Loading summary
Transcript158 lines
- [00:02]
Joel
All right, so I'm excited to continue the new year with just a lot of great speakers. Today I've got Suman Talukdar. He's a founder and GP of AI Sprout. So, you know, excited to go deeper on your background, Suman? You know, I've gotten to know you as you've kind of been building AI Sprouts, and you have a really amazing background as an operator, you know, with just several. Several companies back to back with operating exits. So I'll let you kind of talk through some of that as well. But I think it'd be really helpful to kind of learn a little more about your education, your background, maybe what you thought you'd be studying and learning and how you got into venture capital, because, as you know, it's just an opaque industry. So, number one, welcome to the show, Suman, and thanks for your time, and, you know, just excited to unpack everything that you're building, especially when it comes to AI. One of the things that you told me is you worked on AI when it wasn't as sexy, right? It was just kind of stuff that people were building, but just everyone didn't really understand it from a mainstream perspective. Now it's much more mainstream. It's very frothy. So I'd love to hear your take on just kind of where AI started and where it's. Where it's evolved. They say everything started from Star wars, right? The tech that's out right now, like holograms and AI and R2D2. You know, all that stuff. You know, just watch Star Wars. You. You. You know, you'd be where you are now. But why don't we start from the beginning? Sumon, like, tell me who Suman is. You know, you and I, we went bowling together, we grabbed drinks together. But I want to know the Suman that's. That's maybe three levels deeper. Just tell me about your background, what you studied, and just tell me, you know, kind of a little bit more detail in terms of, like, just your career journey and how you got to AI Sprouts, and I'll. I'll be interjecting in between.
- [01:51]
Suman Talukdar
Awesome. Sounds great. Thanks for having me. First of all, Joel, I've been tracking you for some time, and I definitely think this is a career milestone to be able to make it on your podcast. I appreciate you having me and being able to share a little bit about what I've learned on my journey. But, yeah, AI. Surprise. Just briefly, I focus on seed stage AI. So I specifically look for products and Founders when they have some traction. So just wanted to set the context on that because I think to your point, AI, if you take this very grandiose, futuristic vision, can sometimes feel like it's a little too futuristic or intangible. And I've always had the perspective of, you know, in my career investing in technologies and companies when there's some real like proof, right. I mean, saying it simply like there's a dog that's eating the dog food type thing.
- [02:55]
Interjecting Commentator
Right.
- [02:56]
Suman Talukdar
So I think that's the lens that I have on just to simplify it in terms of like the AI investments that I make. But yeah, you know, I think what got me started on this, you know, one of my first roles after graduating for undergraduate with engineering degree quite some time ago was actually with a gentleman who would later be one of the core founding team members of siri. So this gentleman, Adam Chayer, he's a great friend, he's a great mentor. I'm very appreciative of having his relationship through the years, ever since my post undergrad days. And at the time AI was all research. So he was coming out of SRI Stanford Research Institute, formerly called Stanford Research Institute. And they're doing all these things like agent architectures, multi agents, knowledge systems, all sorts of stuff. Machine learning actually was probably the least exciting thing because it was just not very computationally efficient. So it was like doing a bunch of math. But the rest of the stuff, like using multi agent systems, we were doing it for B2B commerce. That was really exciting. So that was my practical foray into building AI systems obviously after all the standard AI, cs, ee, undergrad stuff. But yeah, that's what got me into it long time ago. And at the time it was just cool, it was like AI is, you can make really cool stuff and a lot has changed since. So that's so just to start with that. And yeah, I mean that was my initial start to kind of my interest in AI.
- [04:49]
Joel
Okay. And then tell me a little more about what happened after that. So you built the interest in AI and you kind of started building an operating experience. So did you grow up in Silicon Valley?
- [05:03]
Suman Talukdar
Yeah, no, actually, you know, I'm a transplant here and I think know I very much respect everyone that's a transplant, whether it's from, you know, the state of Texas, like I was in, or, or from somewhere else. But yeah, I, I, I grew up my formative years post call it junior high high school was in Houston. Okay. And then I went to Rice University. So there's a great network of engineers and technical people out here from Rice. So that's what brought me out to Silicon Valley. Before that, I was kind of a child of the globe, if you want to call it that. I actually grew up in Latin America, etc. But yeah, I moved out to Silicon Valley and I got into the foray of what's really interesting. And AI wasn't as much of a focus in terms of let me keep building this massive expertise. It was more about technology and entrepreneurship and really absorbing everything that Silicon Valley had to offer. And I just kind of, probably for 10, 15 years I just focus on that, like building a great set of experiences and network to, to really find the next big ideas. Right. So that was kind of like the theme. And if you look at my career track record, it's always been about that. Kind of like trying to figure out what the next big shift or idea was and try to get it, get in front of it before others have.
- [06:34]
Interjecting Commentator
Right.
- [06:34]
Suman Talukdar
So you know, that included a bunch of like operator roles, like standard stuff like engineering product. Then I got went to business school, went to Harvard. There's a bunch of like Harvard Rice VC guys. So I felt like that was kind of a path to do it.
- [06:51]
Interjecting Commentator 2
Yeah.
- [06:52]
Suman Talukdar
And then post business school, a bit of private equity stuff, then moved back to the Bay Area and just buckled down as an operator. That was, that was.
- [07:02]
Joel
And tell me a little bit about your time in Latin America. So why were you there? Were. Did your parents have jobs there? Is that why you're in Latin America?
