Be the architect of your AI-driven future at our digital event "Blueprints for Generative AI."

NEW Tool:

Use generative AI to learn more about

Product Launch: has officially leveled up its integration with Snowflake’s new data quality capabilities

PRODUCT LAUNCH: enables trusted conversations with your company’s data and knowledge with the AI Context Engine™

Upcoming Digital Event

Be the architect of your AI-driven future at "Blueprints for Generative AI." 

View all webinars

Another sucky year for startups and investment? with Eva Nahari

Clock Icon 68 minutes

About this episode

If you really want to win, you need to take some risks and be the first with something new. But how do you minimize risk? Eva Nahari, Principal at DNX Ventures, talks to us about the current genAI investment environment.

Tim Gasper [00:00:01] Hello everyone. It's time once again for Catalog& Cocktails presented by data. world. It's your honest, no BS, non- salesy conversation about enterprise data management with delicious, tasty beverages in hand. I'm Tim Gasper, longtime data nerd, product guy, customer guy at data. world, joined by co- host Juan Sequeda.

Juan Sequeda [00:00:18] Hey, Tim. I'm Juan Sequeda, principal scientist at data. world, and always it's a pleasure Wednesday, middle of the week towards the end of your day, and here we are to just spend some time and chat about data and chat about all the good things, but also, hey, sometimes things suck, and I think we're going to talk a lot about that. And I'm super excited to chat today with Eva Nahari, who is a principal at DNX Ventures. Just a quick note, we started chatting with Eva last year and it was like, " Well, let's see what's going to happen this new year." And we haven't chatted to Eva in a couple of months, so this is going to be a big boom. I'm really excited to hear...

Tim Gasper [00:00:56] We're going to learn some things.

Juan Sequeda [00:00:57] ...What's going on in your mind. Eva, how are you doing?

Eva Nahari [00:00:59] I'm doing fine. It's Wednesday, it's Valentine's Day. Who wouldn't be excited about that?

Tim Gasper [00:01:07] Happy Valentine's Day to everyone watching. This is the special Valentine's Day episode.

Juan Sequeda [00:01:12] We're going to talk about how things suck, but okay. Or not, I don't know. So what are we drinking and what are we toasting for today?

Eva Nahari [00:01:23] I'm drinking hot water. It's my new religion. I love it. Riffing. We were joking before we started here, it's the beer without the fulling feeling. So I'm drinking hot water, and since it's Valentine's Day and since most things are still gloomy in the world, I would like to toast to more love and compassion. So there you go.

Juan Sequeda [00:01:57] Love and compassion, definitely cheers to that. Tim, what are you drinking today?

Tim Gasper [00:02:02] Today I am drinking a cocktail that I've had various permutations of on this show, an old- fashioned, but I have a new permutation of it, which is a root beer old- fashioned. It's a root beer symbol, syrup made from root beer and little added sugar and a few other things. And then you make an old- fashioned with it. It's very interesting. It just has that root beer kind of taste to it. It's cool.

Juan Sequeda [00:02:25] How much bitters do you need to put in to actually get a root beer?

Tim Gasper [00:02:29] Well, it's literally the syrup is root beer based.

Juan Sequeda [00:02:32] Oh, it's beer based, okay.

Tim Gasper [00:02:32] And then I do add a little bit of orange bitters to it.

Juan Sequeda [00:02:38] I got to find that one. I'm actually having just very normal, but really delicious, refreshing rum tonic with some orange bitters and lime. Let me enjoy this. And cheers to love and compassion, happy Valentine's, everybody. All right. So our warmup question today is what do you suck at?

Eva Nahari [00:02:59] I think I suck at being patient around inefficiencies. I'm very Swedish in a way. You should be efficient in what you do. You shouldn't waste time. I don't know if you know this about Swedish culture, but when I first came here to the US in the mid- 2000s I was so annoyed people using my name everywhere. It's not something common in Swedish culture to like, " Hey, Eva. Eva, what do you think about this? Eva, by the way." And I got annoyed and then I started thinking what in Swedish culture, what on earth is going on here? And I think it's an efficiency thing. You know I'm here, don't use my name, you don't have to call my attention. I'm right here. And that encapsulates a lot of the efficiencies in my core. I think that goes into work, into family, planning, organizing family weekends. When things are not as sufficient, I get very irritated, so I suck at managing that, I think.

Tim Gasper [00:04:19] That's so interesting. I think there's that book, How to Influence Friends, and in that book, one of the first chapters is use people's name. So I think Americans really take that to heart. They're like, " Oh, Juan Juan. Eva, Eva, Eva." And so everyone's always trying to use your name. I can imagine other cultures are probably like, " Why are you doing that?"

Eva Nahari [00:04:38] Yeah, it feels so personal like, " What do you want from me? Why are you in my personal space? We're just coworkers." It was a very hard adjustment in the beginning, but I've read that book and I realized it's written by an old white male from America.

Tim Gasper [00:04:54] Quintessential industrial, complex American, right?

Eva Nahari [00:04:57] Yes. But it works. I used it. I took a rule from that book every week and implemented it to see if it worked, and it worked. It is an American Bible.

Tim Gasper [00:05:10] It is some good advice on interpersonal relationship, but definitely a biased perspective. What about you, Juan, what do you suck at?

Juan Sequeda [00:05:19] I was thinking I'm going to go down sports. I've never been good at sports. There was one sport I was good at and I enjoyed, which was playing tennis. But then I did a sport that I sucked at, which was snowboarding. I skied when I was a kid, but didn't do that much, and then I went snowboarding and then I think on day three I just broke my ankle and then I never could play tennis again. So the only sport I was pretty good at that I enjoyed, I can't even do it. And all other sports I suck at. So it's kind of sad, depressing.

Eva Nahari [00:05:48] Can I ask a question?

Juan Sequeda [00:05:48] Yes.

Eva Nahari [00:05:48] Are you good at chess?

Juan Sequeda [00:05:48] No.

Eva Nahari [00:05:48] Okay.

Juan Sequeda [00:05:54] I never got into it. I don't even know the rules. I never played it. If I were to play chess right now, I inaudible

Eva Nahari [00:06:00] I was going to ask you if you count that as a sport, but no.

