Invest Like the Best with Patrick O'Shaughnessy - Sarah Guo - The Power of Conviction - [Invest Like the Best, EP.383]

Starting point is 00:00:02 Hello and welcome, everyone. I'm Patrick O'Shaughnessy, and this is Invest Like the Best. This show is an open-ended exploration of markets, ideas, stories, and strategies that will help you better invest both your time and your money. Invest Like the Best is part of the Colossus family of podcasts, and you can access all our podcasts, including edited transcripts, show notes, and other resources to keep learning at join colossus.com. Patrick O'Shaughnessy is the CEO of Positive Sum. All opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect the opinion of positive sum. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of positive sum may maintain positions in the securities discussed in this podcast. To learn more, visit psum.vc. My guest today is Sarah Guo. Sarah is the founder and CEO of Conviction, an early stage venture capital firm built to serve AI companies. She She started conviction in 2022 after nine years at Greylock because she believes AI is the most

Starting point is 00:01:09 important technological advancement of our lifetime. In our conversation, Sarah discusses the challenges and rewards of leaving an established investing firm to start her own venture. She shares her unique perspective on the AI landscape and reveals her predictions for what we should expect on the AI frontier. Please enjoy this great conversation with a very impressive Sarah Guo. So Sarah, where to begin? You're doing recruiting for your firm.

Starting point is 00:01:33 I like starting where people are focused with their attention. What would you teach us about awesome recruiting for an investing firm? I think the hard thing in venture is the sample size is so small. And there is nothing on paper that is going to, at least in my experience, going to predict well if somebody's going to be a great early stage venture capitalist. If you're hiring somebody who is 15 years into an investing career and they have a track record, I think that is different. but the alpha obviously is hiring people before they're obvious. And so I think it's really tricky.

Starting point is 00:02:06 We are recruiting for what we think of as some baseline understanding of the technology transition that's happening in AI. And it's like really more just core technology understanding because AI is like a domain of this from both a product and a technology perspective that you can learn. But it's some context around software and then core traits. We care a lot about competitiveness, team orientation. Judgment is the hardest thing to look for in somebody who's really early in their career, but you can use understanding of businesses, interest in businesses, somebody's thinking about why a product works or doesn't as a proxy. And so kind of the core traits, but it is hard. When you think about early in your career before your judgment could be

Starting point is 00:02:54 assessed through a track record, what was the most edifying investment that you made? Like, what early investment did you make that taught you the most about this practice? I will use two very different companies, and it's hard to say early in career, because the learning cycle on the company is just so long if it is successful. And so the two I will mention would be me and my former partner, Ashim Chana, we incubated a company called Awake. It was a network analysis security company. And this was the year I joined Graylock, we started with an entrepreneur named Michael Callahan, who had come from a math and storage data systems background. This company was eventually sold to Arista, great product, great team.

Starting point is 00:03:41 This was educational because it's just a cult start. And I spent half my time working on building the company. So recruiting co-founders and early employees, talking to design partners, drawing the architecture, working on the substance of building a company from scratch. with the help of a venture firm. And so that's educational because you're doing the work. It's such a privilege. There are a lot of people early on in venture careers that are in a hurry to level up. And I understand where that comes from. I'm an ambitious person. I want to have agency. You want to have the decision-making capability. But I think it was actually this enormous gift that I

Starting point is 00:04:18 wasn't so sure. I wanted to be an investor when I joined Greylock because I got to just work on the things that focus on the substance of the job and helping entrepreneurs. And your time just feels less precious than when you have the opportunity cost of a network of opportunities that you are turning down meetings all the time. And the ability to just learn and do company building with a little more patience was great. So huge gratitude to both Asheam and all the early customers and Michael on that journey. And the other one was just Figma. We probably invest in this company year and a half or something into my tenure at Greylock. I was talking to Dylan about when we met the other day, and this is 10 years ago now.

Starting point is 00:05:01 And so Figma is obviously going to be a very important company. Without the longitudinal exposure, it's really hard to have confidence that your judgment on people and markets that are very early but have potential, it's hard to have the confidence that you're right. The pattern match, this is one of the core challenges of early stage venture, especially with the company that has a long build, like Figma, five years to a public release, you don't get the feedback cycle for a while. And so I learned a lot about just what inexperienced, but truly great entrepreneur looks like at the beginning. And that early process is also a different type of

Starting point is 00:05:39 company than most. I think it's really popular and even trite to say like, oh, product led, user-led growth now. But it's a particular type of magic. And knowing what it looks like when you do have user-love adoption within the leading influential group of a particular domain is just useful for pattern matching. Can you take me back to the moment that you decided to start conviction and tell the story around that big decision? I had the great luck of being at a venture firm, Greylock with an extraordinary history, 75 plus years old, been through multiple technology and internal generational transitions. One of the things that was really useful

Starting point is 00:06:21 is just seeing how unevenly the innovation happens in technology and how important recognizing that had been to the success of some of my legendary partners. For me, I'd been lucky enough to just be intellectually interested in AI. I had a friend who's an amazing academic entrepreneur himself, Andrew Ng, who taught me about deep learning like he taught lots of other people about deep learning. Me too. On YouTube, not directly, but.

