest. And I've seen a lot of enterprise companies and especially when I was one of the leading architects of Salesforce, I got to see the sales part. And sometimes these were very heavy engineering companies. Yeah. The incentives were pretty much towards sales. And so I guess let's start with.

With something that ultimately a company or a startup isn't the business of making money. It's making money for itself. It's making money for its employees to continue paying them. It's making money for the stakeholders and the investors. And would

you say making money is the process or the outcome?

Well, that process, well, it's ultimately the outcome that's going to drive the, the company. I think without getting the right outcome. The company won't survive. So outcome here becomes a survival sort of thing. While in, in many other interpersonal things, we can always say the outcomes and decisions can be sort of, um, separate in, in a lot of the cases here.

I think if we, the company ultimately relies on the outcome, but how the, how the structure of the people involved can be, can be detached and, and you're absolutely right, I think not only companies from their employees perspective, but even the CEOs. Get fired or, or, or let go based on just results, not the quality of decisions they made actually to think about it.

And, and that's something, and the reason, and the reason for that is because we do resulting, so to speak on that and that, and, and it's hard to get away from that because if the company has hit their marks and quarter after quarter, well, then it's a, it's a, it's a tight circle of not being able to get more money or existing investors, um, putting pressure and.

It's, it's, it's, it's beyond just one person's thing into personal. I think Google and technology companies started to incentivize engineers more than say traditional companies were, where they were just kind of one department. And so that thing's definitely changed right now. And so that's what Silicon Valley is all about.

The, the engineering became the Goldman Sachs kind of skills. That, that, you know, nineties were incentivized. So mid 2000s, the high incentives you saw were for engineers, but, but those companies were their entire product was an engineering product, but if we were to look at everything else and also remember, okay, so very important point is most enterprise These companies have a product that makes them money, right?

So an air condition or a car or, you know, any tangible product. Well, a lot of the companies that were incentivizing engineers and not just sales were because the user were the products. They were free products, right? So, so it's a different, it's a completely different, uh, dynamics where traditional companies had to incentivize sales team over.

There, even if it's an air conditioner over their engineers, because engineers can create the best air condition in their, in their warehouse. But if no one's buying it, then that is the best condition that sits there. But in the technology companies, particularly the internet companies, like the Facebooks and the Googles of the world, there's free products.

And the only way you saw adoption was a great engineering. And if you've got a great product out there, that meant. ultimately the outcome of money. So I think it was easy, in my opinion, to incentivize engineers there. It's not so easy if you're selling an actual product, we need to go through sales and the marketing and all.

And I think you need to incentivize that them over, I would say, engineers because they're selling. I'm not agreeing with that. I'm just saying what the dynamics are and how the structures are in place.

I'd love to get your thoughts on digital therapeutics. Specifically, let's talk about the downfall of pair.

It seems like their model was reliant on pharma drug pricing for a SAS service. Do you think the market is ready for digital therapeutics? And if not, when will it be ready? Yeah,

that's actually a good question. Digital therapeutics did have a very exuberant cycle, particularly during pandemic. And now it's starting to kind of.

fizzle down a little bit, but that, but any sort of innovative idea or sort of approach has to go through this kind of cycle. It doesn't mean it's, it's not there. In fact, I feel there's a, there's a really good place for digital therapeutics. The problem, I think, in my opinion, was it's always the over the hype cycle, the old, very, whenever there's a new approach to something, there are a lot of folks who enter the space with half baked ideas and, and then there's money that goes into it.

And especially if this hype cycle happens to be in low interest rate and free money environment, everyone with the PowerPoint will get money eventually even reach. And they're not forced to think business model, really. They're not forced to think, is there really a market that's ready for this? And if so, these are hand in hand question.

If there's a market, how do you get to it? And I think in Paris case was entirely reliant on pharmaceuticals, kind of assumptions that were made that there's a potential customers in this, but I think we're kind of starting to enter the space where I think a lot of that. It's true. Princeton now, I think real.