Juan Sequeda [00:06:06] It's another good question. What is a sport?

Eva Nahari [00:06:08] Exactly. Go to the question behind the question.

Juan Sequeda [00:06:12] Yeah, now we get into the philosophy of things. One of the things I love to do when we talk about knowledge, what is soup? Have you seen that YouTube video on what is soup?

Eva Nahari [00:06:23] No.

Juan Sequeda [00:06:24] Oh, after this you have to go watch what is soup? Or anybody listening, just look up YouTube, not right now. But if you're watching live, not right now.

Tim Gasper [00:06:31] In an hour.

Juan Sequeda [00:06:31] Yeah, in an hour. What is soup? It's also what is a sandwich?

Eva Nahari [00:06:37] I've had that conversation many times.

Juan Sequeda [00:06:40] So what is sports?

Tim Gasper [00:06:42] Oh, okay.

Juan Sequeda [00:06:43] All right, Tim, you go. What do you suck at?

Tim Gasper [00:06:47] I'm sucky with being efficient. No, just kidding. Sorry, Eva. This actually is a little bit of an inefficiency, so pardon me. I'm sucky with being on time to things. Those who work with me and know me, I'm not good at being on time for my meetings and things. I have to work very hard at it, work very hard, but I'm trying. I think I'm getting better.

Juan Sequeda [00:07:16] Well, we have inefficiency, sports, and on time, so I don't know where that would take us. But anyway, let's kick it off with our-

Tim Gasper [00:07:24] Let's take it back to data, Eva.

Juan Sequeda [00:07:27] Honest, no BS, are we really in a sucky year when it comes to startups and investments and data and all this stuff right now?

Eva Nahari [00:07:35] You met me at a very tired point last year, November, December we started talking about maybe doing something like this and I said, " Well, I have no light in the end of the tunnel." And to be honest, I think the spring is coming to California, so I've gotten some sunshine, but that's about the only light that I see. It's mid- February so I feel a little bit more energy is coming back, and I'll explain what I mean by that, but I'm still cautious, very cautious that valuations or the ease of fundraising or finding a golden nugget in all this noise is going to happen in the next two quarters, at least. I could pretend to be hopeful about Q3 for all the fundraising startups, entrepreneurs out there, that's a few signals are telling Q3 is the quarter things will start turning around. And I have some data points to back up that statement other than reputation on the street. For instance, I've seen my friends in the enterprise budget decisions sitting on high chairs, they've started to open up their wallets, but it could be like it's budget planning time and we'll see what they actually spend because everybody's cautious still. It's like when you twisted an ankle and now the cast is off and you're still super sensitive walking on it because you don't want to do it again. I think that's where we are at. And Q3 might bring hope, but then I have the counter contradictory opinion of that that it's election year. So in my 20 years of working with enterprises and product and sales, election year, everybody holds their breath until after the election because then they know what to do. Then they can bet on what regulations are going to actually come down the pipe and what's going to happen if it's going to be a confusing year or not. I don't know if Q3 would be better.

Tim Gasper [00:10:02] It's tough year to make big bets when you're like, "Huh, I wonder what's going to happen."

Eva Nahari [00:10:07] It's cautiously not so sucky. That's my prognosis.

Tim Gasper [00:10:16] This is a weird time with I think what a year and a half ago or so folks were really sounding the alarm and then we seemingly have achieved the soft landing, quote unquote, but this is the only time in the history of these economic situations where we've had a soft landing. Usually there's always been some kind of a recession. So folks don't really know how to react to this right now.

Eva Nahari [00:10:46] I agree, that's been the vacuum situation. It's both strong signals that there's a bullish market and a bear market. What is it? Everybody has been confused. And economic and investing in bedding land, it's not good to be confused. You'd rather wait and see what others are doing or where things are heading. It's also adding to the mix, the macro economical events that very tragic, Ukraine and Israel, I don't want to go down that talking track, but it has this gloomy aftermath on the entire Europe economy, entire US economy. At the same time we have this AI hype that brings promise and endless of efficiency to the world. It's so many concurrent things that are confusing and big and happening at the same time. And in the middle of it, the aftermath of Silicon Valley Bank incident, they're doing fine now and I'm doing business with them, there's no aftermath to be handled at the moment, but when it happened, it caused a stir that we definitely didn't need. It's just so many major events happening in the same short amount of time that has added to the confusion. And if you're in my chair as a VC, early startup, confusion is not what you want.

Juan Sequeda [00:12:36] So you brought up the magic word here, which is the AI. It's a confusion because everybody's like, " Hey, what is this and how much should I actually go do something about it? And should I do little or should I do more because I'm behind?" But it's also an opportunity. So what is your message right now? I'd like to dive into this. What is your message right now to enterprise folks who are looking when it comes to given this current situation and what are the opportunities, what are the possible risks you should be taking? What is your message then also to startup founders who may be listening here saying, " What should be doing," and so forth? So those are the two things that go through my mind listening to this.

