See the show notes for this episode: S 01 | Ep 49 Balancing Act: Immigrant Experiences & Entrepreneurship

0:00:00 - Alex Shevelenko

This is going to be probably one of the first for me, for sure. And one of the first in the podcast business. And certainly one of the most exciting episodes on a personal level. Who is a Chief Business Officer at one of the most exciting AI companies in the market right now called Perplexity? He also happens to be a “chief brother officer” for me. Dmitry, welcome to the pod! Very excited to have you on!

0:00:41 - Dmitry Shevelenko

Great to be here, and it's fun to do our catch-up in front of the entire world.

0:00:47 - Alex Shevelenko

Yeah, I think this is a crazy idea. That happened in a walk and we don't often talk business. But I think it would be fascinating to hear your story again and see how you connect the dots. And I think it was very interesting and relevant that you've worked at some of the phenomenal hyper-growth companies in the Valley, starting from Facebook. Growing from 80 million users to 800 to go on to be in the news reading business which was a great company called Pulse. It got acquired by LinkedIn and then you continue to hyper-growth there. And then Uber is another little success story when it comes to growth and all sorts of innovation.

And now, as a Chief Business Officer of Perplexity, you're effectively in one of the most talked about companies. In half the podcasts that I listen to, they're arguing of what's the future of AI and how you guys are taking out Google or vice versa. So it's very topical for our audience to get your perspective and lessons learned at Perplexity. But also earlier from these success stories. I think it's always fun to do something different, such as have your brother on. So let's get started. What should the world know about Dmitry's story today? I've highlighted some of the companies you worked for, but what's, in your view, the common thread that you've learned that you're applying now in the AI world?

0:02:40 - Dmitry Shevelenko

I think a simple heuristic is to work on products that you like to use yourself. I joined Perplexity because I became pretty addicted to using the product. There was a similar feeling when I was at Facebook. I was a user before an employee, similar to Pulse, similar to Uber. At times when I worked on things where I didn't love using them myself, that usually was not a great signal for the product. If you love using it, you can assume other people do too. So that's been a North Star for me and also it's always more fun to work on things you're passionate about.

I think I love the urban problems that Uber was trying to solve. I thought the time when  Facebook came out, I was an anthropology undergrad. It was just playing this interesting transformational role and social dynamics when I was an undergrad at college. And so that mapped to something that felt like an interesting problem. And then certainly, with Perplexity people get answers. The economy of knowledge and information to me is potentially the most interesting problem that I've worked on. That's been another theme.

0:04:28 - Alex Shevelenko

We can always come back to the previous stories. Why do you find the knowledge and the answers that Perplexity provides so interesting? I think around the time when you were in Columbia, there was this product called Ask Jeeves that was trying to do something where it was using natural search to query to try to decipher what's inside the internet. It didn't succeed even though it was acquired. I think now you're challenging the established leaders like Google. Combining what people love about ChatGPT and Wikipedia. So tell us a little bit about what's Perplexity and what's an interesting innovation that you're bringing to the market.

0:05:32 - Dmitry Shevelenko

I think there are no new ideas. There are just new technologies that make old ideas more useful or less useful. So getting natural language answers would have always been a better experience than getting links. It's just that the technology was not sufficiently sophisticated 20 years ago for that to be a good user experience. So I think a lot of startup success or failure boils down to timing. Perplexity launched right when GPT-3 became available as an API. The ability to succinctly summarize web content just did not exist before that. So if you had tried to launch something with Perplexity six months before, it would have been too early. If you had tried to launch Perplexity six months later, potentially too late. Because we and others were already going down, creating data flywheels and refining the user experience. It's not that there's an original idea.

There's an original application of technologies that just became sufficiently advanced enough to be able to build a new user experience and create a new brand that aligns with that new user expectation. You go to other products, and even if they're trying to show you answers, the user expects to get links. Whereas the user who comes to Perplexity knows what they're going to get. They're going to get an answer. And that's a big advantage for us. Because it's really hard to change user expectations, even if you're qualitatively giving them a better experience. If that's not the experience they wanted out of a certain product or were expecting, it may not resonate in the same way that a new brand delivering that type of experience can do. 