Real products with wealth out of business plans, business models. And now we have experienced learning from what didn't and did not work or did and did not work. I think you'll start to see some real application of digital therapeutics come together. It's kind of similar to what happened with. It, that, that can be applied really to anywhere in nineties, literally anybody with a PowerPoint deck saying, I'm going to turn this into a website.

Got lots of money. In fact, went to IPOs and we all know what happened, but the real sort of products that came out were after the crash. Um, I, I. I think that that's true for tech bio. I think that's generally true for digital therapeutics, which, you know, I think that's also true overall for digital health.

I think digital health went through the similar kind of cycle. So I think this is the right time to be in it. That's my opinion. But I think mental health, for example, preventative medicine. In terms of diabetes, I think a lot of these things have, uh, something that doesn't need sort of a pharmacological interventions.

We could use digital health technologies and and therapeutics kind to speak to, I think, um, has as a role to play, but I don't know if we have fully figured out the business model because ultimately the answer is who's going to pay for it. I mean, the question is who's going to pay. If for it and where the incentives are going to come from,

I'll push back on this a little bit.

I am not a fan of momentum investing largely because it's very anxiety provoking and it seems like a very difficult way to live life for me as an investor, but a lot of money was made investing in romance and cerebral and for him. Will these companies survive? I don't know, but regardless, falling these hype cycles, these bubbles, as long as you can get out the secondary market, we'll get out at the right time.

You know, momentum investing can make a lot of money. How do you think about emerging managers who come to you and say, I am predicting this next hype. I am predicting AI will take off at this point and there will be a bubble. I'm predicting the bubble. It will be a two year cycle. I'm going to get out at 18 months.

Um, do you think it's impossible? You know, this is a common mantra. It's impossible to time the market in the public markets. I do think the private markets are a bit different and maybe that mantra doesn't apply. Um, what are your thoughts in momentum pipe investing?

Yeah, I was going to say that. I think momentum particularly applies to a lot more in public, private.

The reason why it's private, it's called private markets. Not many people know about it or are playing in that space. So start with that. It's a very small space. Money could be quite, there'll be a lot of money in there, but small space, nonetheless. And so in that sense, if you can take in a contrary opinion and go with something that is.

Going to be very new. And if that works out, you will be called visionary. Sure. But I think those kind of bets are very easy, very hard to make if you're, if you have only a small amount of money to deploy. And so, and if you have a small amount of money to deploy, and you're going to just go into one sort of, um, And, uh, I wouldn't call it a hype cycle.

I would call it, um, the next big thing. That's what private markets are about. I think in fact, fundamentally the power law works and, and what private, what VCs have always kind of the model is to go into the next big thing. It is not about finding something that is stabilized between a, after a bubble or that, then you, you might as well be in a private equity model where you're really taking some step, some establishment and turning that, but it's not going to give you those massive returns that VC sort of promises.

So you have to be in some, some form of fashion, be in something that is going to be the next big thing. And you, you, then you make. Some intelligent bets. That's how I see it. But if you look at, um, how, how emerging managers would think about what's the next big thing, like take a look at example of AI. I think if, if someone was to just say, I'm going to invest in companies that are at the intersection of AI and, and all, I think the question that.

And will be, uh, how do you protect against yourself with the potential bubble that we are probably in and, and there are ways to think about. I mean, that's where the strategy that's where, uh, what you present ultimately will allow someone like me was looking at a pitch for the fund to make sense of, for example, if someone is to say, hey, I think AI and.

Is really ripe for precision medicine, for example, and then you show why such, right? I mean, and so that allows me to start thinking, okay, this is well thought out thesis. At the end of the day, these are all that's there's not nothing certainty and and there's a financial management aspect, which you were alluding to.

How do you get out? How do you have secondary markets? This is where I think season fund managers know how to how to do that. It's very difficult for emerging fund managers to think about, Hey, how do I. Okay. Think about how do I provide liquidity? How do I sell off in enough time to make sure that there is, uh, I'm getting out at the right time.

