Apple AI Ambitions Pan-caked with Ann Bordetsky of NEA + Gokul Rajaram and Michael Gilroy | Okay, Computer.

Welcome to OK Computer. I'm Dan Nathan. I am joined again by Ann Borodetzky, she's a partner at nea. Ann, welcome back to the pod.

Thank you Dan. Thanks for having me back. Always love being here.

Well, listen, I really enjoy the convo because of where you sit right now in the, well, I would just call it the Gen AI ecosystem. Some of the investments that you've been making over the last couple years, you and I started talking about them, I want to say early 2023. You brought on Arvind Srinivas, the CEO, co founder of Perplexity, I think a couple months ago and I've seen him all over, he's doing a great job and really not only talking about the company that he's building and the products that he's shipping, but becoming a voice on what is a very nascent, I think we would both agree sector. And he's doing it with the help of you and a whole host of others. And I think the narrative is really good. I'm just curious real quickly, thoughts on that because for you, I think you guys led his Series a in early 2023. Back then no one knew who, no one knew what the product they were going to ship. And now it seems like everybody who's focused on this space knows the product and uses it.

Yeah, look, I think Arvin's getting a lot of attention and the company is getting a lot of attention and interest because they are on a mission to make AI useful for consumers. Right. It's not research in a lab, it's taking the best in class models that are coming to market it seems like almost every single week and trying to figure out how to take that and turn it into an incredible useful experience that will be relevant for millions of people around the world. I think that a lot of people are drawn to that. Not the theory of AI, not the scary sort of dystopian view of AI, but like what can AI help me do today? And I think that is the number one reason that Perplexity is resonating, continuing to grow organically as a company, continuing to grow globally. Now it is truly a useful knowledge assistant and we can see how that might play out over time. But look, I mean people are busy and they want useful AI and I think that is one of the best frontiers for builders right now.

Yeah.

Well as we're recording this right now, it's Tuesday afternoon. Google is hosting their I.O. summit conference, whatever you want to call it, introducing a whole host of products. We're going to Talk about that also OpenAI yesterday got in front of the Google stuff. So we're going to hit some of the things that they released and again how some of that stuff is being perceived. It seems like that almost everything that they put out is an immediate hit. We're also going to talk about some other use cases is some stuff in the hardware space. We're to talk about that Apple flattening ad in the controversy over there because I think you have some stuff to say on that also. Stick around people. I had a great conversation last week. I was at RBC's private tech conference out in L. A and I had this great conversation with Gokul Rajaram who just left Doordash. He is a prolific early stage investor. He's on the boards of Pinterest, Trade Desk and also Coinbase. And we are joined by Mike Gilroy who recently left CO2. Those, those two have invested together on numerous situations and Mike had a lot of great things to say along with Gokul about how they're thinking about the generative AI space and how they're thinking about more traditional tech before everyone was focused on gen AI and how it's changing the narrative. So that was a great conversation. It was live at that RBC tech conference. So stick around for that after Ann and I go through this. All right, and let's first break down a little bit of what you heard from OpenAI. I think a lot of this stuff was sort of previewed. We're, we're thinking about multimodal sort of, of models at the moment. It seems like voice is the rage here. What were your takeaways from what open AI? I think it's the GPT4. Oh, and it's not 4.0, it's 4O as the letter here. And this is, you know, a voice. I mean, again, what I perceived is that people were like, okay, this was what Siri promised us 10 years ago and finally OpenAI is doing it.

Yeah, I mean, the last 24 hours and we're still digesting this with all the Google announcements today are an absolute flat fire hose of new launches from both OpenAI and Google. And the interplay between the two is fascinating to watch. But just going back, you know, 12 hours ago when we were responding to OpenAI's news. Look, I'm very impressed with OpenAI for two reasons. One is they're continuing to deliver on their promise of building foundation models that are not only going to deliver more capabilities to the entire ecosystem. And we saw that with the multimodal launch yesterday. The ability to understand voice, voice and text and images, and to create instantaneous sort of translation between all of these modalities, right? Voice to speech and understanding what's in an image. Being able to essentially have AIs talk to each other, but also delivering on better sort of cost and performance. For their latest model, it's 50% cheaper than GPT4 Turbo. That's going to be important for a lot of startups that are building on top of OpenAI's models and where these cost reductions really make a difference in terms of building scalable experiences for consumers and businesses. But I have to say I was very impressed yesterday with OpenAI for two reasons. One is they clearly tried to steal Google's thunder and I think they did a phenomenal job. But two, I think they really brought a lot of humanity back into the conversation in and around what AI can do for us now. Clearly they have to find a way to continue to engage and bring back users to ChatGPT. And we know they potentially have been struggling with that recently. So I think the voice assistant is a way that they can give their users and consumers a reason to interact with ChatGPT more. But the way that they did it I thought was really clever. It's a lot of humanity that we're putting back into the conversation. It's showing people and inspiring the entire ecosystem, I think consumers more broadly about how AI can play a role in your life. So in some ways the voice assistant is the holy grail. But yesterday we saw a voice assistant be a teacher and you can see how it could be incredibly useful to have an AI tutor that can review your homework and coach you along the way. We saw their voice assistant be used as a social game, right? Social game between friends. We saw the voice assistant laugh at someone's dad jokes. Right?

Never happened before.

Basically never happened before. We're definitely going to need AIs to do that, but also bring more human expression into the communication. And so I think that they really are mirroring back for us the fact that we want AI to feel human, we want it to feel natural, we want it to feel intuitive, especially when we're using it in our personal lives. And they showed a lot of that yesterday and I think really leaned into the consumer experience versus just talking about the latest models and its performance and its utility for developers. So great move for them. And you know, part of me in the back of my head is also thinking about the strategic plays that they're making. And so perhaps this was a way of demonstrating to Apple that they really do have the ability to power Siri or power other core iOS functionality as rumor has it, you know, they might be considering a partnership.

Well, it's interesting, that was exactly my take last night on CNBC's Fast Money when I was kind of reviewing some of what I heard from the company and they demonstrated its functionality is like this is almost, you know, in two days a little bake off for you know, Gemini versus call it open AI. Which is kind of interesting because everyone's focused on this June 10th worldwide developers forum in Cupertino. And again, I have to assume that, you know, Apple is on the 1 inch line of, you know, licensing this technology. But they have to go a bit further than let's say Microsoft did. Because when I looked at a lot of the announcements from OpenAI, I see them really competitive with what Microsoft wants to use this technology for. The idea that they're going to create a search engine. The idea, I mean there's a whole host of things that they've come out with. We're like, oh wait, isn't that what Microsoft was hoping to build out, you know, capabilities with this technology? I think what's different about maybe Apple is that their investment in machine learning and cloud based and their own chip making capabilities and the like has really set the stage where I guess they never had to build these foundational models. Right. If you really think about it. Because that's what a lot of the bulls on Apple, the stock and the company are saying. And therefore maybe just a kind of a layer, a licensing sort of thing gives them the ability to kind of integrate it into their existing technology and then deploy it across an installed base of 2 billion iOS users over world thoughts on that? Again, I'm sure you don't spend a lot of time thinking about Apple other than all the products that you use from them, you know, in the public markets. We spent a lot of time talking about it because again, it's you know, the second largest market cap company in the world next to Microsoft.