Starting point is 00:06:49 Yeah, yeah. We have social friends in common, but one of the amazing things about Silicon Valley, and I know you believe this is just like a fundamental principle, positive sum orientation, is it is not a meritocratic place, but is the closest I have found to that. And an example, I'm not saying like I deserve this, but the marketplace for ideas in Silicon Valley is pretty good. So I wrote some random blog post. It's like an associate at Greylock 100 years ago that was talking about, like LSTMs or something. And Andrew emails me cold and it's like, hey, we should talk about this. I'm like, oh, sure, Andrew.

Starting point is 00:07:23 And that's the beginning of a friendship and lots of other stories that you or I will have like that. But I was intellectually interested in this area. I made a first investment that did not work at all in travel planning in like 2015 or something as a seed, sold the company. And then going back to you see the results happening in academia. You see use of increasingly powerful models. in the traditional machine learning sense happening at all the large internet companies.

Starting point is 00:07:50 And you see Transformers happen in 2017. And at some point, the curve just becomes clear. If you think it is actually a huge change in the landscape of software, not an incremental thing, then the opportunity becomes a little too big to ignore. And so it took me a little while to arrive at that conclusion and also say, I really want to do pure early stage versus a more multi-stage firm. And a part of it is actually nothing to do with Greylock and the opportunity and more just, I still wanted to be an entrepreneur.

Starting point is 00:08:20 This is an important question because so many people listening, I'm sure, have thought at one point or another about starting their own investing firm. So you have this moment. You see this wave. It's worth going after. You want to be an entrepreneur. What are the other steps? Literally, what did you do first?

Starting point is 00:08:35 If you go back to that first week, like, okay, I'm going to do this thing. What did you literally start doing first to get this thing off the ground? What did I do? I guess this is not perfect. It is complicated to leave a venture firm that preoccupied me for a little bit. Say a bit more about that. What makes it complicated? You want to be responsible to the personal commitments you've made. I carried some boards for Greylock still. And I'm very invested in those founder successes. So I think you want to sort out that and just like leave things well. I guess step one would just be I had planned to go take six months off and like hang out with my kids. And I'm allowed to say, like, I'm glad I didn't.

Starting point is 00:09:14 I don't think you're supposed to say that sort of thing about it. Say it all, please. Well, it would have been a shame in terms of the opportunity, if that is your only lens. It is my primary lens right now with lots of love for my family. We launched the month before Chad GPT launched. And then there's some credit assigned in the market for taking a point of view that is actually a point of view versus like following. And I wish I'd done this even three to six. six months earlier. So if anybody who is really convinced that they should do this, but they're like,

Starting point is 00:09:45 oh, it'll be better if I just optimize like another month or three months or something. Like, no, don't do that. The market is moving. It's always going to be hard. And the first thing was I was just like, oh, I guess there were LPs that I already knew who wanted to talk when they heard I was thinking about doing this. And I talked to them out of love and respect in long-term relationships. And then it just felt like I should go raise the fun. So I guess step one, be go talk to LPs. In parallel, I was trying to figure out what the strategy was because I wasn't quite prepared for that yet besides the fundamental orientation of, I'm not going to do this unless we can be truly excellent, tried to be dominant in some scope of asset class or domain or

Starting point is 00:10:26 whatever it is. I'm a partnership-oriented person. I need to go recruit and raise money, I guess. So those are the first few steps. But I actually think it's not very complicated and I wasn't terribly prepared for it. What was your first investment at conviction? Yeah, I guess the first dated investment in the fund would be a company in the legal application space called Harvey. I invested as an angel right before the fund was fully closed and we kind of warehoused it. Tell us about that company. That's a well-known one. Maybe it's actually a good excuse to talk about what you look for since it's an application in an AI application company. A lot of the attention goes to the core fundamental technology, which we'll talk about that too.