0:08:06 - Alex Shevelenko

So what you're describing is, effectively, you've designed a new category. You have particular people who already have adapted. You have this credibility and we'll come back to that in a second. You can have a huge congratulations on that. But I think what you're also saying is it's defensible. Let's say, Alphabet or Google already dominates another category and has all sorts of built-in incumbent advantages there. But with those advantages come the disadvantage that people expect certain things from you and you may not easily move away from those things. What you're saying is that creates an opportunity to add a new experience, a new job to be done, so to speak, for those users where you meet their expectations and they know, “I go to Perplexity for this and I go to Google for something else.” That may be totally fine. Is that what I'm hearing, Dmitry?

0:09:05 - Dmitry Shevelenko

I think there's also a more classic innovator's dilemma around. I mean, there are a million little details in how you render answers instead of links. There are all kinds of trade-offs you can make. When you don't have an existing business to defend, you make trade-offs all in favor of the user versus other constituencies. That's our ability to focus on what's best for the user in the immediate term. Plus just moving, faster shipping every day. It's the only advantage startups have ever had. At Perplexity, certainly, we operate with urgency on execution. We always talk about 1% better every day. It means you have to ship every day, have that urgency, and not have complacency around having a good product. It has to get better in a continual fashion.

0:10:22 - Alex Shevelenko

That's great! I think that connects the dots to some of your previous companies where you worked, like Uber. I think it was the winning part, an incredible experience. At least for me as a consumer, not as a driver, where for the first time you got certainty and clarity of what's going to be happening, your expectations were going to be met. You have the feeling of control both over your time and what's going to happen. It sounds like some of the same paradigms and maybe that's why you love Perplexity. Maybe describe at a more technical level what happens if I type in a query into Perplexity versus what happens if I type the same query into another search engine. Why you're innovating in those areas, where you see the difference for the consumer?

0:11:20 - Dmitry Shevelenko

I think the most important work we're doing for our users is saving them time. When you type a query in Perplexity, we break down that query into many queries. We then find, using our own search engine and web index, trustworthy and reputable sources that have information that's responsive to the query. We extract the useful parts of that web content and then we run it through many technologies. We use retrieval and augmented generation, and we use LLMs to do the summarization step, generation, and then render the answer in natural language. You have to do all that in under two seconds.

But fundamentally there are three product priorities that are evergreen: accuracy, speed, and readability. Those are always in tension with each other. You can optimize for speed. It may come at the expense of accuracy and readability. You can optimize for accuracy. It comes at the expense of speed. You can optimize for readability. It may come at the expense of both. Like, if you have, you know, a lot of pre-calculated templates, it may look better but may have less real-time information.

So there's a constant set of trade-offs that you're making and that's the hard part right is tuning it the right way. That kind of users feel is natural and ultimately, again, saves them time. And you know, saving you time by giving you a bad answer ends up taking more time. Right you, a bad answer ends up taking more time. Saving your time by giving you an answer that's hard to parse, even if it's correct, is also not great. And if your service is so slow that you get frustrated and don't wait for the answer to load and click off somewhere else, that also doesn't end up working well for the user. So, yeah, it's all about managing those trade-offs and continually, you know, making progress on all three fronts without kind of degrading any other factor.

0:14:01 - Alex Shevelenko

So, one of the things that you picked up from your Columbia sociology degree and like a lot of or sorry, yeah, that's right Anthropology, sorry Shows that I'm not very good on brother memory test, but one of the things that you've been fascinated with is is the psychology, right, and I think, um, you know we shared a kindle account at some point and I could have sent some fascinating books that you've been reading around that. So I'm curious, you know, how do you, when you say speed, right, and like, let's say, latency in the search results is an important component, so there's a speed of like your engines and AI, you know, whatever the chips that you're using and all that stuff, and let's just park that for a second but there's often sometimes a very important technology around perceived latency, right, like, how do you create like an experience for a customer where you may be loading stuff in the background but they are not sitting there, you know, frustrated, waiting for the answer to load? And I just kind of recently had an experience with another platform, actually one of our competitors, Adobe. They implemented like some sort of AI algorithm that allows you to test, and check what's in the PDF, and it's taking a minute or two minutes for them to send the query, et cetera.

And then I go on Perplexity and it's instant. Something is showing up, I have something to read, so I'm sure you guys thought about that. You know, like, guide us a little bit on how do you think about creating this delightful experience while you're waiting for the data to come in?

0:15:56 - Dmitry Shevelenko

And what have you innovated around there? I mean, I think the core of the strategy is to make it feel like you're not waiting, and so it is, you know, and giving you useful context. I mean, in the Perplexity UI, we load sources above an answer, and so that's important, you know. That's kind of you know, it's a good example of, like you're accomplishing two objectives. First, you know what the sources are going to be before you have the summarization, and it's important to load sources because that's the context in which a user evaluates whether they're going to trust the answer. Right, like I mean, we have a very journalistic frame that you know your answer is only as good as your sources, and so if your sources are not reputable in general we try to avoid having not reputable sources, to begin with, but if that's, if your sources aren't reputable, your answer won't be, but if your source is good, so, like a McKinsey consultant doing their right.

