# Founder10x - Can AI Beat The Top Investors? - Jan Szilagyi, Co-founder, Reflexivity

An episode of [VC10X - Investing, Venture Capital, Asset Management, Family Office, Wealth Management](https://invisiblenetwork.ai/podcast/1632806986), 28 Oct 2025, 35 min.

Audio: https://anchor.fm/s/a453ddd4/podcast/play/110336529/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2025-10-10%2F412203323-44100-2-24bb812db3897.mp3

Jan Szilayi is the co-founder of Reflexivity, a AI market research platform that is quietly becoming the secret weapon for some of the world's top investors. Backed by legendary figures like Stanley Druckenmiller, Thomas Peterffy, and Greg Coffey, and partnered with Microsoft, Reflexivity is already being used by major hedge funds like Millennium and Soros Fund Management to navigate the overwhelming flood of market data. In this conversation, we explore how AI is evolving from a simple data tool into an analytical partner that helps investors ask questions they didn't even know they should be asking. We discuss the key principles from legendary investors that are baked into the platform's DNA, and we tackle the big question: Could an AI one day compete with the best asset managers on the planet? ⭐ Sponsored by Podcast10x - Podcasting agency for VCs - https://podcast10x.com Reflexivity website - https://reflexivity.com/en Jan Szilagyi on LinkedIn - https://www.linkedin.com/in/jan-szilagyi-12284ab 🎙️ Hosted by Prashant Choubey Follow Prashant on X - https://x.com/ChoubeySahab Follow Prashant on LinkedIn - https://www.linkedin.com/in/choubeysahab Timestamps: (00:00) Introduction to the podcast and Reflexivity (03:26) Johan's background and motivation for building Reflexivity (04:25) Overview of the Reflexivity platform (05:44) Differentiating Reflexivity from other AI investment tools (07:50) Market focus of Reflexivity (public vs private markets) (08:51) Key principle learned from Stanley Druckenmiller (10:43) Types of research possible on Reflexivity (13:02) Democratization of market alpha through AI (14:53) Convincing top investors to use Reflexivity (15:41) Influence of academic training on company building (16:54) Managing a long-distance founder partnership (18:09) Comparison with horizontal AI companies (19:38) Strategic partnership with Microsoft (20:41) Go-to-market learnings and challenges (22:19) Potential for an autonomous investment agent (26:32) Fundraising strategy and investor selection (28:16) Capital allocation approach (29:18) Comparing market prediction to running a startup (30:55) Perspective on AI hype and market valuation (33:05) Biggest learning from building Reflexivity (33:48) How to learn more about Reflexivity For sponsorship or guest appearance requests, write to prashantchoubey3@gmail.com Subscribe to VC10X on Youtube, Spotify, Apple Podcasts.

## Transcript

[0:00] There's also unprompted AI, so even in absence of you actively
[0:06] asking questions, the system is always monitoring a vast amount
[0:10] of data that it's linked to and will be helping you ask
[0:14] questions that you should be asking but aren't.
[0:17] How do you convince the smartest investors on the planet that you
[0:20] have something that they need? Really, just by showing it to
[0:23] them, I think that they convince themselves.
[0:26] Do you think you are competing directly against the bigwigs or
[0:30] the horizontal AI companies like the Open AI or Complexity?
[0:33] We're trying to tackle all these different problems and also now
[0:36] coming up with vertical specific products as well.
[0:38] Let's say Complexity finance is maybe a direct competition to
[0:42] you as well. There is a big difference
[0:44] between how we apply and leverage LLMS versus how you
[0:50] would be using them in a platform like JGPT.
[0:54] You work with the legendary Stanley Druckenmiller.
[0:58] What's one key principle from him that's now a big into your
[1:01] platform's DNA? A big part of the approach that
[1:06] was cultivated with under his mentorship was this focus on not
[1:13] just the micro but also the macro.
[1:16] Do you see if Flex City building an agent in future that's going
[1:19] to compete with the top asset managers for the kind of returns
[1:22] it can generate on its own? You know, I think that.