Yeah, I'd never thought I'd find myself saying this, but I think Apple needs a growth story. I'm obviously a very happy user of Apple's products. It's an incredible company. They have all the resources and all the cash in the world to figure this out. But they are on the back foot right now and they should be worried because is the sort of core of this next wave. It is the machine intelligence, it is the foundation models combined with the ability to really interact with user Data and they seem to be out in the cold. Siri does not have these capabilities. It makes sense that they would be looking at both OpenAI and Gemini. It'll be interesting to see who they pick. But it's fascinating that Apple should be on the back foot and yet I think they might be one of the biggest beneficiaries of the bake off that's happening between Google and OpenAI. The interesting thing about Siri and if and when they sort of launch a better version of Siri powered by generative AI and we'll see who the partner is for, that is what happens to consumer data in that scenario. That's something I'm very interested to see and watch in that sort of partnership. Typically Apple has been the gateway for collecting consumer data, consumer information, that should give it a position of strength. But when that assistant is powered by a third party, whether it's Google or OpenAI or somebody else who ultimately collects and take advantage of that user data will be very, very telling. And I think I'm really excited to see how that kind of plays out in the industry. Then there's obviously meta. You know, they're also well positioned to build an AI assistant and you know, they've made announcements that they're working on that. But yeah, it does feel like a big bake off coming into the worldwide developer day that Apple is hosting in early June.

Just from like looking at it competitively, I would think that OpenAI would be much less competitive than, let's say, licensing Gemini. But again, Apple and Google have had this relationship on search and it seems to, you know, they haven't really overlapped a whole heck of an amount, despite the fact that obviously Google owns Android, which owns a good deal of smartphone market share around the world. And it just seems like that if open AI is open for business to license that technology, it puts a lot of these different players on a similar sort of playing field rather than having some of those competitive sort of issues. So to me, that's kind of my two cents there.

You know, Apple at some point has to start to diversify away from Google, right? They've had a longstanding relationship with Google. They earn billions of dollars every single year by using Google Search. But if I were Apple, I would worry about that dependency over time and I would take this moment to begin to diversify and really create some optionality to their Google relationship. And what a better way to do that than to begin to bring in OpenAI perplexity and other AI challengers right into their ecosystem and I think that's we're all watching to see is do they take this moment to start partnering more broadly across the ecosystem versus continuing to build an dependency on just Google?

Well, yeah, and they haven't done a great job of partnering over the years. And when you think about the relationship with Google, it's literally them paying for exclusive search on Safari on iOS devices, which is a very high margin, 20 or so billion. That just drops right down to the bottom line if you think about it for Apple. And when you think about the smartphone business is not a high gross margin business. Right. So as you know, upgrade cycles have get Elong your point about they need a new growth cycle. I just don't believe that there's an opportunity for them to grow because they say the next iPhone is called AI. It's going to be in the service sector where they get all of their margin for the most part. Right. And so your point is, is like how does. And people forget. So WWDC is a developers forum. So the idea is you get all the developers there, you, you show them what your new operating system looks like, you show them how it works with your hardware, you show them how it works within the ecosystem and spatial computing and wearables, and the list goes on and on. And then you say, okay, go develop, create the best things for our users. And then you know what, we're going to take a 30% tax on every app that is downloaded or paid for or service that are used for. I mean, that's how you get the growth going there. Right? And so to me, I'm really looking forward to see what they come out with. They have basically been at least signaling a great deal more than they ever do in front of events. You know, they just held this iPad event. They spent a lot of time talking about it. All right, before we get too bogged down on Apple, which I think we already have, let's talk about the controversy around this ad last week that came out with this iPad event, because I saw you tweeting about it and you were actually one of the only people I saw tweet something really, really interesting about it. And I just want to quickly mention that you know Ben Thompson, who writes strategically, I know you read him, most people who listen this pod read him. He made a really good point about this. I think he was in agreement that the ad was, you know, really off the mark when you think about what are the reasons that people are drawn to, you know, the Apple ecosystem. But he also made the really good point is one of the reasons why, I guess it's so controversial because it's so true.

Right.

When you think about what these devices, and it's not just Apple, they're just the ones that I guess Tone deftly said it. So I'd love to get your take what your initial impression was, what your impression has been reading a lot of people or talking to a lot of people that, you know, you know, have their finger on the pulse in this space as you do, and. And then talk to me a little bit about a little sleuthing that you did because it's funny, this is not the first time someone's tried to make this sort of argument about technology.

That ad. It really hits you in the heart there something about it. And I'm definitely not the only one. I think a lot of people had a similar reaction of. It does perfectly encapsulate this idea of kind of crushing the creative industry, right. And trying to cram it down into a single iPad, which kind of makes sense for Apple, but it was kind of the most dystopian ad I've ever seen from them, which is both, you know, I think emblematic of like, yeah, we are trying to cram the entire creative industry right onto some Apple application on the iPad. But the way that they did it just felt really dystopian. Like we're going to take all these amazing artifacts and devices that we've loved and that have played a role in our lives in the past. We're going to crush them and make them go away. And it just felt like a real departure from the type of advertising that we typically have seen from Apple in the past, which is deeply empowering, like, empowering to users, empowering to creatives. It's inspiring, it's playful and it's fun. And this felt like, okay, we're just crushing the creative industry and we know it. And so I think it really didn't resonate because it's not the. The type of Apple that we really expect to see. And there was a version of this ad that was done, I think, years ago. I forget, you know, the year in which it came out. 2008 or 30 years ago. 2000. Okay, a little more recent. 2008. That was basically the same thing by an Apple competitor, also device maker, that kind of played on the same theme, but, you know, just expect better from Apple.

Well, you know, it's interesting.

I got two things to say about that. Okay, so the 2008 LG ad. Think about where we were in mobile computing. Like, to me that ad felt actually very optimistic in a way. It was basically introducing the fact that this is all the things that you can do in this handheld device. And think about what percentage of mobile phone users had smartphones at the time. So it was basically saying, hey, come take a crack at this. We got an 8 megapixel camera on this thing. We have a flat touchscreen, you know, like all these sorts of things. So to me, it's just interesting how far we've come since then, how iterative some of these actual hardware devices are. And it's basically telling us that software, in my opinion, is leading the way. But I also want to mention it's interesting that you use the term dystopian 1984 Macintosh commercial. Very dystopian, right. When you think about it, meant to be very optimistic, but it had a very specific tone to it at the time. So in some ways they're almost going back to their Macintosh roots, roots from 40 years ago. And I had a much less negative feeling about this ad and I'll tell you why. I think a lot of the people who are likely to complain on their blogs or on Twitter or, you know what I mean, on their podcast are people that deem themselves to be or see themselves to be as very creative, right? But if you think about the masses, they'd love to be as creative as us or you or whoever, right? And this sort of thousand dollar device levels the playing field for them. They can make digital music, they can do, you know, like create art or experiment with an AI model in a way that maybe they're just not equipped to do on their own to figure this stuff out. So to me, I actually have a very different take on it. I love the ad and I thought it was amazing. And I think we're going to look back and be like, what a bunch of whiny, you know, what's on Twitter, as usual. And it's usually coming from Silicon Valley or the like. And again, this is, you know, me putting my Wall street hat on. So there's my little rant. But again, I just think it's so funny.