0:17:04 - Alex Shevelenko

Before their presentation they kind of set out the framework, the methodology you know, and some of the sources to build the credibility for what's about to come, before you pay, you know, tens of millions of dollars for the strategy outcome. And so you, effectively, you're doing the same thing you could just buy in this particular interaction.

0:17:24 - Dmitry Shevelenko

So, A, you're building credibility and B is you're giving the user something to look at before the answer is ready, right, so, and so that's kind of where you kind of we stream tokens and it gives you time to kind of again, you're immediately getting something before you're getting the answer, and that's.

0:17:47 - Alex Shevelenko

So B is effectively a mirror in front of the elevator while you're waiting for the elevator. You know, instead of buying a super expensive elevator, you're kind of looking at yourself. When you're having a positive experience.

0:18:00 - Dmitry Shevelenko

I mean, I would say there's a utility right, which is you need to know what the sources are anyways, and it's just, you know if we get that information faster, we load that immediately, and so I think there's just a sequencing there. I mean, human eyes can only, like read, you know, parse information at a certain clock speed. And so you can effectively get the perceived latency to be zero if you're kind of loading information as it comes in versus waiting all at once to kind of render the complete answer with the sources.

0:18:42 - Alex Shevelenko

Got it, that's perfect. Well, let's come like let's build on the psychology of the consumer of this content. So I love the sources. It's fundamentally like maybe you know where we we agree that you kind of providing the evidence behind the answer is critical and you guys have these little kind of pointers to the sources. It was in the answer. You know, how are you finding, um, how are you finding the consumers react to that? Is this like a high likelihood that people are going to go click through on the sources and investigate, or is this a kind of a rare occasion and just something that you need to have there to build trust and give them the sense that they could go and search out, or it's not as essential you know, it's effectively not as essential for the modern consumer to go and drill into the sources further.

0:19:37 - Dmitry Shevelenko

Yeah, I mean, it depends on the type of query a user is doing and you know, different query types have different rates of uh click through on the sources and it's not like some if you had to aggregate, I would agree.

0:19:53 - Alex Shevelenko

That makes perfect sense, by the way, and I like it if you have.

0:19:56 - Dmitry Shevelenko

But if you had to aggregate right on the whole, right like we're not going to share any stats but, like it is, the value to the user is sometimes not even direct, like you're not necessarily going to click through right at that moment, but then you'll know, hey, if I need more information about this topic, that was a good source for it, right, and then the user may go separately and just directly go to that source later.

So I would say that there's a lot of utility to knowing what the sources are in addition to just the direct click-through. But yeah, it's not. If you're doing deeper research, the likelihood that you go to a source is much greater. If you're looking for a very quick, you know, informational answer, you know in those situations probably less likely to go click through to the source. But yeah, we're still seeing a lot of emerging behaviors there and what we see as interesting is also a lot of questions have a follow-up question so that we've over 40% of queries have some sort of follow-up and so that's kind of creates more surface area to kind of understand which sources are responsive to the follow-up question and create more opportunities for click-through there got it well, um, you know if we can continue with this theme, right, like, what else are you finding?

0:21:34 - Alex Shevelenko

From a psychology of of the consumer, that's been interesting, that you're pushing towards, because you're moving the world from a link economy to an answer economy, right? I think the other phrase that you mentioned the other time we were chatting that connected with me at least, is that you're kind of like you would be searching the way billionaires do search, right, because you're optimizing the premium on time. We talked a little bit about that, but you're on the sort of bleeding edge of how we interact with knowledge and there's a lot of value, obviously, in that. So what else is interesting that's not confidential that you're uncovering about the modern knowledge consumer?

0:22:25 - Dmitry Shevelenko

Well, I think, going back to the follow-up questions, a good answer leads to more questions. What was interesting about the engagement point of view is people want to go down rabbit holes of learning. So oftentimes, you may end up spending more time on Perplexity than you would on Google. But not because you're having to wade through a lot of junk to get useful information. It's because you get a much deeper knowledge. You're inspired to keep probing and getting closer and closer to the ground truth that you're looking for. So, yeah, I think there's like we try to avoid calling ourselves a chatbot and the role of the conversational piece and the follow-up piece and the natural language piece. It's less about a chat and it's more about creating a space where people can get to the thing that they're looking for. One of our pro search features for paid users asks you clarifying questions. If you ask an overly general question on Perplexity, we in real-time kind of identify what potential clarifying information to help you get the answer, and so, yeah, I think it's kind of this multi-step process of getting more and more knowledge through the back and forth of the Q&A is something that's creating a lot of value for users.