[1:27] Hey everyone, this is Prashanth and I'll be host for the VC10X
[1:30] podcast. And today we have Jon Zilai with
[1:32] us. Jon is the cofounder of
[1:34] Reflexivity, AI market research platform that is quietly
[1:38] becoming the secret weapon for some of the world's top
[1:41] investors. Backed by legendary figures like
[1:44] Stanley Druckenmiller, Thomas Petterfee, and Greg Coffey and
[1:48] partnered with Microsoft, Reflexivity is already being
[1:51] used by major hedge funds like Millennium and Soros Fund
[1:54] Management to navigate the overwhelming flood of market
[1:57] data. In this conversation, we explore
[1:59] how AI is evolving from a simple data tool into an analytical
[2:02] partner that helps investors ask questions they should be asking
[2:06] but aren't. We discussed the key principles
[2:09] from legendary investors that are baked into the platform's
[2:11] DNA, and we tackle the big question, could an AI agent one
[2:16] day compete with the best asset managers on the planet for
[2:19] generating highest returns? So without wasting any time,
[2:22] let's dive straight in. This episode is sponsored by my
[2:25] own podcasting agency called Podcast NX where I help venture
[2:29] capitalists, asset managers, allocators and family offices
[2:32] and starting their own podcast and building a strong brand.
[2:36] Podcast NX is designed in a way to work.
[2:38] Well, with the busy. Schedule of investors where we
[2:41] handle every aspect of podcast production so you can focus on
[2:45] investing. If you're looking to build a
[2:47] strong investor brand, starting a podcast should be right up
[2:51] your priority list. It's the only form of marketing
[2:54] where you can showcase your strong network.
[2:56] In other words, access to limited partners and founders
[3:00] and grow that network further. And contrary to what most people
[3:03] hear, it's still. Very early.
[3:05] To start a podcast, there are at least 10X more newsletters and
[3:09] thousand X more blogs than there are podcasts in this world.
[3:13] You can learn more about our podcasting services at podcast
[3:16] ten-x.com. I'll make sure to add the link
[3:19] in the description below. Enjoy this episode.
[3:22] Hey, Jan, so good to have you on the VC10X podcast.
[3:24] How are you doing? Good.
[3:26] Thanks for having me. Really good to be here.
[3:28] It's a pleasure having you on to start things off.
[3:31] Can we start with what was the specific I've had in that moment
[3:36] as an investor that made you decide to build reflexivity?
[3:41] Yeah. I don't know if I can pin it
[3:42] down to an exact moment, but it definitely was this feeling that
[3:46] as an investor, you always were inundated with data and
[3:51] information and your job was obviously to mine that
[3:54] information for either opportunities or risks to your
[3:58] portfolio. And it felt like the
[4:00] infrastructure available to do that was really outdated.
[4:04] It always felt like you were effectively behind and unable to
[4:09] unlock all of the different insights that seem to be
[4:13] residing right there in front of you in the data.
[4:15] Yeah, absolutely. And it's quite interesting what
[4:18] you're building. So can you give us an insight
[4:20] into what the platform reflexivity actually does in
[4:24] practice? Yeah, of course.
[4:26] So reflexivity, the best way to think about it is plug and play.
[4:30] Zero hallucination investment analysis environment that's
[4:35] powered by two key pillars, which is the knowledge graph and
[4:40] the reasoning layer of large language models, more than a
[4:43] single one. And it enables the user and
[4:46] investor, A portfolio manager and analyst, A wealth advisor to
[4:51] have unprecedented discoverability of the data.
[4:55] So we integrate a number of different data sources and also
[4:58] what I guess I would refer to as analyzability of that data.
[5:02] So we rely on large language models ability to understand
[5:07] what is it that you would like to get done and link that with a
[5:12] wealth of data, as well as a very, very powerful analytical
[5:16] engine in order to be able to really help you dig deeper
[5:20] around ideas, thoughts, risks, scenarios and so on.
[5:24] Absolutely. And I believe that a lot of or
[5:29] almost every investors these days would be using some sort of
[5:32] AI in the sort of evaluation practices or doing diligence and
[5:36] stuff like that. How do you think the flexivity
[5:39] would actually be something different and something better
[5:42] for them to use? Yeah.
[5:44] So you're right. I think there is a lot of, I
[5:47] think there is very strong appetite in the investing world
[5:51] to try and figure out ways in which you could leverage AI for
[5:54] this purpose. I would say in practice, it's
[5:58] proven harder than originally thought.