- [07:11]
Suman Talukdar
Yeah, no, so. So yeah, my dad was actually. So he moved to the US as a Fulbright Scholar. This is back in the 70s, you know, they brought, you know, talented students in to get their postgraduate degrees. And my dad actually got his PhD in geology it from Rice University. And then after that he finished that, we actually moved to Venezuela, where I wasn't born at the time. And, and because, you know, Venezuela was like flourishing, believe it or not. It was like the Dubai back then. There was like all this oil development. So my dad was actually a professor there. And so that's where I grew up. And Caracas was a very cosmopolitan, international place to be. And so I spend my formative years there. And so I'm, I am, you know, someone from lots of different backgrounds, if you know.
- [08:08]
Interjecting Commentator 2
Yeah.
- [08:08]
Joel
What are you. Some of your favorite, what are some of your favorite memories being in Venezuela and maybe looking back, what are some of the things that you learned about their culture?
- [08:18]
Suman Talukdar
You know, it's, it's an amazing any. I would highly recommend it I think, you know, if not just one country, but like if anybody has the opportunity to live abroad. I think it's a very eye opening experience and it learns, it teaches you how to be very open minded about ideas, cultures, specifically Venezuela. It's a, you know, at the time I was, there was a beautiful country.
- [08:43]
Interjecting Commentator 2
Yeah.
- [08:44]
Suman Talukdar
And, and you know, I, there's music, there's art, there's, you know, there was, there was development.
- [08:53]
Interjecting Commentator
Right.
- [08:53]
Suman Talukdar
Because all of this like industry that was happening, super friendly people, beautiful, like, you know, you got Amazon rainforest and you got beautiful beaches and very, very kind of like cosmopolitan population because there were a lot of, you know, expats that were living in the country. There were, you know, corporates that were coming to do business in Venezuela. So that was, yeah, that was my experience there. And, and it was great. I mean I think I learned a lot about having a, the value of having a very diverse kind of multifaceted network, a community that you can, you know, from lots of different backgrounds and the value of that. That was probably my, my biggest learning. And, and I was fortunate to go to great schools and all that kind of stuff and learn the language.
- [09:43]
Interjecting Commentator
Right.
- [09:43]
Suman Talukdar
There was no. Yeah, I spoke English, went to an American school. So when I moved to Houston beginning of junior high, I was like, you know, very well prepared, all the, you know, all the standard stuff and, and could readily kind of compete and, and operate in, within the American context. Right. But I think, you know, coming in as, as somebody that wasn't born in the U.S. i'm a citizen. But you know, I attained that through my, my parents. You know, my dad moved from Venezuela back to the US for, for a job with industry. Obviously it was like, you know, technical geologists, like so in the qualified, something that the economy needed. So then, you know, I was able to attain my right to live here and my citizenship through that. But, but yeah, I think one of the, one of the things that I, I sort of always had kind of the need to like really kind of prove myself a little bit more. Maybe had a little bit of a, of itch to scratch that, you know, I, I can do things the same way as, as, as others that maybe have had longer time, you know, operating in this economy, etc. So that's kind of drove, drove me to like go after you know, like institutions that would open doors.
- [11:09]
Interjecting Commentator
Right.
- [11:09]
Suman Talukdar
Like rice was a great one. And then later on when I went to Harvard it was like, it was all about like opening doors so that I could effectively compete. And, and I found that like, entrepreneurship is kind of like such a great, you know, competitive landscape.
- [11:26]
Interjecting Commentator
Right.
- [11:26]
Suman Talukdar
It's, it's. Once you're an entrepreneur of any form, whether you're a fund manager or whether you're, you're an operator, it's like really, like, you know, the, the initial starting condition is the same for everyone, and then you're really trying to build value. And that's a lot what kind of attracted me to kind of the culture of Silicon Valley. So it's less about, you know, it's. It's less about just the technology.
- [11:49]
Interjecting Commentator
Right. It's.
- [11:50]
Suman Talukdar
I've always been about kind of what technology and, and kind of these, like, new intellectual pursuits can, can do in terms of opening people.
- [12:04]
Joel
So you spend some time in private equity. You know, we have a lot of people in our community that are looking to break into private equity. So what, what are some of the biggest learnings that you took away from your, you know, stint working at a private equity firm and, and maybe some advice that you'd have for people to kind of maybe continue maintaining their good careers in private equity or just people trying to get into the space?
- [12:31]
Suman Talukdar
Yeah, you know, I think when, when I was exposed to private equity with that term, it was when I went into Boston and met all these classmates of mine that had been in there before.
- [12:42]
Interjecting Commentator
Right.
- [12:42]
Suman Talukdar
I was tech and product engineering, Silicon Valley. So I didn't really know what it was. I think now it's changed, and I think it's become more part of our everyday kind of vernacular.
- [12:56]
Interjecting Commentator
Right.
- [12:56]
Joel
And most people that go to business school, I mean, a lot of people, everybody has different goals. And I think that's one benefit of business school. It's pretty broad. But, you know, when it comes to private equity, they don't necessarily always have a private equity track. So a lot of times, like, you get the MBA and then you got to kind of network your way through landing a role, you know, where, you know, like, there's maybe 200 people applying to the same role. Right. So, like, what are some of the skills that you think you developed that really helped you, you know, maybe now in your formative years building a firm, you know, would you say it's kind of a combination of everything that you have to go out and, and raise capital when you're working in private equity? Or was it more like sourcing and screening at. At that stage? I'm assuming that the level that you came in at after an MBA was probably like an associate or principal role. So do you feel like there's some special Skills that you think you honed kind of in that, in that career.
- [13:51]
Suman Talukdar
Yeah, I think, you know, at that level, like I was a principal, you learned the like, rigor, the analytical rigor that goes into making decisions with a lot of capital.
- [14:04]
Interjecting Commentator 2
Yeah.