The question is, who is providing the inspiration today? Because I agree with you, everyone wants to be creative, Democratize creative tooling is one of the biggest opportunities that we have with generative AI. But that inspiration today, it is not coming from Apple, it's coming from startups. And you know, it's coming from OpenAI and Google too. But look at what we've seen in the last couple of weeks with Suno Udio and these other startups that are doing synthetic music. Music generation, right. The idea that anyone can now compose a song, create a song, share it with a friend like that is going to unleash so much creativity. Music video generation, right. The ability to create images with generative AI like that I think really speaks to what will excite consumers and also the potential that we have to democratize creative tooling and make it accessible for everyone. That's not in the iPad moment today. I think that the interesting thing with Apple is what is that next S Curve? It's not going to be the iPad. We already have the iPad. I think the unanswered question that Apple has to speak to is what does that next device look like that will feel magical and new and relevant for this future of both generative AI and AI assistants? We have not seen that from that.

Well, interestingly, they have released that product that I think they're betting the next 10 years on, and that's the Vision Pro, right? So to them Spatial computing is that thing. And let's also remember at the time, you know, when they introduced the iPhone in 2007, I mean Nokia and BlackBerry were behemoth. Motorola still had that cool Razer phone, right? And it was, the adoption of that was really slow. And I'm not making a bull case for Vision Pro, but if spatial computing is going to be a thing, I have to suspect that Apple is going to be at the forefront of it also. So when you think about this AirPods business that they've created, it's a larger revenue business than like I think 90% of the S&P 500s, you know, revenues in a way. And so to me, what, what basically the killer app is for Apple is that ecosystem. The likelihood that they lose meaningful market share to another hardware company because they're using some other app based sort of thing. If it's not an operating system, I don't think that's going to be the case. And think about, about this. You know, Facebook tried to do it over the years, whole host of others. I mean Android, while they have tremendous market share around the world, the average selling price of the devices that they're on, it's like less than half of what Apples are. And Apple has like 80% of the gross margin in smartphones, you know, all over the world. So to me it's not an easy one. If anything, I think Microsoft and the perception about what they've been able to do in the last year and a half with OpenAI's tech, I think that's the thing that ultimately could disappoint in my opinion, if we're just kind of playing stock market, if you will, a little bit here. But again, you and I, we're gonna, we're gonna debate this I think for years to come here. Anything out of Google and I know that it's still going on right now as we're recording. That made you a bit optimistic relative to expectations, relative to some of the other launches that they've had over the last year and a half or so, because again they've been this perceived laggard. You know, it was initially that Bard launch, I think in early 2023. You and I talked about it at the time, then they renamed it and it's Gemini Advanced or Pro or this and that, whatever. Pretty uninspired name. Know these are free products that they're introducing across. They have seven or eight platforms or whatever you want to call them that have over a billion users, you know, like three or four with 2 billion users. I mean if you can enhance that productivity and make those products that much better, you make that ecosystem, you know, that much stickier, I guess.

Well, when it comes to Google, I think today we're seeing them make the obvious move that makes so much sense for Google which is embedded their AI capabilities, task agents, you know, the ability to summarize, et cetera, directly into the products where they have a massive install base. I'm a huge Google user myself, obviously love Gmail. They finally are starting to talk about embedding AI in Gmail. You know, in some ways these are sort of banal AI efficiencies, but these are the things we're going to use every single day that makes sense to us as utilities. Hey yeah, summarize all the emails from my kids school and by the way, like give me a summary and then write a report, apply for me. Like this is the stuff that consumers are going to use every single day. And it makes sense for Google to essentially abstract AI into a set of tasks and assistance and automations within their existing products. And I'm actually very optimistic that that will generate a lot of growth for them, as it should. I'm a little more skeptical about Google really sort of creating entirely new products with these capabilities. And I think the theme of what we're seeing today is just making sure that they're distributing their AI Capab capabilities across all of their best properties. Search, Gmail and Workspace. Yeah.

All right, let's go to Perplexity really quickly here. It was an announcement Last week I saw you tweeting about it. I read the press release. It sounded really interesting to me. And again, I think the theme in Generative AI over the last week has been voice, and I think we're going to hear a lot more about it as we get to WWDC in June 10th. And Siri has to be a focal point of this strategy, if you think about it. But the headline here was Soundhound AI and and Perplexity Partner to bring online LLMs to next gen voice assistance across cars and IoT devices. Now, IoT was hot a few years ago. I'm sure there was some investments that you were making in and around that ecosystem is supposed to be the next big thing or so that is the Internet of things. Talk to me about this announcement and what your takeaway was and really breaking it out. You know, I guess juxtapose some of the things that you've heard from OpenAI and Google over the last 24 hours.

Yeah, I mean, Perplexity is moving fast past experimenting with a lot of different product features and partnerships. This is the latest example. I think the future that we're moving toward is that we want intelligent conversational AI assistant to be available to us in any setting, in any context. And I think the Soundhound partnership is an interesting way to sort of introduce that into the in car experience. You're driving, you definitely shouldn't be on your phone. It's incredibly dangerous. If you have a question or you want to have a conversation, you can do that with the perplexity and soundhound APIs and AI assistants. I think the bigger picture here though is that we have a lot of work to do. And I think this is part of the opportunity for Perplexity to again make all of these new hardware devices in and around AI and AI assistance truly useful. And the killer use case time and time again is information search. Right? You want a recipe, you want a list of the top restaurants, you want an answer to some very specific question. Perplexity can give that to you in a really succinct way. And so I'm really excited about the company continuing to embed Perplexity Search and Perplexity's AI Assistant essentially into all of the different consumer touchpoints that exist. And I think we'll see more experimentation from them on that front. I think I'm slightly less bullish on the Vision Pro. Not to go back to Apple, but I'm not sure the future is all of us walking around with a giant headset on Our faces. I'm much more excited about something affordable. Lightweight earbuds, you know, maybe with wearable glasses. Some combination that is so affordable and so easy to use that, you know, it's available to a billion people. Not someone who can pay $5,000 for a vision.

Yeah, and I agree with that 100%. Especially for a consumer application. I think that's probably first adopted in the enterprise, especially at that sort of price point. And I could see a whole host of use cases for it. And you're already starting to see some develop, you know, for training modules. And I mean the list goes kind of on and on, I think. Consumer device. I agree with you. You know, there was a story out earlier in the week. Meta explores AI assisted earphones with cameras. That is coming from the information. We'll throw that in the show notes. And again, you know, I go back to Google Glass from at least 10 years ago. I mean for all intents and purposes, that was laughed off the tech scene, but that is the sort of device that looks somewhat unobtrusive, had a little camera on it, could probably do a whole host of things that you know, you know what I mean? Like that right now, like think about what you could jam on such something, you know, that size, it's more of an augmented reality sort of glass. But that might have been an easier step.