[6:00] And so a lot of the solutions that we currently see when
[6:04] people are in fact using AI are either confined to a structure
[6:10] data, so document summary, document query and so on.
[6:14] Or alternatively, they will be using platforms like ChatGPT and
[6:18] so on for relatively trivial research questions, meaning, you
[6:22] know, can you summarize this industry and so on.
[6:24] What we are trying to do instead is a lot more ambitious, which
[6:29] is give you the ability to actually conduct very complex
[6:34] analysis with reflexivity as well.
[6:37] So go beyond summarizing or addressing some very, very
[6:42] superficial questions. And secondly, in addition to
[6:45] what I guess I will call prompted AI, so where you come
[6:47] in and you are asking reflexivity specifically to help
[6:51] you dig through a particular piece of information, there's
[6:55] also unprompted AI. So even in absence of you
[6:59] actively asking questions, the system is always monitoring a
[7:04] vast amount of data that it's linked to and will be helping
[7:08] you ask questions that you should be asking but aren't.
[7:11] So in addition to helping you ask your answer your questions,
[7:15] it's also helping you think of better questions, which I think
[7:18] is a much, much harder challenge, but really
[7:21] appreciated by the investors who are using it.
[7:24] Absolutely. So it's basically the investment
[7:26] analyst as an agent in a way that it's going to do the job of
[7:30] the analyst, but it's AI, right. Interesting.
[7:34] Is it only, only relevant for public markets or is it for both
[7:38] public or private markets? I, I think we have had products
[7:41] like Perplexity finance, which had more focused on the public
[7:44] markets. You can do a lot of analysis
[7:46] there. So tell us about what's the
[7:48] focus for reflexivity? Yeah.
[7:51] The focus is at the moment certainly more on public
[7:53] markets, on across different asset classes.
[7:56] So in addition to publicly traded stocks, the system is
[7:59] also capable of handling and monitors currency crosses, fixed
[8:04] income instruments, indices, commodities and so on.
[8:07] So it's very, very broad in its coverage.
[8:10] The idea is, however, that the framework for this kind of
[8:15] analysis is really not limited to public markets.
[8:18] In other words, to the extent that you then plug in data that
[8:21] would be more pertinent to private markets, private
[8:23] companies and so on, the same framework would absolutely
[8:26] apply. The reality for us as a as an
[8:30] emerging business simply was that the wealth of information
[8:33] and the appetite for such solutions was extremely strong,
[8:37] particularly within asset management and investment
[8:40] management, which is why we focused on that first.
[8:43] Your work with the legendary Stanley Druckenmiller.
[8:47] What's one key principle from him that's now big into your
[8:50] platform's DNA? So a big part of the approach
[8:58] that was cultivated with under his mentorship was this focus on
[9:05] not just the micro but also the macro.
[9:08] So for example, if you were looking at and evaluating a
[9:11] particular security, often people will know virtually
[9:16] everything there is to know about that particular stock,
[9:20] right profit margins, what management is doing, the product
[9:23] road map and so on. What I really appreciate it and
[9:27] what really differentiated the approach that he had was this
[9:30] ability to also take more of a macro approach.
[9:34] So look at the environment in which this business was was, was
[9:37] trading, and then usually identify one, maybe two key
[9:43] drivers that are really driving that stock price.
[9:46] So instead of getting lost in this forest of different things
[9:49] that could be moving and then having no idea how to combine
[9:53] them into a coherent bullish or bearish view, he was much more
[9:58] focused on looking at the data, but then trying to identify a
[10:02] single driver that he felt like was going to drive this specific
[10:08] this specific security. And so with reflexivity,
[10:11] organizing the way we have the data currently, that process is
[10:17] a lot easier. You're really able to 0 in on
[10:20] specific bits, specific drivers, specific connections, specific
[10:24] sensitivities that seem to be making the main difference in
[10:29] whether or not the price of a security today will be higher or
[10:32] lower. Yeah, that's interesting.
[10:35] And can you give us an insight into the kind of research that
[10:39] can be conducted inside flexivity?
[10:42] Yeah. In fact, you know the the range
[10:46] and the the breath is extremely wide.