But I would say it's not about psychology, I mean people. It's the, the, the need is like people need information and knowledge to like to live, you know, be productive at work, live happy lives. Uh, we're not like there's no, there's no like manipulation or anything like that. It's like just you know, making something delightful and easy and you know, uh, efficient, you know people will do it more. Uh, so we're not, it's not like there's.

0:24:49 - Alex Shevelenko

We don't do a lot of A-B testing like exact wording and things like that.

0:24:55 - Dmitry Shevelenko

It's really like the core experience delivers on a very essential need. So we're not, there's nothing. There's nothing like artificial happening there.

0:25:07 - Alex Shevelenko

That's a great point and it wasn't meant like that. I actually kind of meant it as more as a compliment, because I think you're trying fundamentally just to think like what would be, as you said, value-adding experience for the user, right, Because you're coming in with a category where there's already a lot of established ways of doing things and you're not beholden to some of them and to me, like what you just described, for example, the ability to ask a follow-up question, right, I've been known to ask, you know, very unstructured questions now and again and I think the guest may follow up and say, hey, break it down, or kind of just focus on. Hey, I'm just going to answer that part of the question here and then I'm going to answer another part separately. So that's a kind of just clarify what the question is all about, and that helps. That's a natural conversation.

So it sounds like you're just basically getting in. You know to be empathetic to the user and, you know, helping them clarify what they're looking for and then the process of getting them to a better outcome. Yep, that's what we try to do every day. Or search query interface, wherein Wikipedia, we've all been known to sometimes go deep into the funnels of Wikipedia, and what's guiding us is not manipulation by Wikipedia right, they're not like really monetizing it. What's guiding us is our intellectual curiosity about a particular problem and clarity and knowing, hey, if I click on that link, I'm not going to go off into some other third space, I'm going to stay inside this. You know this environment of Wikipedia and I'm going to keep kind of querying and querying that further. It sounds like that's kind of an important foundational story behind applying some of that within Perplexity. Do you want to comment on that more?

0:27:29 - Dmitry Shevelenko

I mean, I think the core human emotion that we tap into is curiosity. If you can make it easier for people to activate their natural curiosity, they enjoy it and it's fulfilling for them. It's useful for them. So, yeah, I think it's. You know, that's why we show the related questions and you know, yeah, that's a lot of. The design of the product has a singular focus on getting your questions answered so that you can ask more questions, and that just pays off in many small ways, fantastic.

0:28:18 - Alex Shevelenko

Well, listen, so we've talked a little bit about the product, touched a bit on the technology underneath it, but let's dive into the business and in, in particular, some of the areas where I mean you guys are doing phenomenally well. Right, recently, you raised, uh, over massively oversubscribed round. Um, have some. You know, the names of investors are pretty remarkable. I'll just kind of rally them off so as not to make you embarrassed of who they are. But we have folks like NVIDIA. We have Garry Tan, Jeff Bezos, Elad Gil, Dylan Field from Figma, and Naval Ravikant. I mean these are like who is who of general investors, AI investors, and have great VCs as well involved. So tell us a little bit about the fundraising story and what's behind that and then, like, we'll shift from that into the traction that you're seeing in the market and you know why those investors are interested. Or maybe we could address those in parallel.

0:29:47 - Dmitry Shevelenko

The thing I can talk about that's interesting is just, you know it's a product that goes back to the first thing I said is like it's a very transparent product, like people use it. If they have a good experience with it, they can draw their conclusions and extrapolate. Do you know what is exciting and interesting about Perplexity? And it's not about selling or storytelling. You know about where we're going. It's like you can evaluate it on what it is today and how it's evolved. You know, in the company's you know relatively short history, right?

You know most investors already have a point of view about Perplexity. Yeah, I think like I imagine most good investors would try to use a product before they talk to the company. In Perplexity's case, it's pretty straightforward to do so and that's kind of you know, the conversation is largely centered on the product because it's a pretty universal need and you know, if people use the product themselves, then they would expect others will. If they don't use it, you know, then they wouldn't expect others to.