But you just have to think about.

How Apple operates in general. You know, like to me the Apple watch, it would have made much more sense to look like a whoop or something that was really healthcare focused at a much lower price point. Get much like mass adoption, make it a very easy companion to your iPhone rather than introducing a 5, $600 device now as high as like a thousand, you know, with the Ultra. Etc. So when you think about some of the other things, there's been this humane pin, this rabbit I think that you got. You were doing a unboxing. I haven't seen an unboxing this decade. And so good on you for doing that with the rabbit.

And then when you read stuff about.

Meta, you know, Oculus has not been a home run for them. I know that that's a big part of their Metaverse strategy and what they want to do with generative AI. How are you thinking about wearables in the space? And are you guys, are you looking at investments in the space as it relates to AI power powered devices?

Hardware is hard. Most investors know that. But look, I think I try to approach everything as you know, from an optimistic mindset. I think it's Easy to be cynical and say these things can never work until they do. And so I think that I'm very open to experimenting with different consumer hardware. I try to buy and try anything new that comes to market. There are exceptional teams trying to build wearable glasses. The Rabbit is a really awesome play on, like a retro device device. You know, we've tried the humane pin as well. I mean, it's a. It's a stellar team. And so I think all of these early attempts matter because putting these new AI wearables and these new hardware formats into the hands of consumers is how we're going to understand, like, what's truly useful, what's viable. And I don't think we've seen it yet. So I think as an investor, I'm very sort of cautious about consumer hardware. I think consumer hardware really needs a killer use case, and we really haven't seen that outside of something like Perplexity Search or ChatGPT's AI assistant. So I think that the ultimate form factor that could win here, I could be wrong. I'll be looking back a decade from now and say, like, I really didn't see it coming, but I think it will be something simple and intuitive. And that can also be a fashion statement. Like, no one wants boring tech to put on their body. And so I think eventually we'll move toward a world of, you know, maybe it'll be earbuds and glasses, maybe it'll be a pendant that you wear, but it'll be something where the technology and the fashion statement or the identity statement are kind of fused together in an interesting new way. And so we're proud to wear these devices. They're a statement of who we are and how we express ourselves. And, yeah, they'll also be carrying around a really useful AI assistant.

Let me ask you this, is this a useful AI assistant Grok? I'm not really. I don't use. I don't. Well, no, I don't hear anybody talking about it, but. But, like, I don't use Twitter much anymore. I retweet stuff. I don't use it as like an RSS feed the way I used to. I don't really chat with people. I don't DM anymore. I feel like Elon's reading them when I think about this. And by the way, you know, I was given a premium service. I didn't want it. I didn't want the blue check. I took the blue check off it somehow a week or two later popped on there. But then when I go There, I see there's some Gen AI summaries of content and stuff, and I don't trust it, to be honest with you. So that goes back to trust is really important. And it's funny, like the way that, that Elon Musk is playing in this space. It seems like he's like a lover scorned or something like that. You know, like as we think about his relationship with OpenAI and the way and the speed in which he's focused on creating AI and using Twitter to do it, it just, to me, I don't trust it. But I'm just curious, what are people saying about it? You're still using Twitter. I'm sure a lot of your colleagues and friends are using it. I think engagement's down dramatically. I know I just go on and I look and see, you know, my feed every day, once in a while, a couple times a week. I'm just curious what, what, what people are saying about the product of this because it seems to be left out of a lot of these discussions of some of these Gen AI capabilities as it relates to summarizing content and giving us answers.

I don't have much to say and perhaps that's telling because I don't use Groq and I don't hear anyone talking about it. And you know, I think data quality is a really important consideration in training LLMs and ultimately the output. And, you know, I think an open question for me is sort of what kind of data it was TR and what I trust its ability to synthesize information on my behalf. You know, I love Twitter. I'm still on there. It's mostly for the water cooler conversation that you and I are part of in the tech community and then the startup community. But, you know, it isn't sort of the trusted source of information. And so I think I don't know where Grok is going to go. I'm sure they have a talented team, but I don't think it's breaking through as either a model or a consumer product at this point.

All right, well, you know where I think Perplexity should go? Said this to you a couple. Oh, I said it on CNBC's Fast Money two months ago. I thought Apple, when they were like not talking about Gen AI at all, I thought that they should make, you know, a $2 billion bid and Aqua higher, if you will. And then by the time you and I and Arvin had our conversation, it was a $3 billion bid that Apple should make. And now it feels like something four to five. But again, it's a rounding number for Apple if, if you think about it, or rounding error if you will. But you know, it just seems, seems like that product on an installed base like that would catch on like wildfire and it could quickly become a verb. We might need to change the name of it for become a verb, but.

We need to shorten it a bit.

Yeah, no, but again, congrats to, you know, like everything that's going on there. And I only say that from like a product user, you know what I mean? Because I just see it getting better. I started listening to Discover Daily, which is their daily pod, which feels like really efficient, you know. Know it's just easy at five minutes or so and it's uniform and you know, I know what to expect and I'm hearing some new stuff and some of the other advancements I'm seeing. So congrats to the team. Congrats for you. I know that you and your team are helping them along too. But Amber, Decky, I really appreciate you coming back. I love the conversation. I hope we can do it again soon.

Looking forward to it. Thanks for having me on, Dan.

All right, thanks. All right, stick around again for my conversation from RBC's private tech conference with Goko Rock Rajaram and Michael Gilroy.

I have Gokul Rajaman over here and he recently left Doordash, but he is a prolific early stage investor. And Mike Gilroy here who recently left COTU where he was the co head of growth and he was the fintech lead these guys have invested in together seven times, is that correct? And we're going to talk a little bit about that relationship, what it's like the co investing thing, but we also want to talk about where we think the kind of private tech market is right now. Gokul, you mentioned that a bunch of your companies in your portfolio were just up at rsa. You had a funny tweet. And again, as a podcaster, all you really have to do is look at your subjects, Twitter and you have all the content that you need. Talk to me a little bit about that tweet because you basically said it just, it was a montage of just AI stuff there. And that I have a feeling was not really the case a year ago at rsa. So talk to me a little bit about how much time you're spending on generative AI, just the companies that you're working with right now.

Thank you, Dan for having me and great to be here. The RSA collage that I posted on Twitter was actually from a company portfolio company that was there. They sent a picture of all the pics they took across the booths. About 80% of the booths had AI platform for X or Y. And then someone commented that it's true for every other company conference this year and last year also. Almost, I think at this point in the venture ecosystem it's almost impossible to get funded or to get talked about if you don't have an AI story. That said, I do firmly believe that AI is a tool. And just like we don't talk about mobile, many of us are old enough to remember 15 years ago there was an I fund and all of those things. We are not going to say AI is just going to be an important tool to enable companies to achieve business goals. And so I think it's just a moment in time we're talking about AI, but the reality, reality is the power of all of us investors, all of us who are investment operators have been looking for the next big platform for the several years, last several years. And I think finally we might have found something that is so profound in its implications. And I like to joke that every other revolution has displaced blue collar workers. This is the first one that's actively going to displace all of us white collar workers. If any one of us is generating any kind of content or sitting in front of a keyboard, most likely, likely AI is coming for us. January is coming for us, starting with SDRs and low level engineers first and customer support reps. But then going up.