[10:50] Just to try and give you a glimpse of that range and
[10:54] breath, you can conduct, for example, a full on financial
[11:01] model analysis in the sense that you could say, OK, I'd like you
[11:04] to access all the financial statements for a particular
[11:08] security for a particular stock and help me understand why is it
[11:12] that it trades at a discount to its peers.
[11:15] And so the system will look at the financials for this company,
[11:19] the financials for all of its peers, compare and look at the
[11:22] valuation, and then usually identify things that, based on
[11:26] the publicly available information it has, might be the
[11:29] reason for the gap in valuation between your stock and other
[11:31] stocks. So that's one example.
[11:34] Another example would be that you have a portfolio of
[11:37] securities and you are worried about some downside risk that
[11:40] might be stemming from geopolitics, maybe they're
[11:42] stemming from just overall market risk and so on.
[11:45] And so you could have reflexivity help you think
[11:48] through some hedging structures for that portfolio that would
[11:52] limit the downside risk subject to how much are you willing to
[11:56] pay for that insurance and how much downside you're willing to
[11:59] accept, which obviously will be to sort of competing, competing
[12:03] parameters. And it is very much able to
[12:08] handle that, including performing, let's say, Monte
[12:10] Carlo simulations and so on. So it really feels like you've
[12:13] been given the power of a supercomputer that's residing
[12:16] right there on your desktop. And the surprising thing, I
[12:20] think when we first see users experience it is that they
[12:26] aren't always aware of just how much power the system has.
[12:33] So they'll start with questions that are much more in line with
[12:38] what you'd expect from ChatGPT. And then you sort of dig deeper
[12:41] and you're like, oh, wait a second.
[12:44] This really does give me the ability to perform very quickly
[12:48] some calculations that previously I would have given to
[12:50] my analyst and I would probably then not see him for the rest of
[12:52] the day. Absolutely.
[12:54] Quite interesting. So does it do like reflexivity
[12:58] ultimately spread market alpha around or does it just
[13:01] concentrate it for the at the top?
[13:04] Well, I guess it's going to concentrate it for those that
[13:05] have reflexivity. So we'll see.
[13:08] We'll see whether or not that's spreading it obviously in our in
[13:11] our mind, the ideal cases where it then gets democratized across
[13:15] the across Wall Street. However, joking aside, I do
[13:19] think, you know, a system like this one is not going to
[13:25] magically convert you into a super investor, but it's going
[13:29] to make everything that you're currently doing better, faster,
[13:34] more efficient. So I think everybody becomes a
[13:36] better investor on the back of this.
[13:38] It's also true, however, that the quality of your answers
[13:44] depends a lot on the quality of your questions.
[13:47] In other words, it still is a human in the loop kind of
[13:50] system. And as such, it does mean that
[13:54] what you are going to get out of it depends a lot on what you put
[13:57] into it, right? Like it's not magically going to
[14:00] transform poorly, poorly worded questions or let's say analysis
[14:09] that it's remains very superficial into something that
[14:11] becomes a lot more insightful without you actually using it,
[14:15] relying on it, having interesting ideas and so on.
[14:17] So it does help, but it's not it's not a magic bullet.
[14:23] Yeah, absolutely. Yeah, surely, I think right now
[14:25] AI is at the stage where human in the loop is quite
[14:29] significant. Maybe it changes later wherein
[14:32] AI can do more of the hard work. But right now I think prompting
[14:37] actually depend, it fairly influences a lot of the kind of
[14:41] results that you can get out for AI platform work too, right?
[14:45] Your clients are some of the world's top hedge funds, right?
[14:49] How do you convince the smartest investors on the planet that you
[14:51] have something that they need? Really, just by showing it to
[14:55] them, I think that they convince themselves like I think it
[14:58] becomes the case for this is so strong that I don't think you
[15:02] really need to do a lot of convincing.
[15:03] Again, as I mentioned at the outset, the appetite I think to
[15:07] use AI in your workflow is very strong.
[15:11] I think there isn't a lack of options and knowledge internally
[15:16] for people to build their own systems at the moment.
[15:19] They see the full potential and the capability of one that's
[15:22] built. Like reflexivity.
[15:24] It doesn't require much more than that.
[15:27] Absolutely. So something I've learned is
[15:30] that you have the fastest ever Howard Howard Economics PSD,
[15:34] right? How did that deep academic
[15:37] training influence the way you build this company?