0:31:08 - Alex Shevelenko

Well, and on that note, let's just kind of do some other people in the media that are a little bit more respectable than the Experience-focused Leaders podcast, Wall Street Journal, I'm just screen sharing here. One of their reviews just rated you as a, and I know you don't love the positioning as a chatbot, but you rank number one in the Wall Street Journal review of leading AI chatbots, Perplexity, followed by ChatGPT pulled by Gemini from Google, Claude, and Copilot. Tell us a little bit about what's happening here. Why do you think consistently, you guys are getting that reaction, whether it's from investors or the press, which tends to be skeptical, and how do you continue to do this? Right, you have some pretty established names and competitors. You know gunning after you being positioned here in such a leading bull position. What do you know? Is it that the other folks are just coming in and they are behind the curve? Is there something special that your team is doing?

0:32:29 - Dmitry Shevelenko

Everything boils down to focus and pace of execution. There are very few, if any, moats in life or business. But working with more, you know, focus on what was best for the user and shipping improvements faster is you know, if there's some recipe to kind of you know, and again, that never stops right If we get complacent, like all this goes away. Focus and pace of execution, intensity of execution, that's where we see our competitive edge and hope to maintain that for many decades to come.

0:33:44 - Alex Shevelenko

Well, that connects me to one of the previous companies where you've been driving business activities, Uber, which was also very well known for that pace of execution and being aggressive in securing business opportunities, and business development opportunities in specific markets. You guys have done a really interesting strategy in addition to being consumer-facing and you know publicly kind of engaging with telcos and you know phone carriers and you know placing Perplexity in more and more places. Tell us a little bit about what's behind that. How are you able to accomplish that, given relatively you know short life as a company in your kind of lifespan to get some of these relatively slow-moving historical organizations to be so excited?

0:34:33 - Dmitry Shevelenko

I think it's um a lot of the same traits that apply to the product, kind of focus and persistence. And, I would say, with you know, a deep understanding of what is important to your partners and how their organizations make decisions. Organizations make decisions and in some ways, the superpower we have that is available to everyone is using Perplexity to research the companies and people you want to work with. And so getting that deep understanding allows you to structure potential collaborations, and engagements, in a way that aligns incentives over the long term. You know we're one of my pet lines in having partnerships is, like you know, always being long-term greedy, not short-term greedy. Particularly when you're a smaller company trying to work with bigger companies. Identifying both what's the immediate win for them and then what's the thing that matters to you over the long term and what's the thing that matters to them over the long term.

And yeah, we've been fortunate to have many great partners. But I would say this only works when you have a product that people like to use, in the same way that before we talked to an investor. They've used our product. Before we talked to a partner, they've also, for the most part, used our product as an investor. They've used our product. Before we talk to a partner. They've also, for the most part, used our product.

And you know it's like, to the extent to which I'm, you know, pitching it's, I just do a live product demo, you know, and like it always has to be live right, like nobody for a product like ours. As you know, screenshots and videos don't cut it, it's. You have to ask people, like not pre-rehearsed questions, and see what the product can do, and that's where people see the value. And so just kind of having that you know transparency and you know willingness to kind of almost be vulnerable by this is what it is, and we don't sell to win partnerships. We try to align incentives and show our potential partners how their users benefit from Perplexity and how it then ultimately benefits them.

0:37:26 - Alex Shevelenko

If we can again take a step back now, one of the incentives for your competitors is to maintain their current business right, like in their revenue streams, and one of them is they're worried like, hey, what's going on? This young startup building out a great product and competing with us, and so they're releasing new products. It's in the news. First of all, how does it feel to be kind of in the news cycle for yourself personally, where you have to be, you know, competing with these larger organizations that have you know large PR teams, large news teams, and then you know how are you thinking and adapting your strategy in real-time, given this competitive environment that you live in?

Right, because I think for some of, like some of us, maybe our markets are a little bit sleepier and some of the competitors and incumbents that we're dealing with are not as powerful or not as well known for execution. You're in a very hot and popularly spoken space, and so the success breeds, you know, attention. So I'm curious, kind of both on a personal level how are you dealing with the constant, constantly being in the news, and then you know how, how is it affecting your strategy, or are you kind of sticking through to your original vision and just kind of, like you know, going, going after it and just kind of letting that blow through them, through the, through the back of your hair, as you're moving rapidly towards a shiny beacon on the hill?