Yeah, but that's what Apple is trying to tell us this week, right Mike?

Of the 80%, how many of them are actual AI? Curious what you think that number is today.

Yeah, I think the challenge is that the moats within AI are low. So on almost every category you look at a company, you'll see there's a dozen or more companies being started around the space. For example, example two days or last week I met a company that was doing an AI sdr. SDR are basically the most junior salespeople, sales development reps, they set up appointments, they call companies. I was like that's a great value Prop because for $1,000 a month they are essentially giving you the work of two SDRs. There's about 10,000 tasks they say. So in general it actually is very compelling and if you want to hire more of these AI agents, you can pay more thousand dollar things. But then I did a Google search, turns out there's 15 of these things that are offering AI SDR. So I think a lot. The answer is I think we're still trying to figure out what the motives and what the durability is of these companies and business models.

Yeah. So the fear would be that they're going to quickly become commoditized.

Right.

And then it's just the stuff that's built on top of that. Mike, do you want to talk to that a little bit?

I would say commoditized if you're the app providing it. But we're co invested in a company called Cloudwalk and 12 months ago, the CEO Luis There is always ahead of the trend. He was the first fintech company that we knew of that had stablecoins on the back end, but doesn't talk about it. And he has AI running something like 75% of his customer success. But they built it internally. Right. And so it's not only the bigger platforms that are going to come in and take it, it's people having the ability and technical talent internally to go in and take it themselves.

Two most commonly used applications of AI within the enterprise, customer support and engineering. That's where everyone is going. Every public company, customer support. You can, just like Luis said at Cloudwalk, you can dramatically bring your cost of customer support down with AI agents and then engineering copilots, easily increasing productivity or 20 to 30%.

So we just saw this, I guess that they literally their chatbot from a customer service perspective allowed them to cut 700 customer service reps. Mike, you've spent a lot of time in the fintech space here. Obviously customer service should be a high touch sort of thing when it comes to your money, Right. As a consumer, how are you thinking, thinking about that? Is that number something that we're going to see just take off a little bit or are these kind of customer service reps offshored, which a lot of us consumers don't like anyway, and it's going to ultimately be a better service right out of the gate.

So I think you have to look at categories. Right. If you are talking to enterprise type customers with high value, you probably are not going to lean on AI completely. Right. If you're talking to consumers that missed a payment on a $10,000 loan, or you're talking to SMBs and they're having trouble with the software phenomenal use case. Right. And Klarna I think fits right in this. I think the thing that we talk about a lot is the cost reduction in opex from cogs, from cs. But the revenue improvement has also been pretty dramatic in some of these companies to varying degrees based on how many products they can cross sell or have the ability to cross sell that they've built. And so I think you're going to start seeing it on both sides of the P and L pretty dramatically.

Engineering is right after customer support, the number one application. And so almost every company that I know of that has deployed something like a GitHub copilot to their engineers has seen a 20 to 30% productivity increase. But I think we are now seeing AI copilots or replacements for every function. There's finance copilots or replacements, there's like I said, the sales ones. And so you're going to see AI agents essentially permeate every functional area. And like Mike said, it's about cost reduction on one side, but it's also about improving efficiency, output and revenue. Really that matters more.

Yeah. And Mike, are there other use cases that like, like people here can just say yes, these are things that you should be focused on from an investment standpoint. If you are private market investor, are there some places that you should be going to right now that are likely to be able to monetize?

Yeah, I think one that I'm very excited about. But what I haven't seen break out anywhere yet is underwriting in particular in SMB. If you look about the lending regulations all the world in the US we have something called fair lending which fundamentally limits the data you can use and the analysis you can use on lending that does not exist for businesses. And so at some point we can plug in probably low value loans and have an answer come back out based on new data sets and obviously the analysis around it. And then in particular because of the lending laws here, if you look to other countries with big lending markets, namely Brazil for example, I think there's a massive opportunity that there. But because of the regulation around it, it's going to take a while. But that's something that I think is just slowly sneaking forward here.

How far ahead is the US and what markets are focused on? Because we know that there's a lot of copycats. If you just look at all these other major technological shifts, I'd say in.

This case it's not just, I would say the US but it's San Francisco in particular. I've had enough companies now. They just couldn't find enough talent when they were in New York or Boston or so on. And they moved, the founders moved to San Francisco, Cisco to find the talent. Right now I think it is going to disperse. Just like with the World Wide Web 20, 30 years ago, it slowly Dispersed there were a couple of Nexus hubs but just the quantity of talent trained in AI at Meta, at Google etc is just so high here in the Bay area that it's just very hard to build an AI company without having a research center in the Bay area.

So again OpenAI is $85 billion value situation of a private company. But if you think about the arms rates that exist right now, it's between Microsoft, Google, Amazon, Meta. You know, Apple's going to get in the game here. They're going to be doing different stuff I think than the companies we just mentioned. But when you think about the capital from a strategic standpoint that they're investing in some of these companies. Mike, there's a handful of Chinese companies that have the resources and the political will I think is also really important if you think about it from that perspective.

Perspective.

But I can't see how anyone else is really going to be able to compete with 10 massive companies right now.

This is a very unique trend. Whereas you go back to mobile and cloud really everybody benefited, right? Because of the infrastructure any company can go out and get it. This specific technology favors two fundamental categories of companies. One, the largest data sets and two, the largest balance sheets. Right. The capex output is enormous as we all know here. As an early stage investor and even early growth investors investor you have to be very careful. Whereas you go back to cloud, you can rip in place and say hey xyz, large market cap company is doing this. This is going to be a lot better with this series A company. But today we're looking at it like man there's there are a lot of line mines out there and these incumbents that you talked about are fundamentally advantaged in a bit.

But as an early stage investor are you guys finding opportunities in generative AI?

Because you just think about this like.

A company that a year ago did perplexity, right? And then they've done basically two rounds and this year already the valuation skipped from you know, 100 million to a half a billion to two and it just seems like it's being crowded out for early stage investors. Gokul, how are you thinking about that? Are you finding opportunities in the private markets?

Yeah, you got to pick and choose your spots as an early stage investor and I think one of the most interesting areas is in verticals where there's a lot of unstructured data where essentially there's no. I always think of legacy companies as folk companies which are more than 10 years old. So most of the players, the alternative either is no consumption. There's no one serving them or legacy players. I'll give you an exchange example. If you look at personal injury lawyers who are also called ambulance chasers pejoratively, they basically have to issue what is called a demand letter when there's an accident for their clients, they have to send a demand letter to the company saying hey, here's insurance information, this is the accident info, etc. It took them weeks to pull together this from the data of the accident and send it. There's a company I invested in called Evenup a few years ago that essentially started doing this for these personal injury law starting in la, which is the center of personal injury law. And they basically charge. But what is interesting with all these companies, and this is why I think companies are going to customers are going to benefit, they charge not based on hours of legal time, but based on the outcome which is based on the number of demand that is generated. So literally in this case they are going to these persons and they're replacing tax paralegals at scale. So instead of hiring five paralegals, you're basically just using this agent that generates demand letters for you. And I think one of the most interesting concepts that comes with all of this is pricing based on outcomes versus based on hours of work. And I think that's a fundamental shift to work as it stands today. But I think verticals, healthcare, legal, finance and education, four or five, a lot of unstructured data which we could never access before or do anything with earlier entertainment, of course.