- [14:05]
Suman Talukdar
When you're making like big capital decisions, like 200 million bane cap or whatever is going to make this investment and you're going through all this. So I think, I think that was the lens that I attained.
- [14:17]
Interjecting Commentator 2
Yeah.
- [14:18]
Suman Talukdar
And, and I think it's, it's great. It's a, it's, it's almost like a very complementary experience to whatever you want to do. But it's, it's, it's a process and sort of space of its own.
- [14:31]
Interjecting Commentator
Right.
- [14:31]
Suman Talukdar
Based on like what size check you're writing. Right. So like, I think early stage venture, I think definitely the investment process, et cetera, and kind of what I should deep dive into, kind of benefited from that experience. But usually I'm operating in spaces where there isn't a lot of data. So it's probably a combination of that and kind of like the operating background of kind of understanding where a market is and then understanding how technology evolves over time. Those are probably like the complimentary skills that are sort of necessary in addition to getting that type of experience. But you know, with my journey into venture was again, like operator. I saw how these things work. I had a bunch of companies that did well to various levels, right. Like public, big acquisition, you know, aqua hires, bunch of stuff. I saw how it played out and then transitioned into angels. I was kind of putting my own skin in the game. And then I was like, hey, I feel like I have a recipe, a model, and then I'm gonna now sort of rinse and repeat and build a fund.
- [15:48]
Interjecting Commentator
Right.
- [15:48]
Suman Talukdar
So I think that was my trajectory that allowed me to get the necessary mental model.
- [15:56]
Interjecting Commentator
Right.
- [15:56]
Suman Talukdar
Like, you gotta, like, you gotta have enough training data.
- [16:00]
Joel
No, but I think it's true. I think, you know, I think there's different people that become successful. You know, they all go through different paths. But I just definitely think having some industry experience forces you to have that rigor. As you mentioned, you know, you have a structure, you have an investment committee process, you have a Monday partners meeting, you have your checklist where like, you know, there's a lot of founders that had an exit and they just kind of figure it out and they kind of use the founder model to maybe build a firm, maybe do, you know, think about product market fit. So I think just kind of having that institutional private equity Background. I think that definitely helps because you kind of take a lot of the frameworks that you learned at a large scale firm and you're kind of building a small firm with that scaffolding in place already. Which I think is really helpful, especially when you're trying to approach more sophisticated LP is they kind of see that you have all that structure in place. And I think it also helps that you had some of that information on your resume or on your bio that you kind of did that already professionally at like a larger scale.
- [17:06]
Suman Talukdar
Yeah, I'd hope so. And then I think the other stuff is, you know, accumulated experiences, my operating roles with portfolio companies and great venture funds.
- [17:17]
Interjecting Commentator
Right.
- [17:17]
Suman Talukdar
So like Clearwell which sold to Symantec, that was like a rock star team. It was like the best Sequoia company, like highest performing ebitda. This is coming out of financial crisis. That was like an amazing learning experience. Just like, you know, getting any of like the time of, of like the management team there and they moved on to build like crazy stuff. One went to become a partner at Sequoia and so that was amazing. The next one was like Bessemer, Bessemer Sequoia back company. And so I think learning practitioners firsthand, like going to board meetings, like whether it was, you know, in those roles, I was not the investor. Right. I was like, I was like complimenting the founder but just like hearing and like seeing how they kind of operationalize a lot of the things that or we responded to a lot of the things that were related to the investment pieces was probably just as much learning. But Yeah, I think 100% like the framework, kind of the top down if you want to call it, which is very much like the business school approach combined with kind of like a bottoms up, like realistic assessment of how you build things is a great sort of training ground to understand how to get yourself involved in right opportunities at the early stages. And that's been a bit of the lens that I've had, you know, in terms of like how can I find the best opportunities that fit my thesis for AI.
- [18:56]
Interjecting Commentator
Right.
- [18:56]
Suman Talukdar
I think it's been a bit of the pattern that I follow.
- [19:01]
Joel
Yeah, no, I totally agree and I think forming that thesis is one of the most important things to think about. And I think a lot of people as they're investing in companies are already thinking about this company's exit plan. You've got back to back experiences, operating companies that have had exits. So what are some learnings that you've learned as an operator when the companies are now gearing up for the exit. I'm assuming a lot of those companies are probably on the verge or at some point soon going to possibly be profitable or not. And there's different instances. But what are some learnings that you've had being an operator kind of, you know, seeing the culture maybe shift in prep for like an IPO. And then how should VCs think about not only IPOs, but just delivering liquidity? You know, we've got vehicles like secondary options, there's M and A transactions that are happening, there's a lot of micro pe. And then I'm also seeing a lot of startups now acquiring businesses strategically. So we'll love to hear the founder perspective and then also just kind of how investors are thinking about providing the service or optionality to LPs.
- [20:17]
Suman Talukdar
Yeah, I think, you know, I always had the lens of like looking for companies when I joined them.
- [20:23]
Interjecting Commentator
Right.
- [20:23]
Suman Talukdar
This was during my five operating experiences that I felt would have really solid exits.
- [20:32]
Interjecting Commentator 2
Yeah.
- [20:32]
Suman Talukdar
And then my learning from all those experiences was that kind of what that looks like all depends, you know, on a lot of different variables. I think, you know, investors, depending on who you are, you have like a timeline, you have all this like, you have a number, especially if you're coming from like the, like the private equity track.
- [20:54]
Interjecting Commentator
Right.
- [20:54]
Suman Talukdar
Like profitability. But coming at it from like the venture technology track, what I learned is it's beautiful. Like if it even works, right? If the thing is working like that itself is rare and to that and when it starts working and it works really, really well, then it's like amazing.
- [21:21]
Interjecting Commentator
Right.