I think this, this comes from being seasoned. And so this is why I I've constructed my buckets where the established fund managers will do that. Some fourth, fifth fund managers will probably do that now. And that's why I call the last one very risky because you could be a great investor, but. Managing a portfolio is a different skill set.

You eventually start to learn, but you don't want to take a lot of bets on first time managers because as you know, as they say, I shouldn't be paying for your learning experience kind of thing. Those are my thoughts on that.

Yeah. And you know, I have a million questions about, you know, VCs who hold on to companies post IPO, but I want to ask a different question first.

And famously, Sequoia does this to an extent, um, how much do you rely on structure versus intuition when picking your investments? Do you have a scoring sheet? And I don't, I do not score my investments because I think I can talk myself out of investing in anything. I will never invest if I do that. Um, I do have a structured approach and then I apply intuition after I never invest against my intuition.

I have in the past. And I think, uh, it's too emotionally taxing for me, um, to an extent, how do you think about structure and intuition when investing

the way I think about investing, it depends on the sector, I think. In the areas that I'm, I bring that asymmetrical knowledge. And there are some areas where I think I could be a, quite a thought leader myself and we're standout, which is where I do some keynotes in those areas. That's where I really heavily rely on my intuition.

And of course, but for the structure, you know, the team and, you know, the typical structure that everybody has, but areas that I can, I understand. I get excited about. And the areas that I think this is very interesting, then I'll bring in folks that can do the diligence and I'll let them kind of give their independent thoughts, like write up like a very structured memo and what their recommendations would be.

And I will incorporate that. And then ultimately it's a bet. And so that's where intuitions do come in. But I think you can't just completely rely on intuition. I think if you follow a process, you also make it data driven in some. And so look at lots of data, uh, about the market, do some, um, um, sort of, uh, gathering of that on your own.

Ask the right questions, um, and then have the right folks involved in that particular deal to make sure you're getting contrarian views and independent thoughts on what that is, because that's a very important 100 percent of the times there's something you're not thinking of that others picked up on and that's within the team.

Right? And so that's why I have some mentor partners and, and, um, and associates that that really bring that to the table in certain areas. And so, and then ultimately. Okay. But lead the decisions mark and so this is where with everything that you have the diligence you've done Um, then question it's before even the intuition part where you're executing the bet.

It also is, what are you looking from that investment? Do you see that investment as potentially that's going to be the fund returners? Then you have to start thinking about reserves and the ownership, or do you see that this is a, this is an investment that may not scale to beyond like, say, 300 million.

but could have an acquisition in play. It could, it could be acquired and we're getting so early that if it gets acquired within two or two years, it's great outcome. It may not be more than two X, but it is, could be a good DPI, so to speak. And so my thought process is always, how do you invest a dollar and bring that dollar back very quickly?

I mean, I would be very happy with it. Would you invest against

your intuition?

Well, yes, if it's informed by the opinions that I've gathered from the team, and Have you in the past? Yes, and they have worked out very well and in some cases not so very well. So I think both. In fact, um, um, it's it's one of those things where Or, you know, I think your, how is your intuitions informed is also very important and how are you going against your institution based on what is very important.

Intuition ultimately is driven internally by I, I think, but if that I think can be replaced with some, uh, with some informed structures that you've learned to trust. And I think, I think. The mark of a good investor is to make sure that, um, you incorporate that and not drinking your internal Kool Aid all the time.

And, and there will be some strategies where people say, Oh, I'm just going to stick to what that is, but then you just be a solo angel investor. But if you're going to construct a portfolio, you're going to construct a team around you, then that's your best leverage. Yeah.

For those listening, Kahneman's framework of high and low validity environments.

Um, and I, I'll link to something that I wrote on that. Um, let's talk about first mover advantage or disadvantage. Let's talk about Peter Thiel's Zero to One, where you're creating something novel. And how do you value first mover advantage, building something novel versus the fact that oftentimes you're paying for your competitor's education or paving the way for them?