Eva Nahari [00:13:23] I have so much to share. I don't know where to start. Enterprises, as I said, they've been holding their breath, so they close their wallets, they do a little bit reset and clean up, and you've seen mass layoffs and you've seen it reduced in number of acquisition, you've seen not as much enterprise deal close for growth startups, et cetera, et cetera. So the hesitation is within enterprise economically, macro economically and locally in the US. That has an effect on startups. Their growth are not happening as fast because they can't... The ones focusing on inaudible might be in a different situation, but the ones who go after the big transformative ideas, they've been struggling with growth, and that happened alongside a year of overvaluation, 2021, the hangover of that. So you have startups with high valuations and not so hungry enterprises as customers. It's a very tricky situation, and I think AI came in the middle of that, contributed to the high valuations, sure, but the high valuations also was because there was so much dry powder and so many IPOs and so many unicorns. VCs need to continuously invest and spend, otherwise it's going to hurt them later on. That's how the whole VC fund model works. But AI hype, a lot of dry powder, it got extreme, and now startups are in a very tricky boat of navigating not such a hungry market in general, but also the confusion around generative AI. And now I'm going to double click on that. I've been in machine learning since the last millennial, so to me this isn't entirely new. I think the first generative AI papers were published somewhere in like 2012, and then the Google transformer paper that the whole ChatGPT is based on, was published in 2017, I think. So it's just the research that I'm into, to me this is not new. I've given lectures on Word2vec and the rag precedents, I don't know, I'm not as excited personally. It's just a natural evolution of a long, long journey of machine learning. It actually not should be called AI at all in my opinion. But when the paper came out, I have my opinions about what companies are good on go- to- market with enterprises and who are better with serving engineers, I'm not going to name names, the people who know know, but the intelligent move here was to take innovation and go to market with it. Innovation is nothing if you don't know how to go to market with it. And I have this analogy with Edison, everybody thinks he meant the light bulb or the electricity lights. And there were many versions before that that was in a lab with experts and it was dangerous to light it up and everything. But then Edison came and like, " Okay, this can be democratized. It can be put in homes, in offices, in workspaces, in stores. I just need to make it safe to use." So he invented the electricity light bulb system, the cheaper to produce light bulb, the switch on the wall, and suddenly people could use it. And then the magic was for everyone. And I see that happened a little bit around OpenAI, and I don't try to minimize the research and the eminent team trying to democratize AI, but that's actually their original mission is to democratize what they saw. It's like we want to democratize AI and bring it safe to the world. I think Edison is a perfect analogy for what OpenAI started out as. Then there was a very smart go- to- market company who came in with major investments and saw the opportunity to tie it into business, and the rest is history. But I think what OpenAI managed to do with that partnership, go- to- market partnership, is to bring what used to be in a lab and understood by experts like, " Oh, this is really cool technology," bring it to market is what revolutionized the world. So enterprises may see it as like, "Oh, it's new, risky, you can't trust it." Yeah, the general models might hallucinate a lot, it's getting better, but the technology is actually decades old. You can take the white papers, implement certain parts yourself. You can take open source models and train it over your own data, so you know what went into it. It is safe to do generative AI, but I think why enterprises are so hesitant is because it's so new to them and the general models might come with some risks.

Tim Gasper [00:19:19] I think this is an awesome perspective on despite how exciting this is, also how organic and just incremental this is at the same time. And to that degree, where do you feel like the hype is warranted versus where do you feel like the hope is excessive?

Eva Nahari [00:19:52] Very good question. Let me think. I think I have two answers. Let's see where it goes. I feel it's not so structured, but this podcast is about raw, just what's on top of my, right?

Juan Sequeda [00:20:14] Yes.

Eva Nahari [00:20:16] So bear with me. The first answer would be from an efficiency perspective, not surprising. I think there's no question about how tremendously helpful Generative AI tools, or applications, I should say, are for the individual. So a marketing person, I use Bard and ChatGPT on a weekly basis as an individual.

Tim Gasper [00:20:54] Almost every day.

Eva Nahari [00:20:56] And, of course, I double check, triple check, and go to the source of things, but I use it for the right use cases. It's not like I share my sensitive data, I ask questions about, okay, can you find me the top three, blah, blah, blah? And then when they do that I go and look at it, and it's just not sensitive use cases. And then sometimes it's like, " Okay, here's a long text, can you summarize and give me the top three bullet points?" Efficiency for the individual, no matter what work plus position. I know CEOs who use it. I know marketing people, sales, engineers. It's just incredible what it can do for one person in the role on a daily or weekly basis. Now, from an enterprise perspective of automated workflows and tooling, let's come up with a better claims or a insurance proposal. I'm not sure. Then it becomes a little bit like you want a human mind involved to make sure it doesn't affect your customer or your reputation or your trustworthiness. There are use cases and use cases. So generative AI is tremendously valuable for content and nonsensitive use cases. I think hands down the hype is... What's the right word? In its right. It's a tremendous efficiency revolution. That happened almost overnight. Everybody can use it.

Juan Sequeda [00:22:44] But, sorry, on the individuals side, which I fully agree, but those individuals also working within the enterprise, it is then by transitivity helping the enterprise too, because they're being more efficient in their job.

Eva Nahari [00:22:59] And that leads me to the other bucket. As a startup, I would probably not go into the end consumer. It's almost like I compare it to consumer facing business like what OpenAI did with ChatGPT and Google did with Bard and others. It's almost like a consumer business. I don't want to touch that. That's great, it's efficient, but as a startup I would not go into that space. It's moving very fast. There's already an app store, think almost OpenAI is the iOS and then they open the app store. If you're an entrepreneur, go build an app, and it's going to be the app race all over again that we already know if we lived through the mobile era. So there's going to be millionaires there, but there's going to be a new kid on the block every day and it's going to be up to the marketplace search engine to highlight you, and it's the Apple model.

Tim Gasper [00:24:09] There's always new apps.

Eva Nahari [00:24:11] And that doesn't excite me as much as, okay, the other bucket for enterprises. Let's say you want to simulate different scenarios for machinery or for a logistics route, where I see generative AI could help so much because humans can only come up with that many ideas within an hour or within a day or within a week, and generative AI can support and extend what humans can do in ideation, in taking something and transform it into something else. Here's a blueprint, how do we make it more environment friendly? Here's a surgery plan... God, don't do this. But here's the surgery plan, how can we make it less risky? If you have something and you've trained a generative model on real data that no one else has access to that is super unique in 10 years and super safe, and you have a smart inaudible, I'm super interested what can happen in that bucket, but that's what enterprises are a bit afraid of too. And I think if I were an entrepreneur, I would be a high risk- taker, and generative AI can transform more than for the end user. Do you see what I'm saying?

Juan Sequeda [00:25:31] So let me repeat this.

Eva Nahari [00:25:37] Good luck.

Juan Sequeda [00:25:39] As we take our notes here. Okay. So what is warranted as a hype from, it's more on the individual side? It's like, " Yes, it's making us more efficient. I can do these little tests and so forth." I like how you said it, use it for the right nonsensitive use cases, always triple check things. So there's the obvious things you go do. And that's great for us. And then even, and if you are within your own job by transitivity, you're also applying that efficiency to the organization, but it's really focused for the individual.

Eva Nahari [00:26:14] Yes, correct.

Juan Sequeda [00:26:16] So that hype is definitely warranted. Now, should there be startups and new companies coming around to play in that efficiency game, you're saying, " Nah, because it's too small, this is going to be the app store all over again." Right?