[15:41] I think the main take away for me from Graduate School was it
[15:47] really gave you a large number of mental models that you could
[15:52] apply to solving problems both within the business and also
[15:56] product building. And So what I would say is that
[16:00] although the actual models were obviously academic in nature,
[16:04] the way you would structure your thinking around that particular
[16:07] challenge was very helpful when you were able to then apply
[16:12] principles of, let's say, behavioral economics, or if you
[16:14] could rely on some of the financial models that have been
[16:18] built in order to understand risks and so on.
[16:20] And I guess perhaps more trivially, it also gives you a
[16:24] way to understand the language of not only investors, but also
[16:29] strategist, economist, and so on.
[16:30] So you're able to understand how they might be thinking and
[16:33] therefore when you are telling them about reflexivity, explain
[16:38] things in the context and in a language that is very familiar
[16:42] to them. Absolutely.
[16:44] And you're building this with your former Co CIO who is in
[16:48] London, right? How do you make that long
[16:51] distance founder partnership work?
[16:54] You know, that really never has been an issue, I think.
[16:58] And this is particularly true, I would say, for everybody.
[17:01] Post COVID, we have all gotten used to this hybrid existence
[17:07] where meetings can take place either over Zoom or in real
[17:12] life. In fact, you now need to
[17:15] actually be very explicit if a meeting is going to be in person
[17:18] because the default assumption is that it'll be online.
[17:22] So for us, that hasn't really been an issue.
[17:24] And I think as the organization grew, a lot of our colleagues
[17:27] and partners are in remote locations where we really
[17:32] interact with them, like you and I are now.
[17:35] So that has not been an issue for reflexivity.
[17:39] And do you think you are competing with, you are, I would
[17:43] say a vertical AI company, right, targeting a specific
[17:46] vertical by using the power of AI?
[17:49] Do you think you are competing directly against the bigwigs or
[17:53] the horizontal AI companies like the Open AI or publicity?
[17:57] We're trying to tackle all these different problems and also now
[17:59] coming up with vertical specific products as well, let's say
[18:02] publicity finance is might maybe another competition to you as
[18:06] well. So what's your take on that?
[18:09] I would say for a kind of a very basic layer of questions, there
[18:13] might be a level of overlap, but the architecture and the
[18:17] approach that we're taking is very different.
[18:19] In other words, we rely on large language models, specifically
[18:24] their reasoning capacity to act as an intermediary.
[18:29] Think of it as a translator between you and a very powerful
[18:32] analytical engine. So instead of you being required
[18:35] to also speak, you know, Python or go and so on, you have an
[18:40] intermediary that says, I understand what you would like
[18:43] to accomplish. I will now translate this for
[18:46] quote UN quote, an imaginary engineer who will in the
[18:48] background then construct the correct code, get the data and
[18:53] run the analysis and then give you the answer with with
[18:57] solutions like like ChatGPT or Claude and so on.
[19:02] You are actually asking the translator to also do the
[19:04] calculation, which it's fine some of the time, but it's not
[19:09] fine as the complexity grows because just because they have
[19:11] seen an option pricing formula many times as they translated,
[19:15] it doesn't necessarily make them a quant.
[19:18] So there is a big difference between how we apply and
[19:23] leverage LLMS versus how you would be using them in a
[19:28] platform like ChatGPT. Right.
[19:30] Absolutely. What is the strategic
[19:33] partnership with Microsoft Unlock for you beyond just
[19:36] computing power? It's been a really productive
[19:39] partnership in so far as they are always looking for solutions
[19:42] that they could offer their clients that would make them
[19:48] more, I would say cloud native and also more efficient.
[19:53] And so for us, we have had a very productive collaboration
[19:56] with with with Teams inside Microsoft, particularly on the
[20:01] Azure side in getting exposure to some of the top financial
[20:04] institutions to whom we were able to showcase the solution
[20:09] and contemplate in some cases. But I would say on Prem
[20:15] deployments where effectively you can actually enable the user
[20:19] reflexivity. Inside their own walls, if that
[20:23] makes sense. Which obviously sell solves the
[20:25] issue of data leakage, should solves the issue of security and
[20:29] so on. Absolutely.