Yeah.

And again you look at the arrow growth, some of these companies, it's some of the best we've ever seen. And I think the key thing to not miss is this business model is fundamentally different because the cogs are much higher. Whether that's from the compute or if you're one of these big models, you have to pay for all the data that you're getting. This is obviously a trend that we're reading in the news quite frequently. And so as we think about valuing these companies companies at the later stage and as they get public, we're going to have to think about the gross profit to EBIT conversion very differently than we did for a pure SaaS company prior to this technology.

That's very true. But I think the SaaS, right, SaaS basically was software as a service. You're increasing going to hear this term service as a software, which is basically the reverse of it, which is basically saying you want a demand that is generated. Now you can call an API call and it gives you a demand letter. You want a sales call made to someone, you call AI to an API and better made. So services are going to be offered as software. And I think just like SaaS opened up prior hosted enterprise software, you had to have a data center 20 years ago to use software. With SaaS you can use it in the cloud. Many startups can't really hire a sales rep easily because it costs $120,000. You can't hire 10 sales reps. But with these agents, you can actually now hire 10 sales reps with just making an API call. I think it's going to really democratize. And so the opportunity I would argue for these companies might be much larger. Larger because they can open up these services to many more people. You and I can probably use a sales tip for $1,000 to do something.

You just made a joke. You said entertainment and you're pointing to me as a podcaster. You're basically saying that you're going to be out of a job very soon. And so I want to ask the question about at least the things that we're interacting with right now as consumers. These chatbots, for instance. And again, you can have pay 20 bucks a month and you get. So it's replacing a service that we've had for 20 plus years searching for the web for free. And the numbers are not staggering when you look at like how many people are paying for these things. But I'll tell you this. I had the Perplexity CEO on the pod a month ago or so and I put into Perplexity what are some good questions on a podcast to ask the CEO and they were really good. So talk to me about that. Are we just so in the weeds that we are the ones trying all these things? Gemini Ultra and OpenAI and Perplexity, is there going to be mass adoption of those sorts of products?

You think that's usually how it goes though, right? I would say even with Uber. This was a San Francisco phenomena that started with largely the tech and finance community and sf and it kind of grew from there. I think the level of engagement with Perplexity, I also use it very frequently. It's absolutely going to get to the masses. And in my point of view, I also just think about, as we think about Search and Perplexity, they are two different use cases today. Google's not asleep, by the way. It's a little nuanced today. It'll be most interesting to see how they collapse over time.

Yeah, so Gemini is getting better. They had a rough Launch right barred into Gemini and the like. Mike just mentioned their different use cases. I'm sure Gokul, you're messing around with these sorts of things. I think it's interesting that a rapper an LLM wrapper has gained what I can tell, at least in the tech community and the finance community, a lot more attention than, let's say GPT4 right now. Because you can have cloud, you could have GPT4, you could use any of these large languages, language models. How are you thinking about that? Because that's fairly innovative in a very short period of time.

I think it shows that what wins is not maybe the underlying technology because increasingly technology will be accessible. It's about how you package it and really solve a very specific problem for users. And I think of these tools as almost having an assistant at my fingertips. And when you reframe it that way you're like $20 a month for an assistant who can do so many things. Over the last few months I've started using with GPT4 and Cloud you can start uploading spreadsheets sheets to it and have it do analysis. It's pretty amazing. I was doing a retention analysis today and I was just uploading the raw numbers and asking you to do the cohort retention analysis. It was amazing. It just did it.

You spent a lot of time, you know, the Google culture, they've gotten a lot of criticism over the last year or so within the tech community from Googlers, the finance community. The stock was really lagging in the public markets relative to, let's say Microsoft for the perceived advantage that they had with the licensing of OpenAI. You think about Perplexity as an answer engine. That's the way that, that they describe it. Do you think it has the potential to take a bite out of some of that search market share?

I think it might expand the market share, it might expand people the kinds of things people use search for. But I have to posit, I'm very convinced that Google is going to be one of the top two or three winners ten years from now. I think they have the smartest people. A lot of things they move slower, I should say, than a nimble startup, but the resources they have, the moats they have built with partnerships, with the fact that they have the Chrome browser, the browser views, all of these things owned by Google. I think they are generally a benign company. If you think about Perplexity, Perplexity does use the Google APIs for search. If Google were an evil company, they would first thing have said, let's cut off your APIs and then see what you do. But the fact that they allow a competitor or pseudo competitor to use their APIs shows that they generally I think they care about increasing the size of the search pie. But I don't think perplexity is going to really take a bite. I think it's going to expand the market. You're a user. What do you think? Michael?

I completely agree and right now we're talking about perplexity in Google and the research process in Google versus finding an answer and perplexity. But I think what we're not really talking about is the perplexity use case where I would have called Gokul for an answer. Right. Completely offline. Right? There's answers that other people and you do this for your podcast questions, right? I think in general you end up calling three or four people who know this person in the business, hey, what types of things should I ask this person? And so I completely agree on the market expansion standpoint. And it's not just online, it's offline as well.

What do you guys think? Some of the early beneficiaries obviously, so if you produce the GPUs, you have all the supply and there's demand. That's 2 or 3x of what you can build. So we get it. Nvidia has been a massive beneficiary. This year alone it's gained a trillion dollars in market cap. There seems to be a waning interest in the names that might have some sort of solution. If you think about the hyperscalers have been a huge beneficiary. We saw that in the results. But it was very interesting to me that the narrative around around Meta was that they are using this technology for their business. They were one of the first to really get benefit of it. But the minute that they reported a 2 billion greater than expected spend on basically GPUs for all intensive purposes and then revenue growth was slightly decelerating from what people expected. The Stock was down 15% like that. This is a one and a half trillion dollar market cap company. So I'm curious, how are you guys thinking? Do the public markets have a good beat on this the way you guys are thinking of it out here talking to operators, talking to how they're thinking about harnessing this technology, spending on it, that sort of thing? Or is there another lens that we should be looking through?

Look, without going into individual names, I would say public markets today are still, we're not completely through this dearth of 22 public market investors still want earnings fundamentally. Right. And so the reaction is that I think the promise is so still there for all of these names. But when you come out and capex is multiples higher than what people were expecting, the stock is going to get hit, period. Right. And so, you know, even though it feels a lot better, there's still a lens of cost cutting, you know, across public markets.

And on Meta, I think people, at some point five years ago, people would have said meta has the lowest moats around the company. But I would say the last five years have shown us that their moats are actually maybe one of the most strongest of any of the large companies. Why?