0:39:10 - Dmitry Shevelenko

Yeah. So I mean, the core driver of Perplexity's growth has been people, completely organically, without us asking them to or even knowing that they were going to post on social media that they love using Perplexity, that I would say like my frame on earned media is like, uh, that's an extension of that. I mean, it's people kind of organically talking about Perplexity, um, both what they like and sometimes what they don't like, and so we think that's great. I think it's not. I've lived through cycles of other companies where, like you have to assume that, like when you're a darling, you know like there's natural, just cycles of like getting built up to get torn down and like you don't quite know you know what, uh, um, you know where you are in the cycle, uh, so you can't take anything for granted and it's, um, you know the thing you can control is like always making your product better and ensuring that when somebody tries it they have a good experience and so consistent product quality, uptime. You know that's like that's the actual way to have a good earned media strategy.

Is like ensuring, when people want to try your product out, that it works for them in without just like a, you know, in the demo environment where everything is ideal, right, and I think, the yeah if there's yeah.

So I think that that's how we approach it. It's like, you know, encouraging people to try the product and then come up with their own opinions, and then you know whether they share their experience or not.

We're doing fine with people having more awareness of Perplexity, like, I think, our early users like, probably like the fact that they're early to Perplexity and now it's becoming more mainstream, and so, you know, they use it more because they feel validated that they were kind of, you know, early to a growing trend which is a phenomenon that repeats itself, and they become even more avid supporters of the brand. But we don't do paid digital marketing, we don't do. You know, we have a very authentic brand and we're going to do everything we can to preserve that. Cause I think a lot of overly salesy approaches end up diminishing user trust, um, and that's something we want to be very sensitive to avoid.

0:42:01 - Alex Shevelenko

You have a relatively small team. I think there are some cases, you know, I'm sure it's growing but, like in the larger scheme of things, for the number of years in the business and the kind of the traction that you're getting, you're operating with a small team. Is there anything that you've learned from that experience versus being part of larger teams like at Uber and you know Facebook, that you know is enabled by AI itself, enabled by your product perhaps, and, you know, do you expect to change that? Do you expect to kind of scale up and build a much larger team to take on this big opportunity? Or do you think this lean unicorn approach or like a lean successful company of the future?

Sometimes people say you know one person’s unicorn is sort of the ultimate goal for some companies. How are you guys thinking about that in an age where I think you know there's, particularly in the AI world, there's a lot of certainly like it's expensive from the hardware perspective, but team wise. It's also expensive. There's a limited number of experts in certain fields. How do you kind of process that as a team, and how do you see your overall employee growth changing over time?

0:43:29 - Dmitry Shevelenko

We want to stay as small as we can to realize the completeness of the vision and the products we want to build. Team size is inversely correlated with execution speed and, unfortunately, I wish it were as simple as you add more people and then you ship more products faster because then you know that would be a great formula, but that's not reality. We want to make sure every team member we add is somebody that increases our clock speed, and so that's kind of filtering for people that generally have experience, people that you know come in with high trust and have a bias towards execution trusts and, you know, have a bias towards execution. We've hired very few people, like who aren't experienced because we want people to come in and be able to have an immediate impact. You know part of having a small team is that you need to have everybody be an absolute all-star.

We're big believers in the A players. Hire A players and the talent bar is absurdly high. Because if you're intentionally building a small team, there's no room for lag. And it ends up being demotivating if everybody is hustling hard and then you have some folks that aren't. So, um, yeah, I'm really proud of, uh, the team we have and the team we're building, but, uh, it is absolutely true that, like um, the way you build the modern version of Google is not by being as big as Google or Meta. I think you know whether it's us, hopefully, or you know others. I think we will be surprised at how, how, yeah, how kind of small the kind of next big companies will be from an employee count point of view, because they preserve their clock speed.

0:45:50 - Alex Shevelenko

Great Well, Dmitry, this has been a very serious, very businessy, very kind of formal discussion. Now let's kind of shift gears a little bit and let me ask you a few personal questions that I may have some hypothesis to, but I think that also would be very interesting for our audience. So one of the questions that I ask our employees, or team members that are about to join, is what did you learn from your mother? Sometimes I ask, would you learn from your father?

But generally, I ask one of those questions, and the idea here is just to kind of get a little bit deeper, like what's the story, where the lessons come from, where it's the background of kind of a founding team member or kind of a new key team member. So I'd be curious, right, you have, uh, this tremendous, uh you know track record and very, you know, often, B2C companies, right, and the Perplexity, effectively, for, yeah, at least right now, it's predominantly B2C focused uh, um, I've been focusing a lot more on many B2B companies, uh, in technology space. So we come from the same genetic pool to some degree. So I'm curious, what did you take away, and you know how is that changing over time, right, like how kind of the lessons that maybe you took away where you're a Columbia undergrad is different than when you were kind of an employee and you know, a founder and now kind of a leader in an in the, in a hyper-growth mode.