[20:31] And ever since you started building the company, what's
[20:34] been your biggest learning about the go to market would be any
[20:37] niche and mistakes that you made that you later corrected give an
[20:41] insight into that. And there have been a lot of
[20:44] mistakes that we've corrected for sure.
[20:46] And I think we continue to make new ones that we can correct
[20:48] them later. I would say the, among the
[20:53] learnings, and there have been many 1, is that almost always
[20:57] the, the, the, the cell cycle takes longer than you think.
[21:01] In other words, there's always this.
[21:04] And it's particularly true with AI because I think that there is
[21:07] still a level of anxiety, uncertainty, and lack of
[21:11] confidence in in inability to apply it.
[21:15] You need to be willing to engage in what I would call a thought
[21:20] partnership. You're not just selling
[21:21] software. You're effectively offering your
[21:24] thoughts on where and how their business could be transformed by
[21:28] the application of AI more broadly and reflexivity
[21:32] specifically. And to the extent that you do
[21:35] and are willing to, to do that, I think you generally find a
[21:38] very receptive audience who is willing to try new technologies
[21:42] for the purpose of serving their clients better or being better
[21:46] investors themselves. But it does require a degree of,
[21:52] of, of patience and, and that kind of partnership.
[21:55] Absolutely. And these days we are also
[21:58] seeing autonomous agents also getting built, although I'd say
[22:01] we are still early in that and preferably still a human in the
[22:04] loop in most cases. But still there are some
[22:07] completely autonomous agents being built.
[22:10] Do you see Reflexity building an agent in future that's going to
[22:13] compete with the top asset managers for the kind of returns
[22:16] it can generate on its own? You know, I think that that's
[22:20] definitely the direction we're headed in.
[22:22] It's hard to exactly anticipate what the timeline is for
[22:24] something like this, but if you think about what makes someone a
[22:28] great investor, it's not yet clear that this is something
[22:32] that is a human only quality, right?
[22:35] In some sense, the idea is that you are able to process an
[22:38] extraordinary amount of information, which is what AI is
[22:41] currently helping you with, and then distill from it
[22:44] opportunities that offer unusually high reward relative
[22:49] to risks. To the extent that that's the
[22:53] principle behind this currently, really the piece that is
[22:56] probably missing from the technology that we see deployed
[23:00] around us is that we haven't yet seen the kind of degree of
[23:05] innovative thinking that humans are able to display from the AI
[23:11] solutions. And This is why our efforts, as
[23:13] I mentioned earlier, to help you not just with answering your
[23:17] questions, but increasingly to come up with better questions is
[23:22] so important because it does head in that direction where
[23:24] actually the analyst is no longer there just to be, you
[23:28] know, performing quantitative tasks with a high level of
[23:31] prescriptive information. It's actually actively being
[23:35] able to say, having gone through all of this, here's what I
[23:38] noticed. I think this would be worth
[23:40] looking into more. Yeah, absolutely.
[23:42] Quite interesting. And I completely agree with you
[23:45] that I would say that the ability of humans to effectively
[23:51] use AII think that's limited to an extent because I think humans
[23:55] still haven't completely realized what is possible with
[23:58] the power of AI and what it can possibly do.
[24:01] I think when that realization hits or that imagination
[24:05] happens, then I think we are going to see more breakthroughs
[24:08] happening because of AI and because of the human imagination
[24:11] that, you know, breaks that barrier that OK, now this thing
[24:14] which was impossible earlier, you can now just do that with
[24:17] the help of AI, right? And now talking about hiring,
[24:20] you know, building AI start up right now, it's it's very
[24:24] exciting. You know, everyone is talking
[24:25] about it. Space is very hard, but it's
[24:29] also incredibly competitive, right?
[24:31] I think you have competition from the big rigs, you have
[24:33] competition from other AI companies who might be competing
[24:36] with you for talent, etcetera. So how do you build that hiring
[24:41] process wherein you're able to attract the best people, wherein
[24:44] people are being hired at extremely high packages, right?
[24:48] The meta and all, they are throwing money to get the best
[24:52] people in. So what's your strategy for
[24:54] hiring now? Ultimately, obviously, we're not
[24:58] trying to go head to head with an organization like Meta in in
[25:02] pay packages. The approach that we have taken
[25:07] is you need to get whoever you're looking to bring on board
[25:12] excited about the problems that you're solving.