- [21:21]
Suman Talukdar
And sometimes these things, a couple of companies I was with, cloudflare, Anaplan, even Clearwell towards the end, you know, the, the product, the technology kind of like takes on a life of its own and it's got this momentum and like the, the, the teams, the, the management team's purpose is to kind of like keep that flow going. But if you're able to do that, it just kind of like it just takes off on its own.
- [21:51]
Interjecting Commentator 2
Yeah.
- [21:52]
Suman Talukdar
And so to me in those situations, obviously the exits are tied to kind of a lot of like where the fund is and there's all like more seasoned fund managers or VCs. Like talk about a lot that like, hey, I'm this way into the fund. Should I take capital? I was just hearing kind of a 20 VC where they had a whole show about that. So I'm not going to get into that. But I think in my perspective, if you're in something that's that beautiful, that, that it's like scaling, it's working. Like, I'm in the camp of like, you know, be like having allocation into that is rare. And so you actually don't want to like, take out of it too fast because, like, you're gonna take out, you're gonna like, sort of get an exit from that, get some liquidity distribution, but then you're gonna have to put the money to work somewhere else. So if so my perspective, what I tell my LPs is my job is to help you get into those.
- [22:53]
Interjecting Commentator
Right.
- [22:53]
Suman Talukdar
Like, the fund has access to those through the fun. You maybe get more access to those through spv, sidecars, whatever, you know, Opportunity Fund. But I want you to stay in that because if you're able to find these companies that go from, you know, 100 million, a billion billion, I mean, just hold, right? Hold until like, something, something shows you otherwise. And, and I think, you know, it's interesting because I think sometimes fund managers, depending on the structure and how much capital they raise, I think there's incentives to, to, to not do that. And, and to me, it's like, okay, well, if I can, if I get an exit and I'm able to give those, that equity, that share to my LPs, and I'll let them make the decision.
- [23:46]
Interjecting Commentator
Right.
- [23:47]
Suman Talukdar
I think that's a win. That's a win. That's a philosophical view that I have, you know, in terms of the way that I, I look to find these sort of outliers.
- [23:58]
Interjecting Commentator
Right.
- [23:58]
Suman Talukdar
Because if I find one, right, like one and 30 or whatever are going to be one of those big ones. I just want to stay in it as long as I can. And then as you know. Right. Like, the gains usually happen in the tail end.
- [24:11]
Joel
Oh, sure, yeah.
- [24:13]
Suman Talukdar
Know, it's like, it's, it's, it's not the initial, like six innings. It's like the seventh, eighth, ninth, where things are really. So, so yeah, that's a bit of, kind of how I think about exits and, and, you know, AI 100%. Like, we're in that period now where it's. I don't think it's necessary about the exits. It's about, you know, putting your chips in the right sort of places.
- [24:38]
Interjecting Commentator 2
Yeah.
- [24:40]
Suman Talukdar
And then hopefully those things, like, massively grow.
- [24:45]
Interjecting Commentator 2
Sure.
- [24:46]
Joel
I mean, so let's talk a little more about where AI was when you got into it, you know, early in the, in the evolution of it and, you know, kind of where you've seen it. Obviously, we have LLMs. We have a Lot more, you know, platforms that are going to allow people to do things at the API level versus having to build your own LLMs. There's kind of like the stripe for AI essentially, right? So that's kind of what I've seen. I mean robotics, we saw the cyborg robot just tapping into OpenAI's API. So there's just a lot more ability to just kind of leverage open source libraries and open source platforms. And then there's other people that can build their own LLMs that are maybe more vertical, that they can distribute for a subscription. That's recurring revenue as well. But just unpack for us how you saw AI and where you've seen the biggest changes. I mean, it's probably much deeper than my surface level knowledge. And then where you think it's heading.
- [25:57]
Suman Talukdar
I think that the space is exploding so rapidly that, that I think very quickly your notion of an expert is changing. You know, I don't know who's like going to be an expert in everything at this point. But yeah, I mean, you know, like starting with that initial kind of post, my interest in AI, I mean back then it was like it was a science fair project, right? Like literally it was like science fair project and some amazing like geniuses, like, you know, the folks that I mentioned, you know, they were able to raise a lot of capital and like build crazy stuff, right? But it take a lot of capital before it really became something feasible. Like when Apple bought Siri, it wasn't really like generating any revenue. $200 million sale. So that's where it used to be, right? So it was like really like hard, difficult. And then I would say that it moved to, you know, with the advent of like Nvidia and like GPUs coming in and all that kind of stuff, like 2014, 15, it started to be like, this is interesting, it's starting to get a little bit easier. But it was still a little bit of like a science fair project that was expensive. That was like the 2015 through 19 period. And then, you know, I, I, you know, I've always been interested in it, so that's what got me into it. But since then things just move so fast and it was like unpredictable how fast it's moved. And I, I get the feeling that we're still underestimating like at every step of the way how fast this thing is moving. 2020, it started to be less about science fair and more like engineering and product development. And it didn't, and it wasn't like it didn't take you that much money to a product where you could be like, you know, two, three gals, guys in a room, like building something, you know, and you get a product that customers buy. And that's where I got into it. And I think. And so from that point in time where it became easier now, like, you basically have these massive platforms that are making it exponentially kind of more accessible to build really cool stuff, right? So, like the ChatGPT, these large language models, like applied into like 3D, whatever, all those things are just making it kind of really accessible for infinite number of applications. And so I think now, like everybody's saying like, oh, the space is exploding, like they're saying every year. And it's just even like more like every year. And so I do think that we're in an era now that it's moved from. I guess I felt like it's moved faster than I almost wanted it.
- [28:44]
Interjecting Commentator
Right.