Starting point is 00:11:06 but what is your process for evaluating an application that's built on top of this set of models? The core framework for evaluating software companies remains the same, which is all those things. Distribution and quality of the people and size of opportunity matter, and so you make your judgment on that. If you ask me like what is different in the AI application landscape, I would start with the premise of believing the narrative in the landscape is a very dangerous thing. I'll explain what I mean for a second here. you, of course, remember maybe it's still going on. I think it's like being debunked a little bit. But when there's so much change and so much opportunity, it's very easy for investors

Starting point is 00:11:46 to go ask themselves and each other and entrepreneurs, oh, where is the value going to be? And some dominant narrative to emerge, like everything is GPT wrappers and there's no value in the application layer. So that was like a very popular narrative for a period of time. I think there's still some of this. But when I say it's dangerous to believe in the popular narrative, first of all, that narrative serves somebody. Somebody is pushing that narrative. And second is just it's the opposite of coming from first principles of saying, okay, we have this alien, magical new set of capabilities.

Starting point is 00:12:18 Let's assume that set of capabilities is going to progress in some way. We need to make predictions about how much that set of capabilities progresses, how dangerous or positive it is for us. to have this foundational technology coming from a single vendor, multiple vendors. And then what is the value of the workflows and customer relationships and specific data and change management and distribution that goes on top? I think a lot of that was considered a little bit, oh, like it's last mile. Everything's the foundation model. And it's like, who really cares?

Starting point is 00:12:53 Because all the companies look the same because they're really thin. And I'm like, oh, that last mile, that looks like 99 of the 100 miles to me. It's very similar to a lot of the work that went into building a great company prior. And I'd say what's different is you don't want to build things that the foundation models are going to replicate in terms of capability in the near term because it's wasted effort and then it's commoditized. So this is what you do need to predict on the capability side. But I think the 99 miles is a lot. And in the case of Harvey, I do think another perhaps distinctive characteristic of our portfolios at this point in time, We have a lot of teams with research backgrounds.

Starting point is 00:13:34 This is not all good, by the way. I'd say with total love for our research community and all of the people who are researchers in our portfolio, I'm saying we have chosen to back a lot of people in research and it has some disadvantages because those people have often never met the market. Academia or working at DeepMind or Open AI is a very different environment than having to go talk to customers and build your own distribution and have more limited resources.

Starting point is 00:13:59 But if you think about this traditional diffusion of innovation curve, I think diffusion of innovation also happens with talent. If you have been working on large model training or tinkering with these models for longer than other people, your depth of understanding of them is much better. And so I still think we're like so, so early in the adoption and understanding cycle of these models that it is a significant advantage for people who understand how to build products around them, where they should be advancing the state of the art and making good predictions at this foundational technology layer. And so our teams often, but not always, have more research DNA than traditional

Starting point is 00:14:40 software companies. If you ask me, is that going to be true five years from now? I'd say, I really don't know. But I think at this point in time, there's overrepresentation that I is conscious. So what has it been like to watch that team build a product from scratch using new technology? Like what would you say since you were in it early, like the discrete phases of that business so far and anything that's portable to others that want to build using this technology from having watched them? The combination of DNA and technology fit for purpose is very obvious where Gabe is a researcher. Winston comes from the legal domain like White Shoe Law firm. And then they have the core entrepreneurial traits of overall velocity, work ethic, and customer orientation, not just customer

Starting point is 00:15:23 orientation, but ability to go out and engage customers and convince people to take risk with them. The phases for most companies are like, get a prototype out, go get first customers on board and show them value, build the product suite and expand your lead and improve the organization and it's like scale. We're early in that third phase now, but the company is in publicly reported tens of millions of runway with a bunch of leading customers. And I think it's just credit to Winston and Gabe and the team so far that one of the hardest things is law is a. more conservative industry. And so your ability to take something that is fundamentally probabilistic and sell it to a relatively conservative or like risk-aware, non-technical audience is challenging. And so they had to do a bunch of work to get people to understand how huge the economic

Starting point is 00:16:21 benefit would be and go build that credibility. And I think if it be Alan and Nguer, or PWC or some of their early partners, they're building that credibility in their lead. What do you think the simplest way is to understand what that product does today for its customer? You had to explain to someone that had never heard of it before. How would you describe the product today? Yeah, I mean, assuming that they have heard of co-pilot, I think the simplest thing would be to say, this is co-pilot for lawyers that does increasingly sophisticated tasks.