Eva Nahari [00:26:30] Yeah.

Juan Sequeda [00:26:31] There will be the small development companies, the consultancy things that they'll be doing like they're doing fore these apps, they're writing the apps. But you'll have a small consultancy team making millions of dollars because they created this cool thing and that's it. But there's like, " Am I going to invest millions of dollars to go create a startup to go do that?" Not worth it right there. So inaudible.

Eva Nahari [00:26:54] Exactly

Juan Sequeda [00:26:55] Now, from the enterprise perspective, you're saying, " Oh, if you really use generative AI to do something transformational," and the transformational idea, one of the things that you brought up here is ideation, for example. If you are going to use generative AI, either you're going to bring in and train your own model, whatever, focused on your organization and it's going to help you be more ideas that can help you come up with new things and so forth, how can I make this thing more X or less Y and so forth, that can really transform your business?

Eva Nahari [00:27:29] Exactly.

Juan Sequeda [00:27:30] The issue here is people are not taking that risk and they're like... So basically it's like I'm putting myself in your shoes. It was like, wait, if some company says I can really transform this industry, they've been doing things this status quo way, and we can really transform it to do this new thing where all this transformation is going to be gigantic. Wow, if we really accomplish that, that would be huge. Now, the real struggle that we have is getting the enterprise to really play in that transformational game?

Eva Nahari [00:28:02] And getting the data to do that, because the enterprise sits on it. So here we come into the conundrum, which is to do those generative models you need that kind of workflow data, you need access, and only the enterprises that's been around for 10 years... I'm exaggerating to make a point, but you get it, industrial companies like Deloitte audit or manufacturing blueprints, who sits on five years of that data and knows the workflow in and out is the enterprises, and they're not innovating as fast. And it used to be that a startup could come and innovate and serve the enterprise, accelerate the enterprise business, and then acquisition or IPO, but because this is tied to data, and now we're getting to the gist of why we're talking at all, because you're data guys, startups don't have access to that data. What startups have access to is general data, and that is not viable, competitive, long- term mode, then you're serving the end user or you're an app entrepreneur, which can make millions, but it's not what VCs like me would be excited about. So it's like I want the enterprise sensitive data to partner with startup entrepreneurs somehow for this other bucket to really happen, and I don't know how to make that happen. I see so much value. That's the transformation I see with generative AI. And I can't help it.

Tim Gasper [00:29:44] That's fascinating. This is a really great perspective, because I don't think it's something that gets talked about enough.

Eva Nahari [00:29:50] Well, maybe I'm wrong, maybe I'm the only person who thinks this way.

Tim Gasper [00:29:55] I don't think you're wrong. It makes a lot of sense. I think your analogy around iOS and the mobile explosion that happened 10 years ago, let's call it, although it's still happening now, but it was hype 10 years ago, there weren't that many billionaires that came out of the mobile explosion, but there were a lot of millionaires that came out of the mobile explosion.

Eva Nahari [00:30:16] I'm trying to think of a few though to not put them down.

Tim Gasper [00:30:20] There was Instagram and things like that, but for all the Instagram's there were way more Scan Pro, whatever, that helps you make your documents get scanned and stuff like that.

Eva Nahari [00:30:30] But Tim, let me ask, how many apps do you have on your phone?

Tim Gasper [00:30:35] I have probably 200 to 300, maybe more than that. I've had a phone for a long time.

Juan Sequeda [00:30:40] How much do you pay for them? Well, you pay with your data too.

Tim Gasper [00:30:43] I have a lot of free apps, but I do pay for some apps.

Eva Nahari [00:30:47] Okay.

Juan Sequeda [00:30:48] inaudible right now.

Eva Nahari [00:30:51] Sorry, I interrupted you, but...

Tim Gasper [00:30:53] No, you're good. Well, anyways, I think that analogy works out well. But then I think a question is so if you follow what you're saying here, who has the advantage? Is the advantage in the favor of the enterprise who owns the data who needs to bring the brain to them? Or is the favor in the case of the startup or the technology company that wants to somehow broker data or maybe has access to some unique data? There's a lot of companies out there that have interesting data sets that they sell or things like that. I'm curious about how you think about the power dynamics.

Eva Nahari [00:31:40] Then we come to a parallel track. So first I want to answer what you said. I think the advantage is where the data is, unless the startup can broker a partnership and get the data, but that's really hard because data is the core and IP, and many enterprises it's like, " That's our gold today. Data is gold, so we're not going to share it." Hence, I am going to round up to the enterprise on the advantage. And then there was one more thing which is the startups are nimble, and that's what enterprises are not, but I don't know how they're going to solve that to actually innovate. But it draws to another parallel that ties into why this is a sucky year. It has nothing to do with the year. It's a sucky situation right now because if the enterprises owns the advantage, guess what happens if the enterprises innovate? The startup ecosystem will starve. And that's why I as a VC am a bit worried that there's too much innovation or the platform for innovation is either by the giants or the enterprises right now, because I've seen so many pitches and it's all not unique data, and then it's going to be commoditized by the big ones. That's just how I feel.

Juan Sequeda [00:33:15] Basically you're saying if you're seeing companies, startups saying, " Oh, we have this new model on this particular domain, whatever," it's like, " Yeah, good luck."

Eva Nahari [00:33:28] If it's in the public realm...

Juan Sequeda [00:33:28] It's already old.

Eva Nahari [00:33:29] ...It will be sucked up by the giants.

Juan Sequeda [00:33:31] This is a very fascinating point. You're saying, if I got this right, enterprises have the advantage and it's because it's their data, so they have that advantage, therefore they're in the position to innovate. Now the question is are they nimble enough, will they be able to innovate or not? Now, enterprises who will innovate, then from your perspective, well, then startups, they're going to be in a situation they're going to be starving because all the innovation is coming from the giants, and that's actually something new... Usually innovation comes from the smaller firms.