- [28:44]
Suman Talukdar
Like, if it moved a little bit slower, I could sort of been like, I told you this, I was mentioning this. Like when I started, people were like, really like, AI ML, like, okay, you know, that's kind of like this, you know, like, why do you want to focus on that right now? You know, it wasn't like clear.
- [29:00]
Interjecting Commentator 2
Yeah.
- [29:01]
Suman Talukdar
And then now it's like, oh, every, like you, you know, we need to do this. But the, the plus of that is that you're going to get a lot of experimentation and, and you could, you know, as long as you're not just like putting all your eggs in one basket, there's going to be lots of different winners and it's going to be unclear. And it's ne. It's not necessarily clear who that person's going to be in the front end. So I think it's a very open landscape. Yeah, there's gonna be the huge players that raise a lot of money first to market that are gonna set themselves up. But after that I think there's gonna be. It's unclear, like, who the real return generators will be. And like, if you look at SaaS, like when Cloud came along, I mean, there were companies seven, eight, nine years after AWS set up or whatever that people didn't even know they were doing that well. Right. And I feel like AI is going to go through that transition now. You know, specifically for the fund, you know, we've invested in, you know, enterprise software. That's about most of the bread and butter. Then there's, you know, I was never shy to invest in things where to acquire the data you needed to have like some sort of a gadget, like hardware or whatever. But it wasn't about the hardware. It was like. So I did a bunch of computer vision stuff and, and that's been really interesting.
- [30:26]
Joel
So was it like an offline device? I've seen, I've seen some pretty interesting computer vision stuff where like there's, you know, like there's data that's collected on the edge. So you use your phone, you scan some type of image and then on the phone it's processing some of that imaging offline and then it pushes it to the cloud. But like, I think it's really powerful where you don't need the Internet. Like an elevator, sometimes there's not connectivity. When you're in an elevator, you might have to have some type of hardware that can collect that data offline. And then when it hits the network, then it pushes it out. It doesn't necessarily need to be real time. I think you just need to at some point collect the data so that it can sync and train the model when it catches the network again. But any opinions on that in terms of the edge and collecting data offline to kind of train the models?
- [31:18]
Suman Talukdar
Yeah, you know, I, I, what, what I saw from the entrepreneurs and I have one company that was a great example of this company called Hayden AI that's actually, you know, doing a lot of stuff in New York. And it's more about the total like, solution. And it wasn't like, okay, where I'm gonna process, it's more about how do I get this thing to work.
- [31:37]
Interjecting Commentator 2
Yeah.
- [31:37]
Suman Talukdar
And what they were doing is they, they basically entered the market. They figured out that right around the 2020 timeline, GPUs had gotten small enough and sort of cheap enough to be able to put in buses. And they literally put these things in buses. So now you have this processing capability on the bus. And now the buses are able to see the road and they are able to recognize license plates and lots of interesting stuff happened. And then they figured out the whole architecture from that. So it was more driven by, okay, what's that problem? Right, like that they're trying to solve with the AI. But yeah, 100% like the computational power, whether it's on the cloud, right. Like data centers, or whether it's going to be distributed in your laptop, on the edge or some device or whatever. I mean, it's just proliferating.
- [32:28]
Interjecting Commentator
Right.
- [32:29]
Suman Talukdar
It's becoming ubiquitous. And so whether it's trained a thing, you can do it cheaper, or whether it's actually run the model, I think you know, developers are going to have tons and tons of horsepower out there. But it's again, it's, in my perspective, it was like the stars that align so that as a developer you could start to find these problems, right, that you could solve. And now you didn't need to raise like $10 million to solve them. And that's been the reflection and I think now you can even do more.
- [33:05]
Interjecting Commentator
Right?
- [33:06]
Suman Talukdar
Because I was sort of market pre with the fun one. Now I'm on my second fund. But like pre chat GPT and, and then so post chat GPT now you have all these things that chat GPT can help you with. So I think you're gonna have kind of more like what's always been the case, not just one approach, but like a mixture of approaches. And the founders are going to call it their own thing, but it's really what's been existed through the years. When people are doing these old school AI systems, they're trying all sorts of different techniques or approaches and seeing how it worked together. I think it's a great time, Great time.
- [33:53]
Joel
I mean, there was a founder of ufc, I think his name is Dana. So he was just saying how he was suffering from sleep apnea and you know, he was just saying that he doesn't even go to doctors for his regular checkup. I think obviously if you need surgery or you, you have something serious that's going on with you, you definitely want to talk to a doctor and, and get seen. But just to check your blood work, you know, a lot of those stats, there's some type of recommendation that you can get based on the stats because the stats are essentially just a value. So that denotes if it's normal, if it's at risk or if it's borderline. And then there's some type of logic or some type of recommendation that could be generated. Like, and I've complained about this in the past, I mean I've gotten my blood work done and I like would DM my doctor. I had to DM him in the app and wait for him to respond. And his feedback was super generic. It's just like, oh, you know, cut down sweets or cut down, you know, these, these, you know, cut down fast food or something like that. Right. So I think the bandwidth of human physicians is getting thinner and thinner and I think the required output is increasing. So that impacts quality over time because you're getting more pressure to see more patients and have more throughput. And I think there's a sacrifice of being able to respond to all your patients and give them the care that they deserve. So if some of that could be augmented by just saying, look, you know what, based on your ethnic background, you know, this is kind of the normal state that you should be at as a benchmark and you know, based on your weight, your height and just based on maybe your family's history, you need to step up these things. And here's some behavior that you should improve. And I think if you could track that dynamically, almost like, you know, like we do in finance, right? We have betterment and we have all these dashboards to improve, you know, kind of how we're managing our spending. I don't think there really is something out there unless I'm missing out in terms of kind of dynamically tracking your health and like knowing what healthy means. Right. You have a bunch of stats that you get when you do your blood work, but you don't really know, really know what that means. If there's some type of easy type of benchmark that helps you accountable because a doctor is not a typical doctor is not going to have that time to be able to do that.