And I'll kind of say, you know, Google was the 18th search engine, which is a, that's repeated famously over and over and as, as pushing against first mover advantage.

Well, I think first mover advantage is only an advantage. If you are going to be kind of cutthroat in the sense that your mentality is to become a monopoly, as I think, uh, Peter Thiel famously points out in zero to one.

If, if you're going to be a first mover in a space and you're thinking incremental stuff, I think you will be replaced. And then you've essentially created market awareness and someone's going to come in. scoop that up. But if you're going to come in and think I, I need to become the dominant and this is going to be a kind of an eventually a monopoly, that's a different mindset.

And then there are different metrics to go after if it's a, if it's a, one of those products that rely. Heavily on users and how are you going to best capture the network effects? How, how is it that your IP structure comes into the play, which, which will allow you to kind of be in a dominant position where the trade secrets and the team that you start building ultimately, you just need to be, the founder has to have a very, I would say a really unique skillset.

To be able to execute some of those, I think those, it's not easy to see that at the time, but you have to pick up on that. Do they have that X factor, especially if it's going to be zero to one company, but you know, if it's going to be one of those where you have an incremental change and you get acquired, that's a different thing.

Otherwise, yeah, first mover is a very tricky situation. You could ultimately. raise good money and be a company. But then what ultimately will happen is someone else will capitalize on where you lacked and because they're going to be lean and things are much better in market awareness, they might outpace you.

If you look at Uber, Lyft was never able to come in and take their first mover advantage. Right. That's one example. But if you look at MySpace and Facebook, it completely flipped. And the reason, you know, if you look at both the reasons, I think MySpace was never thinking that we need to be a monopoly. I think they were thinking, let's just keep capturing, uh, teenagers who just like to have fun online.

And, and that's why they allowed anonymous postings, anonymous profiles. Nobody was who, I mean, you could be anybody and whoever. Well, Facebook said there's a bigger market to capture, but if we give them some security on what the profile, that's why they made it very private. You had to have a college email address.

You had to have email. You had, they made this sort of, uh, a culture to have a proper face profile in my space. These days, people used to put their band's profile too. Uh, and nobody was. Fully comfortable. I mean, there were folks, but as you kind of became older, you were not comfortable putting all your information on my switch.

It was completely open. Anybody could see anything. So this idea of private networks, um, was captured well, and that's, they lost their first movers. And, you know, we know what happened Facebook, but look at Uber though Uber on the other. I was thinking, Hey, this is us. We're going to think like this. We're going to invest so much money and, uh, getting lawyers to go and fight the city municipals and the medallion.

Okay. Things in the cabs, they were not bogged down and so Lyft has come, you know, I think there's, this was a space where there could be multiple winners for sure, but Lyft never came and kind of became the Facebook of that space. So, so kind of some examples, uh, it all depends. I think there are some markets where there could be multiple winners.

There's some markets, there's only going to be one winner. I think it all depends also on what the aspects are.

This reminds me of one of my favorite books that I regret not reading before my startup, the hard thing about hard things by Ben Horowitz. And he talks about wartime CEO, peacetime CEO, ruthless, formidable, radically honest founders.

Do you think all pre seed startups need wartime CEOs?

I don't know, actually. I don't, you know, I don't know. I think pre seed is such an interesting time period for an idea, really. Really, it's an idea that it really tests. As a founder, your resiliency and your commitment, I think if you're fully committed and you're resilient, you will, if you can survive that, uh, then, then your idea will start to take form or shape or form precede.

In the other sense can also be that you might latched on a pretty brilliant idea and you got some, some exciting funding, but you have, but you didn't get focused on execution and, or may not even have that skill. So I think Preseed is actually a very interesting one because there are some founders who can execute things very well, but they just can't fundraise.