Eva Nahari [00:34:04] It's a whole new situation in my entire three- year career of VC... No, but seriously, I've talked to other VCs and they feel this is new, there's a starvation, but remember, we're talking about generative AI. It's not like all startups are inaudible We started talking about it's the hype, rightfully a hype. I'm like, " I'm confused." It's starving the startup ecosystem right now the way it's going, and of course it's business and it's fair and good for the giants to take so much of the room here, and of course they're going to be app millionaires, we can agree on that. But for VC startup, transformative for enterprise, that's the swim lane I'm in. It's a very new situation, and should I focus on this big golden pot that I know is there if it just happens what needs to happen, the data partnerships or the nimbleness? All I can say is I've started collecting enterprises that I want to keep close and I am going to be some kind of matchmaker maybe. I don't know. It's just very hard to see how this phenomenal technology that is mature, but you need to know what to do with it and give it the universe it should control and that you can trust. How do you give entrepreneurs access to that data? I don't know.

Juan Sequeda [00:35:44] So for the data leaders, executives who are listening, for CDOs and folks that we know that listen to the podcast...

Eva Nahari [00:35:51] Call me. I'm a matchmaker.

Juan Sequeda [00:36:00] The message I'm hearing here is that, look, there is this amazing opportunity right now for you to actually be an innovator, which is not a common thing, but to really succeed in this innovation with this wave of AI, the foundation really is on data. You will be able to innovate because the only way we're actually going to provide value here is because it's based on the data that you have and your own. Therefore, if you really want to be a leader here, you need to have the strong foundation of your data, because if you don't, then you're not going to be able to innovate. So this is actually a call, a huge motivation saying where is your data strategy, how are you actually putting this all together. That's how I'm seeing things right now. If I'm listening, shit, hopefully I'm now being able to organize that message to my leadership to saying, " Look, we can innovate, we can be better than competitors, but this can only happen if we are investing in the strong foundations of just the core principles of data and metadata dimension and all that stuff."

Eva Nahari [00:37:12] And if you're nimble enough or if you partner with a number of startups to help you out, it's just like I don't know if it's a I can do it all myself kind of strategy that will work. I think many ecosystems need to be facilitated, and I think I can help with that, but I also fear that in most companies, people don't know where their data is. It's in all these different silos, there are no standardized way of handling acquisitions, rapid growth, restructuring, new chief on the top of the hill, every enterprise is messy and their data is just another reflection of that. So I don't know if it's even possible to get to that conversation that you were suggesting, which I want to happen, because they don't know where their data is or it's such a mess, how do you even do this in an enterprise? And I think that's the missing hand. I have two hands, but the third one is missing.

Juan Sequeda [00:38:29] You need to solve these basic things.

Eva Nahari [00:38:33] I'm not selling your product.

Juan Sequeda [00:38:34] No, no, no, but I mean-

Eva Nahari [00:38:36] I did not get any percentage of this transaction.

Juan Sequeda [00:38:40] No, but I mean what you're saying is you have this opportunity to lead, you have this opportunity to innovate, you have this opportunity to beat your competitors, but it's only going to work if you have the foundation of data, which is going to start. The basic stuff is you don't even know what data you have.

Eva Nahari [00:38:58] Where's the third hand? No, but you summarized it well. It starts with the data. Enterprises are sitting on gold. They're in a mess, they're not nimble enough, but it's a very unique situation for enterprises to innovate. Now, I don't like that as this early stage startup VC, I need to find a way for startups to help these enterprises.

Juan Sequeda [00:39:31] So what's missing in the landscape for this? I'll be honest, I'll be honest, I honestly think we don't need any more companies, more tech. From a technology perspective-

Eva Nahari [00:39:47] Everything has been embedded. No.

Juan Sequeda [00:39:48] Yes. Honestly, there is no need. Up to now it's how do we deal with the situation is that you are able to have that aptitude to go say, " I'm going to go innovate. There's a reason why I'm going to go focus on this." It's all just people, process, culture. There is so much technology out there. There argue there's no more need for this stuff. Maybe how do companies want to go set up their own LLM and... Okay, maybe on this new space, but for the foundational data stuff, no, I will call BS if we need another data type of company.

Eva Nahari [00:40:21] I feel like I want to argue with you.

Juan Sequeda [00:40:25] Perfect. Let's go. Tim is being very quiet, so I don't know where his head's at.

Tim Gasper [00:40:29] I'm just going to take my cocktail and inaudible

Eva Nahari [00:40:31] We need to drill Tim at the end of this. There's going to be a quiz, Tim, pay attention. No, but I think I want to argue with you first is the wheel always gets reinvented, so give it five, 10 years and there's a new kid on the block doing-

Juan Sequeda [00:40:48] Of course, but that's the problem that we're just doing the same shit over and over again.

Eva Nahari [00:40:54] I was in that view set, but it isn't the same. If you look at... I lack an example now on top of my head, but if you look at data warehouse, it was on- prem, it was clunky, it was impossible, it needed experts, and then 10 years later there's a cloud service version times 10. So it's the same thing, but serving the use case and the change in the market of more volumes of data, more rapid pace, better. So it's always better. And the best investment you can make is something that there's already budget for. There's a purchase process, people know what it is, there are skills already there, but it's better. Think about it.

Juan Sequeda [00:41:52] I give this talk, and I start out the talk giving just a copy and paste of someone says, " Oh, this bank is trying to go innovate, blah, blah. They have these problems about integrating data because you don't have the shared identifiers, blah, blah." And I shared this example and people are reading it and everybody goes, "Yeah, yeah, yeah, that's exactly the problem I have." I copied and pasted this from a paper from 1992, 30 freaking years. Do not tell me that we have not been able to solve this problem inaudible

Eva Nahari [00:42:23] You can get as upset as you want about this, but if you use that same diagram and go and ask for investment, I bet you get funded too. It's just how it's happened.

Juan Sequeda [00:42:34] Then we're not solving the problem. We're not solving the problem. People are making money. People are making money, the problem is not being solved. And this is the thing that frustrates me all the time. This is inaudible

Eva Nahari [00:42:47] Go build another startup then.

Juan Sequeda [00:42:49] No, this is more about change management. This is the whole people process.

Tim Gasper [00:42:53] Well, maybe what's interesting here is the innocent bystander in all of this, I get to just contemplate is, Juan, what I hear you saying, and Eva what I hear you saying is that... So the problems are the same. We still have the same problems. We don't know what the data is, we don't capture the meaning in an effective way, the quality is a problem, the data is over there, but I really need it over there, and I wish I could query it faster. The problems are the same. But what changes is technology evolves, form factors evolve, mediums evolve. And so it was Informatica yesterday, it's Fivetran today, and there's going to be a new data integration thing tomorrow, and it's going to take advantage of the new trends.