- [36:17]
Interjecting Commentator 2
Right?
- [36:18]
Suman Talukdar
Yeah.
- [36:19]
Joel
So I think there's a huge opportunity in the healthcare space to, you know, recommend that. I know we're talking about agentic AI, so having agents do different things, I think that's kind of interesting too. So I don't know if you have any. I'll stop there. And I wonder if you have any reactions to what I said.
- [36:38]
Suman Talukdar
Yeah, no, I think healthcare, you know, just like other areas, you know, like in addition to three, like climate tech. I mean there's this notion of sort of like being AI being kind of in the mix of things, but really it's AI adding kind of a human in the loop, you know, that really amplifies the end output. So I think that's one way to think about it. And so health, like if there's like sort of me in the loop, right, like with, with the doctor and, and data, like I was talking to a gentleman and they have, you know, they have these like patches that people wear, but you can wear them. He had kind of adapted one that people wore for, for diabetes, but he was like, I'm just wearing it for my own health because I just measuring my keeping tabs. But I was like, that's really interesting because like you're getting that data and then now imagine that data is going into your, you know, like your Kaiser or whatever and then like you're getting like real time monitoring. The AI is kind of Telling you stuff and if it goes above a certain level like your doctor gets in. I mean those things are like no brainers.
- [37:42]
Interjecting Commentator
Right.
- [37:42]
Suman Talukdar
And that type of system, I mean it's actually can be applied to anything. So I've got a company that is doing that for like heavy equipment, machinery.
- [37:51]
Interjecting Commentator
Right.
- [37:52]
Suman Talukdar
Like if you're a petrochemical plant you don't want your things to break down. Right. Because it costs like millions and hundreds of millions of dollars in like lost production time. So they basically have like sensors that they measure and the things about to go down. It's going to be like hey timeout. You know, they send alerts, they ask remediation. So I think that as a sort of approach can be 100 health. All these different areas where you sort of need to sort of intervene in a smart way and then hopefully get you on the right track are definitely big opportunities. And yeah, I mean you know with the agentix stuff the, everyone's into the LLMs basically know.
- [38:36]
Interjecting Commentator
Right.
- [38:36]
Suman Talukdar
They know things.
- [38:37]
Interjecting Commentator 2
Yeah.
- [38:38]
Suman Talukdar
Ask them questions they don't do and, and they have a hard time doing. So I think that's where like the agentic paradigm is coming in. It's now all of a sudden these things will know and then with that knowledge you're gonna be able to in a way teach them what they should do.
- [38:58]
Interjecting Commentator 2
Yeah.
- [38:59]
Suman Talukdar
And obviously that's going to be a multiplier.
- [39:01]
Interjecting Commentator
Right.
- [39:02]
Suman Talukdar
If I can teach them to do the right things then it's going to save up my time. But I got to find the right slices.
- [39:08]
Interjecting Commentator
Right.
- [39:09]
Suman Talukdar
So if I try to teach them to do the thing too early then the customer is not going to be happy about it and, and, and then it's not going to work.
- [39:19]
Interjecting Commentator
Right.
- [39:19]
Suman Talukdar
And, and a lot of like what I've seen is companies trying to sell that their sort of middleware under the hood thing is better than somebody else. And I think that's really hard because. So I think we're going to go through this time period where a lot of people are going to get excited about this new approach method but then at the end it's going to be the ones that actually work, that actually deliver the thing. But yeah, moving from knowing to actually doing is very exciting.
- [39:56]
Interjecting Commentator
Right.
- [39:56]
Suman Talukdar
I think that's step in and then translating that to be able to do it in a way that's really easy. I can type something and not have to spend hours and hours figuring something out. That would be super powerful.
- [40:13]
Joel
Yeah, no, that's really helpful. And I know we got about 10 minutes left so I Want to talk a little more about firm building? You know, walk us through kind of the origin story. After being a successful operator, kind of seeing these exits. What was the origin story of building your first fund? And you know, what advice would you share from, you know, now graduating to kind of thinking about your, your second fund?
- [40:40]
Suman Talukdar
Yeah, you know, I. The origin story was I kind of lifted the Internet. Like I moved out here. I kind of saw how that played out. And then I experienced the cloud. And when I was experiencing the cloud, like post business school, I saw a bunch of funds that were built starting out first, writing like angel checks. It was a pretty long list. And when I felt like there was a time for AI and it was kind of the right time in my career to pursue this, I've been trying to think about ways to like build a legacy, right? Not, not in some kind of grandiose way, but just like, you know, do something really meaningful that leaves a mark in some sort of a significant way. I was like, hey, like I've been doing this angel stuff for a while. If I can turn this into fun and turns into philosophy around like what I look to invest in, the types of founders I look to support, the way I want the companies to be built in this sort of time when AI is about to like explode or I think, I think that's like a once in a lifetime opportunity.
- [41:47]
Interjecting Commentator 2
Yeah.
- [41:48]
Joel
This is also provocative question as you're talking about firm building. Look, I've seen a lot of AI funds even just come through our past cohort. So how can fund managers as they're building their firm not only focused on AI, but just differentiate themselves? You know, I've had a handful of healthcare funds come through our program a lot, you know, handful of AI funds. So what are some things that you should think about as a fund manager when you're building your first fund? How have you seen great funds, including yourself? Like, just differentiate yourself.