They're not good storytellers. And so there are some founders who are really good storytellers. But when it comes to execution, you just can't, they just don't have that. So I think precinct investment is so risky and it's, and I have made. Pre seed investments, but let me just tell you, I've not made any pre seed investments without actually establishing a relationship with the founder, meeting them in person, having a meal, just got to understand beyond what that is.

There are these sort of, uh, soft skills that needs to come into play, uh, because that's what the stage of pre seed is. It's, it's, it's about, it's about soft skills. It's about. Being resilient. It's about, uh, getting, getting the funding in. It's about executing it on your ideas eventually. And this, and so that's why I think, by the way, I was reading some stats that a lot of the Y Combinator founders who kind of don't amount to much and they fold.

And I think one of the things that that's come to bear is that, um, it's, it's the execution or the focus. A lot of founders are also not ready to start. Managing money as well. There's just not. And so you get distracted by lifestyle choices, you know, especially if you're out of college and all of a sudden you have three, four, five, 6 million in funding.

And so are you even mentally ready for that kind of stuff? And it turns out not many are. And so the brilliant idea goes from could be the game changer to amount to nothing. Yeah,

I think, uh, why combinatorial healthcare is interesting, and I'm not sure. Healthcare is so specific where equity of access is paramount.

So a lot of the principles you will learn in business school and in the startup don't apply to healthcare. And a good principle is the 80 20 principle, or focus on 20 percent of your customers who deliver 80 percent of our value. You can't focus on 20 percent of your patients, you have to focus on 100 percent of your patients.

And move fast and break things. Like, there's a lot of mantras that, you know, are dangerous in healthcare. Um. Healthcare,

I guess, also depends what you're saying. Is it B2B focused? Is it, is it consumer tech? I mean, consumer then falls into the traditional stuff we're taking. You look at, uh, ORA rings or Fitbits of the world there.

I would define them as health products, but it doesn't have to be administered. But the whole idea of this health revolution is that it comes to our bedsides. And so in this, in that case, these are traditional tech companies, but I absolutely agree, I think health, if you, if. If there's a healthcare company that's going to help with EMR process or, or helping with detecting, um, postpartum bleeding with an, I mean, these are, yeah, those models absolutely totally upside down there.

You got to think of how the insurance comes into the play. What are the CPT codes or regulatory? Those are different skill sets and different than, than I think the typical, I guess, tech startup, which is, which can just be focused on finding product market fit and the users will network effect will carry you forward.

Yeah, I think the other common adage I'd love to get your opinion on is innovation happens from outsiders. Uh, I'm currently reading innovators dilemma and it kind of touches on this as well. Um, do you think that applies to health care? And a faith made example here is the Wright brothers versus Samuel Langley when they were developing planes.

Samuel Langley had tons of funding. I think he's from the Navy and the Wright brothers were these two brothers with a strong why, but very little funding. Um, do you think this rings true in biotech? Do you need industry expertise or can an outsider be successful in a biotech startup? As a founder.

Yeah, it's how it can be successful.

We've seen this with pharmacyclics. We've seen this with gen and tech. We've, I mean, we have, we have examples already. In fact, um, you continuously see examples of physicists getting into the space of biotech and being very successful. You've now started to see. AI talent come in and be successful, be it academic papers or pushing the field a little bit forward.

Look at DeepMind. I

mean, do you look for clinicians on the advisory board or founding

team or no? Oh, absolutely. I mean, I think if it's one thing to develop a technology and the other is the deployment of it and how, if it's particularly focused towards delivery of care, right? So DeepMind is solving a different problem with AlphaFold, right?

How do You find those structures and, and you can make it. Case that clinicians don't fit into that picture, but if it is, I'll give you the best example. If there's a company that is going to, uh, say we're using AI and to read digital pathology scans and we can identify patterns just from basic. H and E stains that, uh, if chemotherapy or sorry, uh, immunotherapy is going to work or not, which will be phenomenal digital biomarker to, to write a front.