0:47:35 - Dmitry Shevelenko

Yeah, I think, yeah, we've, we've been very lucky with our, our role models. So, yeah, I think, you know, from with our mom, I think one thing is like, don't be afraid of hard work, that's one thing, that I probably feel even more connected to now than I did earlier, and just thinking back to, like you know, her, her hustle, her drive, her like late nights and, but doing so in a way that, like, at least I, I don't feel there was ever any trade-off in terms of her ability to kind of be there for our family. So I think, um, this idea that there's some sort of trade-off, particularly for women, um, you know, like you can't both be a rock star kind of business leader and a rock star mom. I think she disproved that. And I'm a proud father of three daughters. So I think it's it's not just for me, but like I, I love, you know, her example for both my, my wife, was a professional and for our kids, um, so so that that that one sticks out. I think she another one is just I feel like she really sweats the details, um, and so I think kind of you know that's another thing from her. 

The third would be like a reality check, like you know is the thing you're doing, does it actually make sense, right? And I feel, um, you know, kind of this don't get caught up in your own story. And like touch grass, I think that's her. Through her career, and you know how she parented, some lessons came through to get it interesting.

0:49:49 - Alex Shevelenko

It's funny, like from an outside perspective. Hearing you say this, don't disagree with it. Not sure would have picked the same, the same lessons, right so you have one of your three.

0:50:01 - Dmitry Shevelenko

What are your three?

0:50:02 - Alex Shevelenko

I was just thinking, I mean I would take them but the one that sticks for me is just the ability to connect with people, I think, as you know, on a very human level. Sometimes I think some people may think that I'm extroverted, but I'm like, compared to her the ability to come up and just jumpstart conversation was a stranger in a different country and you know, whatever, like I'm an income poop still and I think I've grown and had the privilege of absorbing that. So that's kind of one and I kind of on a relatively early level, but then you know, deeply connecting with people professionally, personally, over a long period of time. And I think you know one of the stories that kind of sticks in my mind is some of the alums of other companies that she worked with, or kind of teams that she's led still gathering to uh together many years after that company no longer exists. That means that as a leader you're kind of able to create a following and then that's not just a following about you, it's not that sort of like the person. That's all about me, the ego-driven thing it's. It's actually you built a community that other people around that, like the mission, like the vibe that's been created. I think that's remarkable.

I think this is kind of the second thing. Is that what you're describing kind of doing as an immigrant, a refugee to the United States? I think to me that's very fascinating. You know we had slightly different ages when we entered the United States. But Silicon Valley is well known for being a place where people that are not born in the country are kind of doing, you know, either kids of immigrants or immigrants themselves are doing really well. So what's your take on kind of that journey and how does that influence you to know the way you've tackled different roles? From being on the lending team for Facebook. You know other kinds of expanding Uber everywhere around the world. Like, guide me a little bit on that kind of immigrant journey and what impact it had for you being part of an immigrant family in the United States.

0:52:48 - Dmitry Shevelenko

Yeah, I mean I think there are several things together like a healthy chip on your shoulder. That then is like motivating to like swing bigger. Yeah, I think sometimes you know a desire to associate with well-known brands, so like kind of going between riding off of the brands that establish you and then trying to create your own. You know brands and identities, and I think kind of that. There's probably some of the immigrant mindset of wanting to fit in that, like you know, makes you want to be part of better-known things. And then there's like the risk-taking side of it that then, like you kind of venture out.

So I feel like in my career I've kind of jumped between those modes. I mean I arrived in the country at three, I think it's more, it's less about the experience of an immigrant and more about the experience of having immigrant parents. That’s the defining. I mean I wouldn't consider myself like I have enough scars to really like wear the immigrant badge. You definitely know some formative experiences of seeing your parents kind of go through immigration and adjust to a new country.