- [42:24]
Suman Talukdar
You know, I think it's not an easy answer because especially when you have such a big opportunity space.
- [42:32]
Interjecting Commentator 2
Yeah.
- [42:33]
Suman Talukdar
Like, AI is so massive.
- [42:34]
Interjecting Commentator
Right.
- [42:35]
Suman Talukdar
And so I think the answer is, you know, find the focus that fits you that's like true to you, and then just like understand that when you have that focus, you're going to find things that are like perfect, like right in line with your thesis. And there's other things that are going to seem exciting, but they're going to be like not quite what you're about. And so I think that's my approach to differentiating has been that it's like, kind of like the Bruce Lee thing. Like, what's your one kick? You know, it's like, it's, it's like. So I think it's a small, like I'm a micro vc, small fund manager, I'm running billions. I think it's about, for me, it's what's my one sort of key thing that I look for. And it's always been kind of this like, idea of like breakthrough AI with traction, which is not necessarily an easy thing to find because sometimes when you break through, it means you need a lot of money, right? And sometimes when you're trying to get traction, it's like, well, you might not really be differentiating the technology or product as much as you could. So I think for me it's been that. But I would say that regardless of what fund you have, whether it's like a robotics fund, like you say you're just doing robotics, or just health or just climate tech, I think true to finding what that one thing is that fits your experience is, is I think the right hypothesis around getting that answer.
- [44:00]
Joel
And to quote your blue and to quote your Bruce Lee quote, because I just looked it up. I fear not the man who has practiced a thousand kicks once, but I fear the man who has practiced one kick 10,000 times.
- [44:12]
Suman Talukdar
Yeah, I, you know, I 100% believe that. I, I think it's, you know, and in sports and stuff, you see that, right? Like, you don't have to five different things. I mean, over time people develop expertise, but, you know, like X player, 10, I'm tennis player, whatever. It's like a great forehand and like they get, they get great by having that and that. And I feel like when you have limited resources and you're kind of a new entrant and there's just like a finite space, got to find the one or two things that you do really well. And I think differentiation comes from that. Otherwise it's positioning, which is like, is it really real?
- [44:49]
Interjecting Commentator
Right?
- [44:49]
Suman Talukdar
Like I'm going to position myself this way or that way give you an advantage, I don't know. But if it's like, you know, that's kind of my, my current.
- [44:59]
Joel
And I would say the final decision because, I mean, look, there's, there's performance data that is like table stakes. There's, you know, your access to your deals. Let's say you got all these great deals so you could have all these great things. You have a great path pedigree as well. You worked in private equity, you went to Harvard. So all Check, check, check, check, check, check. Right. But at the end, do they like you and do they want to stick around and work with you for the next 10 years? That's not something that is really that quantitative. And there's so many people that have already gotten the pedigree, they've got the track record, they've got, you know, all these hot deals, you know, so what really tips them over at the end is just like, look, you know what? I really like Schuman. And I, I think he's just a good person and he's just someone that has the same values I do. And I think I'm gonna, I think I'm gonna go. Go ahead and, and take this, make this commitment with this person. And I think that in my mind, I think for all sectors, you know, they've had so many climate funds a couple quarters ago that, that went through, you know, several different transitions. Right. I've seen some climate funds pivot into like, AI climate. So I think there's definitely pivots in the investment management space on strategy. But I think at the end of the day, I think the people that do the best are the people that can authentically maybe do storytelling and help people understand what their mission is over time. And I think a big thing too is there was an LP that came on last week. It's also kind of the succession plan. So I think for you, kind of also thinking about having a bigger mission. You said you wanted to have some type of legacy, so, like, you know, it sounds like you've already kind of thought through, like, what the next couple vintages are going to be that also play on top of the evolution of AI.
- [46:44]
Suman Talukdar
Yeah, no, for sure. I think to just respond to a couple of those points, like connecting, being able to connect to great founders. When you meet kind of like that next, like Elon Musk or whoever that is, and that skill is like number one. Because at the end, my job is to be helpful.
- [47:04]
Interjecting Commentator
Right.
- [47:05]
Suman Talukdar
So I think that's been something that I've learned through the years, just through my operating roles of like working with founders. But yeah, I mean, you know, in terms of the longer term outlook, I think, you know, when I started this, I was like, hey, AI is just happening first time in the history of mankind, which is crazy.
- [47:23]
Interjecting Commentator
Right.
- [47:24]
Suman Talukdar
I want to invest in like a hundred companies.
- [47:27]
Interjecting Commentator
Sure.
- [47:27]
Suman Talukdar
And I feel like if I invest in a hundred companies, you know, and like five of those are going to be, like, working. I think those five are going to be like, big, you know, so it's like, so I think. And if those five are going to be big because they're really like doing something interesting.
- [47:44]
Interjecting Commentator
Right.
- [47:44]
Suman Talukdar
That's going to change some trajectory for, for some industry or the way that humans do something. So I think the first legacy is that just like being able to like find those hundred companies and be able to shape the ideas that are being created out of this technology.
- [48:03]
Interjecting Commentator 2
Yeah.
- [48:04]
Suman Talukdar
And doing that well, hopefully kind of returning capital to my LPs showing that there's like a system behind kind of the approach that's like repeatable then I think, yeah. How does this become a longer term kind of thing without me will be, will be clear, but I think that's down the road for a long time. I think what's interesting is I've sort of started the fund at a time where I can use AI for myself. Your funds could not have.
- [48:39]
Interjecting Commentator 2
Sure.