Okay, great. This is a fundamental technology that can be developed by an outsider with, with getting some understanding of. Of how the pathology images are and can just take a view that I'm just going to disrupt this now comes the hard questions about, well, how is it going to deployed? What is the regulatory for?

How are you? How are you going to do clinical trials on this? What are some of the nuances that you missed? Uh, and so this is where the clinician. In the loop has to be very important. The ones who bring the domain knowledge and, um, and guide that through. Because without that, ultimately, the tool that's been developed here.

I'm just giving 1 example is for the clinicians to, to make better, uh, predictions of the outcomes. Right? And so if they're not in the loop. Then, then I think it, it, it, uh, it becomes a problematic thing. So absolutely in that space or anything that involves, uh, diagnostics and prognostics and delivery of care, got to have somebody that is, if, if not a co founder, then prominently present in some operations or advisory, uh, otherwise you're going to iterate on something that you may not find actual use cases for.

Last question, Sair. What is your contrarian truth? And we talk a lot about can AI replace physicians, but as a thought experiment, I'd love for you to give your thoughts on what happens when AI replaces physicians.

I think that the, I've thought about this question, but you know, I came across something that Richard Freiman said in the 80s.

The question about Replacing a human with some technology, be it AI, that question fundamentally is flawed because the question that planes replace the birds, planes are flying, but, but the concept of flying between a bird and a plane fundamentally are, but it's not the same plane doesn't have to flap the wings and plane can carry, uh, massive amounts of load or bird naturally what it does.

Right. And so pros and cons. And so I think that's where. Okay. I kind of start to think about what AI is doing. I think a lot of folks are getting bogged down on, is it going to become more intelligent, super intelligence? Because the, the word intelligence, we don't have a good definition for when a technology and I'll, I'll just keep it a high level technology, not even AI that replaces a human function.

It typically will do that if, if it really outpaces everything that we can do and it has to do it better. And fundamentally better, not only in automation, which is where I think everybody's focused on, but then there's a societal aspect of it. And I want to give you one example. When I was in high school or entering high school or middle school ish in Houston, next to my apartment, um, we had, I don't know if.

You are aware of Kroger's, but Kroger's is like this massive grocery store chain that we have in, I think, Texas and, and Ohio. So Kroger's had just employed these things called Uscans, which you can self scan these things. This is, I think, um, late nineties, early 2000s. And there was an article that came in Houston Chronicle that said, we have these self scanning machines, the future of cashierless.

services are here. And that article went on to make a prediction that his next five years or six years, we'll see less cashiers and self scanning what has happened after 20 years or even 20. We still have cash. In fact, more cashiers, you don't see self scanning as much. In fact, Amazon go did it take on these, you know, they had these airports.

I went back and they're all gone. All I'm saying is just because there's an automation, there's a technology doesn't mean it replaces the human function. And so when I think of physicians and clinicians, I put them in two categories and this might sound sort of crude. One is that are dealing with the humanity or humans that interactions.

And that it's, it's the same question as, do you want to talk to a human and deal with human? Because there's medicine. It's about healing and healing has a human touch to it. And then there's this back end operations of medicine. So radiology, pathology, blood. All of the other things that a patient doesn't have to typically interact with, I think those will be first to get automated and I think maybe potentially completely replaced at some point, but when you think, when you think of a physician that actually has to, that human touch, that's going to be very difficult.

It's the same example. If you want to look at what happened with cashier less systems. If you look at, if you ask most of these stores and I have, because I did Salesforce for them, simply the thing was just having a human present there. Makes better sales. It makes a better interaction. So I, I see a very hard time seeing physicians getting replaced by AI because we are at the end of the humans.

And at least we're not there yet. That was completely want to just sit at home and not no other interaction. But I do think that there are some functions of medicine and health. Delivery that will be fully automated for, for, because they will bring such a huge value add that just outpace what humans can do and the error rates and things like that.