Eva Nahari [00:43:39] Maybe I'm launching one.

Juan Sequeda [00:43:41] inaudible it was streaming, it was a reporting... But effectively it's all the same things.

Eva Nahari [00:43:48] But, Juan, I don't agree with you. As I said, the whole cloud migration happened and it's not the same thing. It is the same thing, but with new requirements and new circumstances. And that's why I'm going to bring up the other challenge, when it comes to the data pipeline...

Juan Sequeda [00:44:05] Let me stop you. I just want to pause. Depending on who you ask and what problems they're tackling, they will solve it's better. But if you zoom out and at the end of the day we're trying to go make money and save money, and people are asking these questions about how many customers do I have? And I still can't answer that, that problem has not been solved. And even though internally if we solve a bunch of technical problems, but the bigger one is it's still the same and that's what frustrates me. inaudible

Eva Nahari [00:44:31] I want to give you a little hope from the Silicon Valley glass is always half full perspective. Every cycle of reinventing the wheel to solve for the same macro problem comes with slight improvement. Things have become better, but the world has become more complex because now we're all on our phones all the time and more data is generated everywhere. As Tim wisely as the observer and coming in with the key points here, there's technology not necessarily in the space you're in, but other technology transformation that shifts and suddenly there is a need for a new cycle of the new kid on the block. Same stuff but new. And I think that's a very good business to be in both as a VC and an entrepreneur and enterprises know how to adopt it and it's known. And how enterprises buy is with least risk. What can I do that gives an incremental and significant improvement during my career stage in this title? Because big companies are more career people, that is least risk for getting fired. No one has been fired for buying IBM. Maybe that's not true anymore, but there's always like, " Okay, I need to do something in two years to show transformation to get to the next game level," and therefore people want to buy the same things. That's just human nature.

Juan Sequeda [00:46:20] I fully agree with that. I think it always boils down to the incentives, human incentives.

Eva Nahari [00:46:25] But it's not bad. I'm not saying it's bad. It's great because with every improvement, every new cycle that's new stuff, new improvement, new careers, new people involved that can grow, it's business and it's good and healthy, and your heart is sad because inaudible

Juan Sequeda [00:46:50] My heart is sad, I agree with you, but I'm sad with having to agree with you. And that comes from just my academic scientific training.

Eva Nahari [00:47:05] I'm there with you, my heart is sad.

Juan Sequeda [00:47:06] As an academic, I'm sad that that's the status quo reality. inaudible

Tim Gasper [00:47:11] Maybe that ties to your feelings too in terms of efficiency. I see, Juan, that you're like, " But wait, but there's a more efficient path."

Eva Nahari [00:47:20] But I think you can change that sad heart to Valentine's heart again, and see it as that is the never ending pain and glory of the entrepreneur, to try, try again. And the wonderful thing of serial entrepreneurs is like, " Yeah, the last time we did it this way, now I'm going to try it this way." It's a beauty in that of learning and building on the past. I love that you read papers from the 1990s. The mobile phone was actually in Graham Bell's lab 100 years ago, it's not new, but suddenly technology happens so you can realize it.

Juan Sequeda [00:48:03] Let's see hat have over here. This is the proceedings of the first international inaudible

Eva Nahari [00:48:08] We needed a little bit of academia in this.

Juan Sequeda [00:48:11] So I do this inaudible The funny thing is I'm reviewing papers for inaudible 2024 right now.

Eva Nahari [00:48:18] Okay, there was a second bucket I was going to give.

Juan Sequeda [00:48:21] Sorry to interrupt.

Eva Nahari [00:48:22] I'm trying to remember. It's like reinventing the wheel.

Tim Gasper [00:48:30] I think when I mentioned Fivetran and Informatica and pipelines, you started to go down a path.

Eva Nahari [00:48:35] No, I said I might be looking at one and then I left it there intentionally.

Tim Gasper [00:48:43] Oh, okay. There you go.

Eva Nahari [00:48:43] Cliffhanger. No, but I think imagine them being an entrepreneur coming up with something completely new. My journey with Cloudera, Hadoop was turning the whole ETL to something else, first bring the data then do the ELT, BLT. No, but it was an educational sell. It was an uphill. And then they should change their whole organization to have this data team. And it was organizational transformation because they needed new people. It was educational because the whole thinking that people have built their careers on was turned on the head. And once they realized that, they're like, " This is amazing. How do I get better at that? How do I build a career around it?" But it took years to get budgets and understanding how to make the technology adoptable and survive, because it needed organizational support and people trained on this. It's a heavy lifting. Poor entrepreneurs. I was in the middle of that and I know the pains and the never ending struggle on getting people to understand the ones who've done things a certain way for 20 years now should change? Uh- huh. So choose your battles. If you want a smooth path, you might not get to satisfy that academic heart, but you build a beautiful business that actually helps enterprises incrementally. Or you go on the hard path of, yeah, oh my gosh, I have to create a market or I have to create a budget or I have to create a buyer. Oh my God, please don't do that to yourself. But that is also inaudible transforms the world. So as a VC, I'm open to either,

Juan Sequeda [00:50:46] I have to say, Eva, that our back and forth where we disagreed but ended up agreeing, but I agreed with a sad heart, but it's Valentine's, so I'll turn that into a nice part. I think this was one of the first times we've had some interesting disagreements on this podcast like this. I appreciate having this discussion. This is the honest no BS discussion that we need to have more. And thank you for that. Really, thank you for that.

Eva Nahari [00:51:12] And I appreciate you guys. This has been awesome. I get to vent all my fears.

Juan Sequeda [00:51:18] I get to rant over and over again about this stuff I'm annoyed about.

Eva Nahari [00:51:23] It's a very confusing market, but there's always hope. And then some new innovation happens and everything is back inaudible

Juan Sequeda [00:51:29] We could keep talking more and more. And I really do want to, but we got to go to our next stuff, lightning round questions. I'll go first. Given enterprises have their own data, do you feel eventually enterprises will train and manage their own enterprise LLMs?