- [48:39]
Joel
I mean you could do. You know, it's optimizing a lot of workflows as well. Right. So like, you know, you had to hire analysts and now you can upload a deck and it could write a memo for you. I know a family office that has upgraded to the premium version of Claude and Chat GTP and they've been able to combine some of the features of both and replace their headcount by two. So they've like, they've actually eliminated, not eliminated the analysts, but like eliminated the need for more headcount. So I think there's definitely some augmentations that can help. But what you can't replace is just going out to cocktails and building a relationship with people, talking about their family and bonding with people. So I think that's still there until, you know, civilization gets replaced with cyborgs that can do the same thing. Maybe, I don't know, like there's a. Yeah.
- [49:31]
Suman Talukdar
I mean I, I believe the relationship thing is the key component.
- [49:34]
Interjecting Commentator 2
Yeah.
- [49:36]
Suman Talukdar
All the other stuff, it's going to be hard to be way better than anyone else. Like you can automate this like in my automation is better than yours. Like. Okay, I don't know. But like the automation is going to supplement the key thing which is like the human relationships.
- [49:51]
Interjecting Commentator 2
Yeah.
- [49:52]
Suman Talukdar
Like your instincts. And I think that's the longer term sort of play, especially for smaller boutique shops. I think the more you can stay true to your kind of like the intangibles, that's what will give you an advantage. Because people I still think will want to work with people and like you want to be that first call, you want to be that first break for an Entrepreneur because they feel like when they get your check in any way possible into their cap table, all of a sudden people look at their company differently. Like, I've had that impact from my LPs, like some of my early LPs, like when they came in, they're like, or my current LPs, people are like, oh wow, they're in. And it's, it's, that's, that's an amazing thing, right? To be able to have that type of value. And I think that's not something that automation is going to take away and you got to build that over time.
- [50:48]
Interjecting Commentator 2
Sure.
- [50:49]
Suman Talukdar
So I think that's, that's very much like an old school advantage. But the difference is now I can, my, my data set is not just like 500 companies, right? It's like thousand companies, right?
- [51:02]
Interjecting Commentator 2
Like, yeah.
- [51:03]
Suman Talukdar
And in some way I'm digesting that data differently, but I still know like, okay, where I should spend my time, where I should go. But, but that network that you built over time, like, you know, meeting you, knowing you over the years and like compounding that and then building a reputation, I think that's an intangible that's really hard to automate away.
- [51:28]
Joel
And I think there's some high stakes decisions that still, at the end of the day, the trigger pulling comes on feeling. And I think the two are really romance and finances, right? So when you're thinking about an investment, you looked at all the data, right? There's data that you're synthesized, you got some feedback from a friend of yours that invested. Obviously they may have a little bit of a bias because they invested, right? So they don't want to say this is a horrible deal, I invested and they're trying to get you to. So there's some bias there. Then you've got some of the, you know, the stats, right? The revenue, the, you know, if you're looking, if it's a consumer company, you're looking at like the, you know, cost to acquire customers. So there's all this hard data and then, you know, there's maybe some things that you've observed from the founder that could be a yellow or green flag. But then beyond that, at the end of the day there's a final feeling, right? It's like, you know what, I think based on all this, I'm going to decide and same thing I think when it comes to romance, right? Like, I mean, if you're trying to, you know, choose, you know, someone that you want to kind of, you know, go in the long haul with or even go on a date with. There's some type of feeling where, like, if you're just going to, you know, choose a sandwich, maybe it's just less of a feeling and it's more logical. It's like, look, you know, the Chipotle is right around the corner, whatever, I'll just go ahead and do it, you know, So I think there's definitely some type of, you know, instinctual feeling that, that is finally the, the decision maker. But again, that's just my opinion. So.
- [52:57]
Suman Talukdar
Yeah, and, and you know, we, the, the math has. It kind of helps you with that.
- [53:01]
Interjecting Commentator
Right.
- [53:01]
Suman Talukdar
Like, I think you're not just trying to make one, you're trying to make a bunch of them.
- [53:05]
Interjecting Commentator 2
Yeah.
- [53:06]
Suman Talukdar
And, but yeah, you know, it, it depends like, like whether you're, you know, the only check or like, I personally like to be in situations where I feel there are others in the room that understand the space, you know, even better, more seasoned. And I look at them as just another sort of great ally for me to help vet. But at the end, but yeah, you know, you get a gut reaction based on everything that you gathered, whether this is a situation you want to be in longer.
- [53:40]
Joel
Yeah, I mean, it's just, look, it's the same thing. The, the consensus is the same thing at the founder level and at the LP level. Right. The found, you know, like as a vc, like, oh, who are your, you know, is. Is. Is lightspeed coming in in this one? Or, you know, who, who are your. Who's leading the round? And then I, you know, even at the LP level, it's like, oh, who, who's kind of your first couple early LPs, are they well known founders or are they. Is it. Do you have like one fund to fund coming in? You know, so definitely it's, you know, definitely kind of seeing some type of social proof and I think getting those first people to kind of come along is definitely probably the challenge. I mean, it's challenging the whole way throughout, but like just kind of getting a little bit of like product market fit in the beginning definitely helps. So. Yeah, anyways. Yeah, well. Hey, Shimon, this is amazing. Thanks for staying over a little longer. I could probably go another hour with you, but I want to respect your time and next time. Yeah, I appreciate you being generous with all of your knowledge and it was great for the community to hear more about your story. So thank you again and appreciate the friendship.
- [54:42]
Suman Talukdar
No, thank you, Joel, for having me. And like I said, it's. I very much appreciate being part of your community.
- [54:50]
Joel
Yeah, absolutely. It's a lot of fun. So we'll catch up soon.
- [54:53]
Suman Talukdar
All right, take care.
- [54:53]
Joel
All right, take care, everybody.
- [54:55]
Interjecting Commentator 2
Bye.