[25:14] And to the extent that those are unique and they aren't able to
[25:17] just go and solve them elsewhere and that they have a certain
[25:20] degree of interest or maybe even passion for financial markets or
[25:23] for some of the challenges that are that that investors
[25:28] encounter in these markets, that usually is a much more powerful
[25:32] driver than trying to compete just on the financial metrics.
[25:37] I will say that one nice thing about being in financial markets
[25:43] specifically is that they do capture quite a broad level of
[25:47] interest from people. I think that there are a lot of
[25:49] people who are generally have always been following markets.
[25:52] They're interested in it, they're invested in it and so
[25:54] on. So they do come in with a
[25:56] certain degree of knowledge already.
[25:58] And if you can tap into that and get them interested in what
[26:03] you're looking to contribute to that space, that's usually been
[26:07] the best way to hire people who can contribute the most and stay
[26:11] with you the longest. Absolutely.
[26:14] Now let's talk about fundraising.
[26:16] I see that you've been able to on board some incredible
[26:19] investors, right? So tell us what's been your
[26:21] fundraising strategy and what did you have a targeted approach
[26:25] to the kind of investors that you wanted on your cap table?
[26:29] Give us an insight into that. We've always been looking to
[26:34] have a mix of investors who could help us in various aspects
[26:37] of building the business. I don't think that there was any
[26:41] particularly, I don't think that there was any any secret sauce
[26:46] in how we ended up with such a stellar group of investors.
[26:49] I think it was much like the question you asked me earlier
[26:52] about hiring with investors to you want to bring them along to
[26:56] for a journey that they feel like is exciting and solving
[27:00] some problems that ultimately of course would result in the
[27:03] company growing fast and becoming a big and valuable
[27:05] enterprise. However, the goal really is to
[27:12] show to them the scope, the challenge and the value of the
[27:16] problem that you're trying to solve and the importance of it
[27:18] being solved for the audience that you're solving it for.
[27:21] And so we've continued to do that over time and as we build
[27:24] reflexivity into a product that is used by an increasingly wider
[27:28] audience, I think that a lot of those initial goals have really
[27:33] have really proven to be correct as far as the nature or the type
[27:38] of investor. I think we've ended up with a
[27:42] combination of, but I would say financial investors that can
[27:45] often help with some expertise in terms of operation and so on.
[27:49] And then strategic investors who have been instrumental in
[27:54] helping us with distribution partnerships potentially
[27:59] becoming clients and so on. So it's been, it's been a good
[28:02] mix. Absolutely.
[28:03] And talking about the capital allocation strategy, how much of
[28:07] the funding that you have raised goes to, let's say, hiring
[28:10] versus product development or something else?
[28:12] So give us an insight into that. I would say the the two major
[28:18] expenses that we have are really data and of course the top
[28:24] talents. I would say that those are, you
[28:26] know, I don't without without specifically going to the
[28:29] ratios, those are by far the two biggest expenditures that we
[28:33] make. So in order to deliver a product
[28:37] like we want to to our investors, we always want to
[28:39] make sure that we have not only the best capabilities, but also
[28:42] the highest quality ingredients, which is primarily data.
[28:46] And we really do believe in the garbage in garbage out, gold in
[28:52] gold out is true also of AI. And so we, we, we take great
[28:57] pride in bringing together the best data with the best
[29:00] technology to offer a product like reflexivity.
[29:05] Absolutely. And giving your experience as a
[29:08] previous CIO as well and now building a startup, what do you
[29:11] think is harder? Is it predicting the direction
[29:15] of a market or running a tech startup?
[29:19] You know, I think building a company is definitely harder,
[29:23] partly because when you're an investor, you get to pick the
[29:28] problems you're solving. In other words, you look at
[29:30] things that are interesting and things that you think you'll be
[29:32] able to analyze. You analyze it, and then you
[29:34] have an investment strategy and an investment opportunity.
[29:38] When you're building a business, you have to solve the problems
[29:40] that come to you. And I think as the CEO, you're
[29:43] almost always the filter for all of the worst stuff, right?