Eva Nahari [00:51:46] That is already happening. I think they will be more specific and smaller, but it doesn't exclude the use of larger general LLMs in parallel. Both is my answer.

Tim Gasper [00:52:03] Second question, with all this power now moving to the enterprise, because they have the data, do you see that investments, the money, the budgets at enterprises going into data as a percentage is going to increase? Is data going to increase as a percentage investment for companies?

Eva Nahari [00:52:29] I want it to. I really want it to, but I am not sure when you work inside an enterprise that you have the external perspective. We have this executive advisory board for enterprises that gives them a little bit Silicon Valley external perspective maybe, or internal Silicon Valley perspective if nothing else. I think it's hard to look at yourself from the outside and know what you have. So I'm not sure it's going to happen overnight, but I hope so. Their own data is what they should consider very, very high value.

Tim Gasper [00:53:15] I like your answer there. Maybe I'll add to that and say it should, if there's so much value we can get out of it that it should. But the question is will it?

Juan Sequeda [00:53:25] All right. Next question. Do you see the biggest opportunity for generative AI for enterprise around industry specific data and use cases? Or will this be more horizontal and broadly applicable use cases?

Eva Nahari [00:53:39] So I think we touched on this really well. For the efficiency side of things, for individuals inside the enterprise, it will bring a lot of tremendous efficiency gains, but for transformative gains of new revenue, mass shift in how you do business, or whatever, it needs to be in the other bucket, I think.

Juan Sequeda [00:54:05] I like how you categorize this, efficiency versus transformational gains.

Tim Gasper [00:54:12] So efficiency, more horizontal, and individual for transformation more industry specific and enterprising?

Eva Nahari [00:54:19] I think that's a big takeaway from today.

Tim Gasper [00:54:23] That's a good distillation. All right. Fourth and final lightning round question. We didn't end up touching on this too much in our discussion today, but there's this whole movement and focus around AI tooling and MLOps and ML pipelines and things like that, do you see strong opportunity for startups that are focused on that space?

Eva Nahari [00:54:45] This was the other point that I dropped.

Tim Gasper [00:54:47] Oh, okay.

Eva Nahari [00:54:49] So reinventing the wheel was one point. The other point was the DevSecOps or DevOps pipeline, how will that need to look different when data is involved MLOps has been going on for 10 years, that was not new while I still worked at Cloudera, but it's like I think with the new hype and more about trusting ML and accuracy of data coming into it and not sending out any random prompt, there's AI safety that needs to be baked into the MLOps, whatever you put into that, but the data pipeline and the whole to productionize ML. Some of it has happened, but I think there's some new opportunity to rethink because enterprises want to trust it.

Juan Sequeda [00:55:47] It's that trusted. There is so much that we went through. Tim, take us away with takeaways.

Tim Gasper [00:55:59] All right, I'll try to distill it quickly. Eva, this was awesome, amazing chat today. We started off with honest, no BS, why is this a sucky year for startups and investment, so we jumped into it, and you started off by giving us your perspective on the macro environment and how there's good things and bad things happening at the same time is very confusing. There's indication that Q3 and beyond might get better, some wallets are starting to open up. The stock market is in general doing pretty decently, tech startups have mostly rebounded, but at the same time you've got enterprises holding their breath. They're not sure exactly what's going to happen coming up. It's an election year, difficult world situation, a lot of conflicts going on. And then yet AI, hype, promise, excitement, efficiency for everybody. It's like everybody's Popeye and they just ate spinach, that's what the AI will do for you. So it's a very confusing time. And you also talked about how that's impacting the world of startups and VCs because if enterprises are holding their breath, then that's going to impact startups who are going to struggle more with their growth in sales. And of course that makes things difficult for investment and VC and things like that, because where can they make bets that actually will have not just a return on investment, but a rapid return on investment, so it's a very different time. And GenAI makes this all quite confusing because it's not new, as you noted. It's based on things, research that's been going on for over a decade, it's a natural evolution of machine learning. We probably even shouldn't call it some new thing. Kudos to the person who was the marketer or whoever who said generative AI-

Eva Nahari [00:57:44] Go- to- market.

Tim Gasper [00:57:44] Oh my god, you chose a great phrase, congratulations. But really it's just machine learning. It's statistical models taken to the next level. So you gave the analogy of Edison and how he created the light bulb, but he didn't really create the light bulb. He helped with the go- to- market and the democratization of the light bulb.

Eva Nahari [00:58:02] And the safety system.

Tim Gasper [00:58:04] And how to make it safe, how to make it approachable, how to make it affordable, so there's a lot to that. And ultimately enterprises are just trying to avoid risk, which is an important aspect and dimension here, but so much more. But, Juan, over to you. What were your takeaways?

Juan Sequeda [00:58:21] Tim posed a very interesting question, which was where is the hype warranted and the hope excessive? That ended up to our large discussion here. So this is where we go into the efficiencies. This is where we really see how individuals can be much more efficient, use it for the right nonsensitive use cases, triple check your work for summarizations, and what you highlighted here is that, hey, for startups, it's not worth getting to an efficiency gain for an individual perspective. This is going to be just for app developers, people, small consulting firms will get into that. So the opportunity here is in the enterprise and you're talking about like, " Hey, this is going to help people to go for ideation, just really transform what the enterprise can do." But there one issue is that to be able to do that transformative work, you really need to have the data. So the enterprises have the data, the startups don't have that data, and if they do something in the industry for data, that's still going to be old because you need to really have the data for that enterprise. So really here is what's presenting is an opportunity for the enterprises to really be innovative. But are they nimble enough like startups? They're probably not. So how is this going to work out? So at the end of the day, enterprises have the advantage because they have the data. So they are and they will be the ones innovating more than startups, which is this brand new situation, because then you're not going to have startups innovating as much because the enterprises are the ones who are going to have that advantage. However, to actually innovate, you need to have that strong foundation of data. What happens at these large organizations, they don't even know what data they have. Every enterprise is messy. So we just had this whole discussion about how you really need to start investing in the foundations of data. And then we even had our very fascinating discussion between Eva and myself where we had a disagreement on really how much are we reinventing the wheel and how much is actually needed or not. My position is that we don't need more technology. That's how I started. And Eva disagrees with that. But at the end of the day, I think where we agree is that even though the new technology coming around, which may be seen as reinventing the wheel, is always advancing things little by little, maybe not as much as we want it to. And I think where I do agree with that, but leaves me in a sad position is that we should be doing more on some stuff and less than some other things, but that's just me coming from my academic scientific point of view and trying to be sometimes too idealistic. There's a lot of nuggets there. And I just want to have people know that they should just listen to our fantastic discussion and disagreement. Because, at the end of the day, we need to have these civil discussions, we need to be able to agree, and be able to put our positions, and at the end that's how we advance.