Because.

Because they have identity and they have cross platform identity and the family of apps they have used for communication throughout the world is unparalleled because I think Android and iOS have established themselves as two OSs. And the way you bridge is with the meta family of apps where you don't have to worry about whether you show us a green bubble or a blue bubble, but use WhatsApp to communicate transcending that. And I think WhatsApp now is emerging as a really massive revenue driver for them, which is under reported. And I think it's a very volatile market. I feel these stocks are being almost like not penny stocks, but when companies are losing 15% of the market cap, it just seems like very short term.

I think part of it is they become very crowded trades. Right. There's a scarce set of names that you can invest in that are going to be able to deploy this in a meaningful way. And it's not in 2024 really, it's. You got to start now thinking about 2025 and 2026 and the way we went through this when we looked at the investments in cloud or any digital transformation right. Over the last kind of 15 years ago. All right, let's switch gears a little bit. So how are you guys thinking about how will this make it to the public markets? What does the roadmap look like over the next couple of years? If you were here at this conference, throughout the day, it seems like there's some optimism. There's massive pent up demand by companies who want to come public. There's massive pent up demand from investors, investors in public market, for new stories because of the crowding in some of these other names. Right. So the companies that you guys are invested in, are they starting to get interested in coming to the public markets and what would be their apprehension? I guess maybe. Mike, I'll start with you.

It's Growth and earnings fundamentally. Right. Improving that operating leverage over time. And certainly once you're a public company, there's a lot of talk about. We were just talking about it offline here about the scale. Right. And if you look back at 2021, many companies went public way too early. David Einhorn had a great letter this quarter talking about, I think he said the public markets are fundamentally broken. There's so much passive money going after these mega cap companies. Being a value investor in stuff that has been picked over but is too small, you can't make money in anymore. And so I think there is some truth to needing real scale to going public. And the reality is, I think to get new names public right now based on where they're at from a profitability side standpoint, that's going to be tough. I think we have some time.

I think durability of growth is one of the most important things that decides value in the long term. And scale is the other thing. And I have fortunately many portfolio companies in that 60 to $100 million revenue range that two years ago would have thought of going public. But right now they're growing at 15 or 20% at 100 million. And really PE might be the right answer. And there's not a coincidence that Thomas Bravo has raised a $34 billion fund and then some of the other folks have raised $10 billion funds. I think you're going to see a of lot lot of PE exits.

We co invested in a company, they're $550 million top line, growing 40% wildly profitable CEOs not even thinking about going public today. It's just not even on his mind. He's going to sit here and wait this out and continue to get scale. And by the way, if you're profitable, you're in complete control of your own destiny. And it's viewed as a complete distraction today if you don't need liquidity and you don't have investors that are pushing you to do it.

Have, have you guys gotten a lot more patient or, or some of your peers about this? Because again this private for longer.

It seems to be a theme.

Tell you what, watching these companies go public too early and just get absolutely demolished in the public markets has instilled a lot of patience and all of us.

Yeah. What about you Coco? I think I'm an LP in many venture funds and I am seeing a preponderance of letters asking for continuity. So they basically typically have 10 year funds. And as they're seeing the portfolio companies stay longer, they're either doing continuity funds or basically just asking for a two or three year extension tension. So I think people are still probably, I think no longer is 24 considered. We're going to see a ton of public companies, I think probably 25 here.

I'm going to help my banker friends.

Out in the audience here a little bit.

Listen, the Nasdaq's in all time highs people. And if you think about over the last 27 years or so, it's been pretty cyclical, the downturns, right? So if you were Amazon in, in the late 99 and you guys were VCs and you saw this thing come public maybe too early, right? And there was a lot of hype around it, you could have made the bet that this company is going to transform retail, it's going to transform a whole host of other things. But from its highs in 1999 or early 2000, the stock lost 85%. Now you may say, well that's an outlier. But I'll tell you that in November 2021, when Mark Zuckerberg went out on that analyst meeting or that earnings call and made that big proclamation about changing the focus of the company and changing the ticker and the name, that stock sold off 75%. That was a trillion dollar market cap that had real earnings, a real durable business. And then I could go into the financial crisis where lots of tech companies got hit as hard as a lot of the financial companies too. So I think like the idea of getting cut in half in the public markets, you better get used to it no matter what. And so I guess my question also is at some point, if you're stripe, if you are SpaceX, you're going to have to get there. So Mike, how are you thinking about that?

I think the difference between these two periods is today Nasdaq's at 27 times earnings, right? And so these private companies aren't necessarily looking at that number and saying, man, this is a great time and particularly relative to the last growth round that I did, so why go right? And not to talk about these two companies, but if you don't have earnings or near term visibility into earnings where you feel good about, you know, a Nasdaq's on average growing mid teens percentages, these companies are growing much faster, but it's just not a great multiple setup to be entering the market.

It's interesting that you mentioned that because again, the growth is coming from those large companies that are a disproportionate amount of the weight of the S&P 500 in the NASDAQ. But Gogol, you, you might have some thoughts on this. If you look at a lot of these SaaS names in the public markets, they are trading very poorly this year. There's a lot of dispersion and I would almost categorize some of those names as AI adjacent. When Adobe a snowflake, ServiceNow or Workday is it just so much focus on the names that are making the investments right now and have the products and are able to tell the story. What's going on with that disparity?

I think there was a recent analysis that someone forwarded to me actually yesterday where it showed that there is a almost a barbell effect where the large cap SaaS companies are getting a much higher multiple to small cap SaaS companies. Everything else being equal, they have the same growth and profitability profile. But the multiple that the market is ascribing to the large caps is much higher than to a small cap. And I think the reason is twofold. One is the profit pool assumption that the large cap has more durability in their growth and the second is they're going after much larger profit pool. You're right. I think the thing is these small cap companies are what are going to be large cap. Look at HubSpot, right? They went public 10 years ago, what was it a billion dollar market cap then and now they are 30 billion. They just have to prove themselves multiple earning cycles and earn the trust of investors. Because the reality is now there's what, 100 plus public SaaS companies. There's a lot to choose for.

Yeah, I, I completely agree. I would also go back to market structure and all the passive capital and these ETFs that are only going to pick off the mega cap companies, right? So they fundamentally have demand on the stock that these small cap companies don't.

Yet. What do you guys think? There was some discussion of this earlier today. Again the strategic investment in some of these name like at Amazon and Google tripping over each other to invest in anthropic. Right? And you're seeing this valuation go much higher. They're not focused on value valuation. Right. How are you guys thinking about that as venture investors if you're a later stage? Because these are companies that obviously are not going to be profitable for a very long time. The revenue bases are very low right now. They have the potential to obviously really hockey stick. How do these companies grow into these valuations? Because we've never seen this early, this high, this quickly.

One thing to note to remember Wait.

Wait, Mike, why are you laughing?

My. What I wanted to say was hopes and prayers.

Yeah, all right.

Substantive answer.

Yeah.

No, but remember, a lot of the investment, a huge chunk is in the form of compute credits. Right. So you're basically just, I shouldn't say round tripping, but you are getting that.