And then, and just the speed of things. But I think the human, the physician that brings the human touch, the, the probial healing, I think that's, that's, that's going to remain. Do you have a hard stop, sir? I can go for maybe another, uh, 10, 15 minutes.

Okay. Um,

Let's talk about, is VC a game of picking the best founders or making them? And there's two kinds of different varieties of venture capitalists here. The founders front variety where it's really about picking the best founders and getting out of their way or the A16Z variety where, you know, you help the founders commercialize or exit as well and hire a team and help them as an operator as well.

Where do you fall in this paradigm? I think it

starts with

Uh, uh, uh, a firm's own capabilities and what they can bring to the table. It's very important for the investor to understand that and then kind of go after that model. I think A16Cs of the world are obviously much bigger with wider networks and they can certainly bring to the table, the value proposition.

And so being, being really tied to the hip with the founder obviously helps. But funds who don't have similar level of capacity, uh, You, you can bring some level of support system. Everybody has a way to bring that. And so I think it starts with internally evaluating what is it that you can do for the founder.

Of course, the, the most important thing you're doing is bringing the capital, but be. Being a strategic investor and being able to help them could be in many capacities. One, it could be helping them in thinking about business strategies, think helping them scale, helping them with next round in capital investments again, because every startup needs more money and helping with potentially getting customers.

And so everyone has that. Capabilities, no matter how big or small. And an example I'll give you is that say there is a company that is, um, bringing out some kind of a robotic automation in dental industry and in a large fund might bring a commercialization officers and things like that. But dental industry is very much mom and pop shop owned as well.

Yeah, there are bigger conglomerates, but there's this individually owned. And so a smaller fund might have Big relationships with these, these, um, sort of family style owned. Dentistries that own hundreds of them or some private equity firm that actually owns, you know, a few hundred dentists place, those could be potential customers.

So I think everybody has their skill set and, and, and capabilities that bring to the table. So I think it starts with understanding what that is and being able to, uh, communicate with that founder. And, uh, I think where you're referring to is that if you're a lead investor and if you're a lead, then you're on the board, that's a different ballgame, I think then you're.

They're on the board, you have to be sure you have to be very involved and not the operations, but just really be involved in what the company's progress are more than you would with any other portfolio of yours. You may rely on just a quarter of the update and you know, their updates, but when you're on board, your responsibilities are different.

Let's go deeper in the AI conundrum. Um, and I love talking about intelligence versus consciousness because we can see AI as being intelligent. Plants are intelligent, but they're not conscious. Um, Yuval Harari talks about this in his book Homo Sapiens quite a bit. I'd love to get your thoughts on consciousness.

What is consciousness? And do you see a future where AI possesses consciousness? And if so, will it pass the Turing test?

Well,

I don't. No, if I have some very well thought out things on consciousness, because that itself is quite a paradoxical if I'm consciousness, I'm thinking about consciousness kind of thing. And that probably what makes I think that my definition of consciousness is that. Beyond my sensory inputs and and just beyond just processing of things and myself my have I become aware of myself in light of everyone around me and I think that is a very different proposition on top of an intelligence.

So I think to me, intelligence and consciousness are different. Um, I think intelligence ultimately being. Is about how do you process data information coming in and to make certain decisions certainly consciousness can be mixed in, but consciousness about being aware of you and where you are in surrounding of the others.

And so example is plants, you know, they, they have an amazing process to stay alive and they do, but they're not, they're not self aware of. You know, in the jungle and things like that, they may have roots and that's what, but again, it's all about sense collecting sensory data consciousness, uh, arises from when you become self aware and, and you can start to observe things beyond just your data, data gathering capabilities.

So I think to me, that's at least what I, I have not done more metaphysical thinking on this intelligence though. The current problem that I have that the LLMs and what's happening, I mean, I think at the end of the day, are they reasoning that's where the everyone's divided on the matter. My thought is that it's not quite reasoning capabilities.