0:54:20 - Alex Shevelenko

So dive into the last kind of thing that you brought up is like okay, so yeah, you're successful, you're part of these great success stories. And then you want to go and you know, do the difficult thing right. Start a hardware robotics company. You know, and kind of guide us a little bit through what were the lessons and you know the resilience muscle that you build. In that kind of going through the process, because you know from an outsider I really admire kind of the process and the integrity was which you approached kind of the journey and the kind of transparency was your team. However, I'm really curious. You know we've focused on kind of the super successes, but one of the great successes is to have tried something that may have not worked and then bounce back and go on to do new great things where lessons are learned. So walk us through a little bit of kind of what your takeaways from Tortoise.

0:55:28 - Dmitry Shevelenko

So Tortoise was a hardware, robotics company focused on delivery robots and mobile vending robots. We started actually in robotic automation for shared micro-mobility. I think the number one lesson for a potential hardware founder is don't do hardware. It is incredibly difficult and capital intensive but we think a lot of it boils down to like finding good people that you want to work with. I got to know my co-founder well and we had a really good complementary skillset and we had a unique kind of perspective on the market. We originally actually did not want to be in hardware long term. It was kind of an initial ingredient towards becoming a pure software company.

But then COVID happened and we shifted strategies and I think there are two lessons of Tortoise. One is like we tried everything, we pivoted two times and we left it all in the playing field, but we also knew when to not chase. You know like good money after bad, not kind of like we could have kept it going, raised at super dilutive terms, and kept the train running. Let go of everybody, if the founders themselves don't feel it in their hearts. It's not the best use of anyone's resources.

And so I think when the zero interest rate phenomenon expired, and so did our prospects, I think we made a judicious choice and tried to do everything we could to secure, the best possible exit. However, it was a difficult environment. But for me personally, the experience of being a founder was invaluable and I think that mindset and ownership bias is part of why it was really easy to kind of slot into the Perplexity culture, and it's kind of whatever road leads you to the situation you're in, if you're happy with it, then you know that became the central path. So no regrets.

0:58:07 - Alex Shevelenko

Yeah, and I think that's exactly what it feels like. I think you have kept the hustle, maybe even you like, you set the chip on the shoulder of like hey, but now you're bringing it an environment where things are clicking now, last kind of wrapping up the family section right, so we started was what you've learned from your mom. You mentioned you are a dad to three girls. What do you think are the lessons that you are bringing to them as a dad, from your professional career? I think we all can have our journeys, etc.

But here I think sometimes people say, well, if you're in the startup, you're anti-family because you're all committed to the, to that world, and I think this sort of like, not to me strikes me as not over. You know sometimes it requires you laser focus and some moments in time, but it feels like there is a better answer to how you could apply the lessons from the kind of fast-moving technology world, the unpredictability of it, to being a parent and I think in my view I've seen you in a parenting role and I admire you, know what you bring to the table there. You show out fully. I think even for me and our audience it would be great that are still early in their career. What can they learn from being in the startup that they can bring when they're building their family startup?

0:59:54 - Dmitry Shevelenko

I think the part that I don't feel good about is they’re not startups. Actually, you have more, especially in kind of the modern, post-Covid kind of work dynamics. It's not like it's not always about office time. I don't feel like I have a lot of work or travel, but I don't feel like I got a lot of hours of being in the same physical space, if your brain is somewhere else, that's the part that I feel I still need to work on right. It's like when you're there be present, and kids have incredible bullshit detection skills so they can tell when you're like thinking about something else, even if you know it's you're ostensibly there, I think there's a balance of communicating your excitement. You know that there's like something really special about working on problems that are meaningful and that other people think are meaningful, and that you know the kind of engaging my kids, especially as they get older, like talking about you know what Perplexity is doing and how we're doing it, having that be part of the things we talk about and kind of bring them in as almost advisors on the journey, whether it be our marketing or product feedback or all those things. So I think there's a way to integrate it.

But I think, hopefully, like the thing that sticks is they sense, you know you can be really like excited and passionate about your work and that creates meaning for you and so if that's a takeaway they have and that you know it's ultimately, I think the immigrant mindset of like you know we ultimately work hard to like provide unfair advantages for the next generation, and that's certainly part of my drivers is like you know, like we're working hard, so you know like my kids can have resources and things that can kind of help them realize their full potential. And you can't communicate that to them and it's tricky, but I mean certainly there's no silver bullet. Whenever I leave for a trip, it's awful and they make me feel bad and I feel bad and there's still a lot of work to figure out how to manage it all.

1:02:38 - Alex Shevelenko

Well, I am confident that you'll make your ancestors proud, Dmitry, so good to connect in this mixed capacity. Thank you again for joining us and sharing your story of Perplexity and beyond!

1:02:58 - Dmitry Shevelenko

Okay, this is fun!