Who did that 25 years ago? Sun Microsystems. Sun Microsystems was the Nvidia of that time. They probably weren't getting equity in exchange for it. But at some point in the year 2000, somebody jammed one too many banner ads on a website and the whole thing came crashing down, you know what I'm saying? So I worry about that a little bit because Nvidia has been so successful, they have no competition. And so their ability to invest in the ecosystem, it makes a ton of sense. But the fact of the matter is, and I don't blame the company, I blame investors that they've added a trillion dollars in market cap to this company that is only going to see deceleration in their earnings in sales and their margins. Because once the competitors come, once this stuff becomes more commoditized, margins are kind of come down pretty fast. So thoughts on that, Mike, is I know that you have a good sense also for the public market.

Yeah. And you know, and it's, it's hard to look into the future because even 18 months ago we had no idea this wave was going to come. But to give Nvidia credit, prior to this they had a big run up against intel because they were supporting bitcoin mining. Right. And so they are really thinking forward on every single trend. Maybe there's something around the corner. I don't know. I tend to agree with you more broadly just in terms of the trend and it potentially being overheated in the short term, valuation wise. We don't know. Maybe there's something else coming.

Listen, if you just look at the company over the last five years, it was data center gaming, bitcoin mining, and now generative autonomous driving, autonomous drive. I mean, like so. And again, the stock was very volatile when it started to decelerate on each one of those trends. That's the one, one thing that I would say. All right, we only have a few more minutes here, but I'd love to get from each one of you something that you believe. It could be private, public, about tech, it could be about anything, actually. Something that you guys believe that you think consensus has not come around to yet. As you think about investing, as you think about the advice that you give companies that you're invested in. Is there something that you want to share with this audience?

Usually we get time to prep for questions like this. We, we didn't, we were talking about this earlier, Gokul and I. One, one major thing we talk about these software companies that have been beat up pretty dramatically in the public markets and many of them today are one act and Gokul and I have invested at the intersection of financial services and software quite a bit. There's still a lot of learning particularly in the public markets around what some of these names can do and cross selling payments in particular. And so that's one trend that we're very focused on. I don't know how controversial that is but I think it's something a lot of people aren't baking in.

Yeah, for me it is something that Kotu and Mike published a few years ago which said financial services are 25% of the GDP of every country in the world. If you look at the fintech is one of the most beaten down sectors. It's the opposite of AI over the last year. And so it is crazy to me that it's beaten down because again it shows the short term thinking because in the long term if you look at the largest fintech companies, their market caps, their financial company, JP Morgans, etc. They're all half a trillion dollar companies. There's multiple half a trillion dollar companies and these fintechs like Square and others are only like I think what 1, 2%? So I think fintech is probably next to software one of the biggest investing mega trends over the next decade. I think even though there's ups and downs just because the opportunity is so massive across the world.

Yeah, just on the fintech stuff Mike, because you've spent a lot of time in that area. When you think of the pull forward forward, the behavior during COVID stocks like PayPal and Square, a great use case within that was upstart, right? That was a few years ago before we were talking about generative AI. Are you seeing opportunities in some of these names? This technology evolves and becomes cheaper for them to integrate into their products or services.

Yeah, 100%. And I think maybe something that is controversial. I've been investing in fintech for a decade and the thing that drives me crazy more so than anything is lumping fintech as like a single bucket. If you look at the multiple variants within financial services it is wider than any other category in all of business. So if you don't know each individual business model and product line and how they should be valued and what the risks are, you can get in a lot of trouble very quickly. And this happened a lot in 2021. And so I think that's number one in terms of the pull forward. A lot of this is public market investors learning this. Right. It's a decade old. And so the fintech companies that are public today, the private market ironically have a much better understanding of the business model and what's going on. And they sing all. When all, all of these companies went out. You mentioned a pure lending company. There's many pure lending companies that were valued just like a, like an adjunct, for example. Right. And then if we just peel back the onion a little bit as we value companies, you go back to fundamentals and basics.

Right.

It's quality of revenue, it's EBIT conversion on that. What does retention look like and how big is the base growing with inside of the business? And by the way, are you going to get hit with defaults in a downturn that's going to whack your entire book and P and L overnight and that has to be baked in and it just simply was not.

Yeah. Goku, really quickly away from fintech, is there a sector that you are focused on that is going to be almost an immediate benefit of integrating generative AI into their business model so they can leverage it from one of the hyperscaler. There's a whole host of things that they can do that we are going to see monetization in the near term and therefore a rerating, let's say of the whole sector or at least a handful of leaders in the space.

I think the whole notion of AI agents, I alluded a little bit earlier to this whole notion that you can hire an AI agent to do work. I think there are companies that are building horizontal platforms because ultimately if you're an enterprise, you have say 10,000 employees. You want to figure out systematically which of these employees can I augment with or replace with AI. And the reality is a compliance manager at a square is going to be different, slightly different work than a compliance manager at an adyen. And so both of them want to train an AI compliance manager, but they want to do it with their data, their workflows, etc. So companies that are helping enterprises transition to an AI human hybrid workforce are going to be, I think, very interesting. Overall, I think there is going to be an AI human hybrid. It's already there in engineering, in sales, in support. Every function is going to be a human hybrid.

Well, let's end on this. Like how are you guys thinking about about what this means for the human race. And let's start near term versus long term because near term it sounds like you guys are both in the camp that there's going to be a lot of job displacement. But ultimately like we've seen in every major technological shift, there's a whole host of new jobs that come online about it. Near term, do we have something to fear? Are the robots going to take our jobs?

I think one of the things I'm excited about, to be honest, is the emergency emergence of the importance of creativity in education. I think our education system has become too STEM centric if you look at college admissions. And I think we're going to see that as programmers and engineers are affected by AI, you're going to see the people who are really going to be able to use the AI are going to bring the left brain and right brain together. And so I think the importance of creativity in education is going to be re emphasized again. And to your point, every revolution over the technological evolution of the last century has first led to painful displacement. But then we always thought what's going to happen? Turns out people come up with new jobs, new ways to do things, etc. And so I'm an absolute 100% optimist.

Yeah. And you know that. I completely agree. And this goes back to the assembly line way back when. I think it's the first time we were talking about this or I wasn't talking about it, but they were. But I think what has been fascinating to see are these kind of like micro SMB businesses, right. Because of these models you can spin up an app just sitting at, at home. And what we know about this next generation is that they are very creative based on TikTok and the ability to completely, you know, immerse themselves online and particularly on some of these open source models. I think that's fascinating to think about was the micro entrepreneurs that are the. Maybe our generation was on Shopify and the next one is going to be building business on some of these models.

Thank you to Goko, thank you to Michael for being here. Thanks to RBC for putting on a great day. It was really. I don't know about you guys, I learned a lot. I met a lot of great people. People. And when we talk about optimism, when you see some of these companies that were here and what they have to say and the problems that they are attacking and then when you think about just the pent up demand on both sides to get this to a broader investment base, I'm really optimistic over the next few years also. So thank you guys, and thank you guys for being here.

Thank you. Thank you, Dan.