I think it has capabilities of of finding right substructures in this high dimensional space where. where you're, you're starting to, uh, to see where things fall in. So you can do the next predictions and you're aware of where, where things are. So you can start to kind of gather long sequences as output.

Is it reasoning? I think it's to be seen. It may appear like it's reasoning, but that's what predictions can do. It can actually, the stochastic processes can give you this perception that it's. The reasoning, so reasoning to me will be a little bit different than what's currently happening with, it's a phenomenal progress, so don't get me wrong, but ultimately at the end of the day, what large language models, the base models that we get are just capable of predicting the right, right next word with a certain view on the context.

That's kind of what we have. And then on top, we have this kind of the RLHF layer, which. Does the prompting and all of that, but that's just a nifty engineering, the reinforcements and all of that. I think it's still yet to be seen, uh, if we get full reasoning, because reasoning involves a proper reinforcement loops.

And so I'm really excited about reinforcement learning. I see that as the thing. And the last thing I'll say is that the other fundamental thing is that a lot of our models are given an objective function now, just to be a little more technical about it. Intelligence doesn't rely on a very specific objective function.

Our intelligence, if we're now starting, we were not given A particular objective function, we have multiple objective functions, and they all work in tandem in a pretty incredible way. And so an objective function where you're born is to make sure you're being fed milk. But if, if you're not, you will, you will have this thing to cry.

And, and so eventually that objective function over time learns to start becoming self sustaining, you know, I'm going to eat myself, but that's not the only goal. There are so many of these objective functions. So, I guess, ultimately, what I'm saying is, if we are to go toward. towards AGI. My, my, my opinion is that not only these models can really hone in on objective function, but these models will be able to tweak their own objective functions.

That's when we'll start to see real AGI, if that makes sense. This is where we start the reasoning. I can reason. Therefore, I can change my mathematical objective function, but that's not what we have right now. What we have is gradient descent, and it needs to have an objective function. Otherwise you when do you stop and the stop is local minima and such.

So I think it's a it's a it's a pretty catch 22 situation. But I think that's where I would be really convinced that we're moving towards. Yeah,

it would be interesting to see if acts in a social capacity like we do. If it interacts with other AIs, like we interact with other humans,

um, I, I think the social construct is very different.

I think I'm just really focused on just purely operating at a data point of view, social construct of an AI will require. Many different things. It's one thing to interact with Alexa, you know, in a shapeless, formless manner and a response comes back that will eventually become part of us, our, our, our life, kind of like how it.

Having a tv became part of our lives. If you had shown this concept of a portal where you're seeing things about 300 years ago, people would be like, this is not, this is not normal, like to sit there and like stare at moving things on a box. But now, if 100 years from now, maybe speaking to a box and having a proper relationship with that might be a normal thing.

But I think where you may eventually be going towards is sort of these machines that are now. In form of humans, like the humanoids and all. Perhaps there, there is. They have some room to be this, um, to be collaborative machines within ecosystem. And I think this is where really the social construct will come in.

Can they interact with being in the, being in the homes like a human would? And so this is where I think the human centric designs and things like that would come in. Yeah, I think that those are my thoughts on that. Uh, so social, social part is totally different in my opinion. We tend to mix that, but I think at the moment AI is just focused on can I look at large amounts of data and find rules that govern where these things were, how they were generated.

So in that sense, this model could inhibit what human reasoning could be. That's all it's doing right now. Right. I mean, Stack Overflow and Wikipedia, these are human thoughts. So the source, it's not, it's not creating a brand new source. It's just looking at what the human thought process has been. And through this mathematical operations of non and adding some nonlinearities, ultimately with, with massive amounts of layers in the middle, it's starting to say, Hey, I think these are the set of rules that govern these human thoughts.

And so hence I can predict the next word may look like reasoning, but I don't know if it is truly.

Well, thank you, sir. Um, and this has been amazing. Thanks so much for coming on the podcast.

No, thanks. This was a very good conversation, actually. I really enjoyed it.