[29:46] Things that come to me are things that usually require
[29:50] immediate attention and are very often bad news.
[29:53] So you can't say, OK, why I would like to just not deal with
[29:58] this, you have to deal with it. So from a challenge point of
[30:01] view, it is definitely an order of magnitude more difficult than
[30:07] than than the investing side. Of course, intellectually
[30:10] speaking, I think investing is incredibly, incredibly exciting
[30:14] and extremely rewarding. It just has that additional
[30:18] component that you have some degree of freedom as to where
[30:21] and what challenges you solve, unlike with the business that
[30:27] doesn't give you that option. Yeah, absolutely.
[30:30] And let's say you had to put your investor hat on for a bit,
[30:34] right? Looking at the AI hype right now
[30:38] and how the market is behaving, what what do you make of it?
[30:42] You think it is over hyped or over valued in general?
[30:45] Or do you think there are some reasonable pockets still there?
[30:49] Or do you think even the overvalued companies are still
[30:51] going to give a massive return in the medium to long term?
[30:54] What's your take on that? I think we have ample evidence
[30:59] from prior cycles that's no matter how overvalued a stock
[31:02] can always go higher. And so the issue currently is
[31:06] that the excitement around AI is real.
[31:10] And although valuations themselves relative to history,
[31:15] you could argue are very elevated, it really depends on
[31:18] which companies you're talking about.
[31:20] South, for example, you know, you take company like NVIDIA and
[31:24] so on and you realize that they really are actually making a
[31:29] tremendous amount of money as part of this.
[31:31] This is not like some of the previous cycles or some of the
[31:34] previous analogs people try to draw in terms of bubbles where
[31:38] you had a lot of hype but no revenue.
[31:40] There is real revenue being generated by many of these
[31:42] companies. There are real solutions being
[31:45] built. And if I think about what stage
[31:48] we really are, forget whether or not there could be, there always
[31:52] are corrections in the market, right.
[31:53] People get excited, then you have a little bit of an air
[31:56] pocket and there's a correction and then a rethink and so on.
[32:00] The larger point or the take away for me is this which is at
[32:05] least in the space of investment management, we are really only
[32:11] at a pilot stages in in so many cases, right.
[32:14] Institutions are really just dipping their toes into this.
[32:17] So we have not seen anything yet in terms of true impact, true
[32:22] required compute, true power that is going to come to bear on
[32:27] the problems that we're facing. So I have a high degree of
[32:33] confidence that we are in the very early innings of this.
[32:36] Again, that does not preclude some very violent market action
[32:40] and rethink. However, when it comes to the
[32:43] trend of AI, this is an exciting time to be and it really does
[32:48] feel like the ground floor. Absolutely.
[32:51] And not making any investment advice your own analysis, right.
[32:55] And to move to a last question before we try to wrap this up,
[32:59] What's been your biggest learning building Reflex City
[33:02] till now? I think the biggest learning is
[33:06] that ultimately the most important asset that you have
[33:08] and what makes a business successful is not, you know, a
[33:12] flashy vision, not sort of the mission that you talk about.
[33:17] It really is the team and the execution of that team.
[33:21] So if you gather really great people and that team executes
[33:27] well on the goals, that is by far the most valuable capacity
[33:32] that you can have as a company and the most valuable thing that
[33:35] you can aim for as the CEO. Absolutely.
[33:38] That's a beautiful learning there.
[33:41] Now to wrap this up, do tell our audience where they can learn
[33:44] more about what you're building at reflexivity and where they
[33:47] can follow a journey. I think the best place to start
[33:49] is really just to go to reflexivity.com.
[33:52] We have a lot of material there that can help people understand
[33:55] better what is it that we're aiming to do, what our audience
[33:59] looks like, and whether or not this is something that would be
[34:01] of interest to them. And if it is, we'd love for them
[34:05] to contact us and we can show them the solution, talk to them
[34:08] about some of the use cases and are happy to engage in a little
[34:11] bit of a back and forth on the direction of AI in investment
[34:14] management. I'll be sure to put the link to
[34:17] your website and your LinkedIn in the show notes.
[34:20] Thanks so much for making dive for this Jan.
[34:22] It's been amazing talking to you.
[34:24] I wish you all the best. Thank you very much.
[34:26] It's great to be here.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.