Tim Gasper [01:01:18] Ideally with a drink in hand.

Juan Sequeda [01:01:19] That makes it even better.

Eva Nahari [01:01:20] Ideally with people who are smart and entertaining. Kudos to you guys for hosting a wonderful podcast.

Juan Sequeda [01:01:29] Anything we missed on takeaways?

Eva Nahari [01:01:32] No, I want to say some app developers, consultants, I think there are going to be a couple of winners there on the GenAI app side. It's more that I'm not looking there. I'm looking at the transformative for industries to come. Just a minor afterthought, I guess.

Tim Gasper [01:01:56] Different angle, different level.

Juan Sequeda [01:01:59] So to wrap up, three questions. What's your advice? Who should we invite next? And what resources do you follow?

Eva Nahari [01:02:06] My advice to whom? To

Juan Sequeda [01:02:12] Whatever you want, whoever you want, about data, about inaudible

Tim Gasper [01:02:16] ...And all the listeners

Eva Nahari [01:02:22] In hard times, some of the best ideas and startups are born. And if you're in it already, navigating hard times and being resourceful is what's going to become the next level of entrepreneur leadership, and it's only going to give you better tools for the next game level. So it's a shitty, sorry, sucky year that we're coming from, but if you survived it, pat yourself on the back. It was not easy for anyone.

Tim Gasper [01:03:08] That is very good advice.

Juan Sequeda [01:03:10] And you see a lot of big innovations at companies being built in times when they're hard.

Eva Nahari [01:03:16] Exactly.

Juan Sequeda [01:03:20] Who should we invite next?

Eva Nahari [01:03:22] I would say I'm a fan of Denise Persson, the CMO of Snowflake. And she's been out and about lately, maybe she's up for it. I like her. She's cool.

Juan Sequeda [01:03:41] That's part of the Swede data...

Eva Nahari [01:03:47] Yeah, she's a fellow Swede, and the Swedes are slowly taking over the world inaudible

Juan Sequeda [01:03:57] We've had Emile, a friend inaudible on the podcast. We've also had Erik Bernhardsson from Modal, he's also a fellow Swede, you. So we need to get more.

Eva Nahari [01:04:12] Well, I can call my Swedish mafia.

Juan Sequeda [01:04:17] All right. And then finally, what resources do you follow, people, blogs, conferences, magazines and newsletters?

Eva Nahari [01:04:38] I follow Alchemist and PitchBook and all that for my job. And then there's... What's his name? I don't even remember, Michael something. I can send it to you guys, you can post it somewhere maybe. What popped into my head, I listen to the Swedish podcast, Swedish radio podcast that is called the Trend Spotters. And each week they get to come up with a new possible trend. They find three data points to back it up. And I think although it's an entertainment program about what's coming and there's been a lot of generative AI trend spotting, it's opening my mind. And I think this is what I want to highlight with that resource is I tend to listen to other things outside my zone, outside my area, because that brings me the most new perspective on things once I read what I'm supposed to read. So I don't know, my advice is go listen to something else. It's the cross- pollination that matters.

Juan Sequeda [01:05:59] Wow.

Tim Gasper [01:06:00] That is good advice. It's easy to go into your one silo. Sometimes the best ideas come from inaudible

Eva Nahari [01:06:07] Listen to something you're curious about, not your space, and then you get new ideas. That's been my recipe. Listening to molecular biology development in the last 10 years. It's not anything I'm interested investing in, but I learned a lot, and that center of the brain where you feel the aha moments triggers other parts that makes my job more exciting, but also maybe I see things differently. At least that's what I believe.

Tim Gasper [01:06:47] New connections get opened.

Eva Nahari [01:06:48] Yeah. Massage the curiosity part of your brain on a weekly basis. That's what I mean.

Tim Gasper [01:06:56] I'm adding that to the advice here.

Juan Sequeda [01:07:00] inaudible Right there, massage your curiosity part of the brain. Have you massaged a curiosity part of your brain today, or this week?

Eva Nahari [01:07:07] Talking to you guys, that was new.

Juan Sequeda [01:07:09] Phenomenal.

Eva Nahari [01:07:10] What else did I do? I went and talked to a Stanford student, a young lady, a freshman whose dad wanted me to connect with her while she's here. Not her dad, her uncle, sorry. And that was really exciting, getting new generation perspective on things. Oh, wow. That was so valuable to me.

Tim Gasper [01:07:36] That's another good tip.

Juan Sequeda [01:07:37] That's another good tip, yeah, to see where's the future going.

Eva Nahari [01:07:40] Should I ask her to come on this inaudible She would.

Juan Sequeda [01:07:43] Eva, this has been such a pleasure. And I'm back because I've just missed having just great conversations, debate where we don't have to agree, this is super important. That really just fuels me up because that's how we learn, that's how we get out of our comfort zone.

Eva Nahari [01:08:02] Well, I'm glad. That was a small, tiny disagreement.

Juan Sequeda [01:08:06] It was small, but that's the one I remember about this one. Actually, I'm going to post about this on LinkedIn. Everybody like, " You got to watch this because we disagreed." Eva, thank you. Thank you so much. Just a quick reminder, next week we have Karen Mepin, who's a director at inaudible talking about data governance, and now we're calling it AI governance, and we're calling BS on all this stuff between governance and AI and all that stuff. So it'll be a fun one. And as always,, thank you so much. Lets do this every week. And happy Valentine's to everybody. And again, Eva, thank you so much.

Eva Nahari [01:08:34] Thank you guys.

Juan Sequeda [01:08:35] Cheers.

Tim Gasper [01:08:36] Cheers.

Special guests

Avatar of Eva Nahari
Eva Nahari Principal, DNX Ventures
chat with archie icon