Starting point is 00:16:53 And the product makes the application of this co-pilot increasingly automated. And so more and more of the work that used to take a junior person in a legal firm hours and hours should be something you get an answer back automatically for. And that could be search. It could be writing. It could be precedent. It could be hard, grinded out legal tasks. It can do in faster, more automated fashion. Yes. And so I think that's a significant piece of the product. If we zoom out and look at this versus our general lens on what is valuable in AI applications, There is making people who might be billing $2,000 an hour more productive, hundreds, two thousands of dollars an hour, more productive. And then there is do things that you can't do at scale today. The perhaps simplest version of this is, I want to understand if a term exists in these 20,000

Starting point is 00:17:48 contracts, but it's not going to be written exactly the same way. And if that costs you X tens of thousands of legal hours in or out of. of house, you're not going to do it. If it costs you a Harvey rod, you might. And so I think these are two of the forms in which the product is offering value today. Can you say a little bit more about this term I've seen your writing, which is to avoid the path of incumbent strengths? And maybe Harvey's a good example of this where probably unlikely that anthropic or open AI are going to build like a purely legal-based model. So it's probably not part of their big roadmap and ambition. And so it's off that path of incumbent strength. But say a bit more about

Starting point is 00:18:27 that and how it relates to company picking and investing? I am very willing to bet that very large, very capable incumbents are not going to go win every market. And so part of the judo here is being smart and really intellectually honest about what incumbents, let's consider some of the core foundation model companies, incumbents now in terms of their resourcing and ambition. They're not going to be good at everything. The technology might be general, but if you think about the layers on top and how that technology lands with an end user, the distance is pretty far. So understanding where that distance is really far and then what these companies most care about, you need to make those predictions as an entrepreneur. And so I'm a big fan of Arvind and perplexity, but if you look at the search GPT announcement that just came out, I think it was pretty predictable that they were going to do search at some point.

Starting point is 00:19:25 And that doesn't mean don't do it. It just means be aware and don't be surprised and have a strategy against it. I still think entrepreneurs should go after these opportunities, but the ability for an incumbent to be truly great at solving a problem or a product that is very secondary in their business doesn't drive the enterprise value of the business. Let's say for Google, it might be anything out of ads and then cloud. If you want to go after ads, they're going to attack you violently. there's real reason for them to. I'm going to offend somebody out there with this.

Starting point is 00:19:58 But if you're working on the like 19th most important product at Google, maybe that's okay. Sometimes people look at incumbents and they say like, oh, this thing is so scary and so powerful on them. Like, oh, man, like that thing has a lot of political challenge. You see projects end all the time based on how priorities shift. And if you look at where this product idea stack ranks in the economic impact to an incumbent and what the quality of people that are working on it in the organization are,

Starting point is 00:20:29 maybe you feel a lot better. Entrepreneurs should mostly focus on the problem they're solving for their customer. But if they are going to worry about incumbents, making richer predictions about what players in your ecosystem are going to do and care about, I would encourage. Can you explain your notion of minimum viable quality? There is still this theory of in B2B workflow software you need to do enough tasks, and it might be like enough operations on a particular object type in order to fulfill core workflows where there's actually clean abstraction from the next task or a piece of software, or you can just like replace an existing piece of software, a CRM or something.

Starting point is 00:21:14 If you're going to do that in a particular segment, you might need to assign reps and quota and accounts have some queuing or allocation function and show BI against that. There's a lot more in CRM, but let's keep the table of customers and assign reps to accounts and track it. And so you have this idea of minimum viable product scope. What is a different concept that we use in AI is mostly you're not deterministically writing the source code for specific workflows. you're manipulating a model to produce an output that you can think of as some step in a workflow

Starting point is 00:21:53 process that somebody had, like an end user business customer had. And it doesn't even have to be like a business customer. We can talk about Tesla and video output as two different examples. But you can use machine learning models to do many different things. And the question is, are they good enough to sell? I think really having a strong point of view as an entrepreneur about that. and validating that with the customer is going to be a huge part of finding product market fit for companies and being creative about how to improve the quality of the model, use user experience

Starting point is 00:22:26 and traditional product to wrap it such that the experience quality of the model is better, dealing with failure cases. There are many things you can creatively do as well as model improvement or model system improvement. But if we're more concrete about two examples, I'm on the board of this company called HeyGen and they generate video with people in it today. And people use it for all sorts of commercial use cases from like a McDonald's commercial to a creator on TikTok making influencer product videos or whatever. But a North Star for the company in terms of how deep the video problem is, is well, is somebody going to make a 30 second Super Bowl advertising spot with this? Not yet. That's pretty far. People want a lot of control. They want a lot of creative degrees of freedom. Everything needs to be invested in. And one of the ways this company, Hey Jen thinks in working with their customers is for some type of video asking customers, like, are we good enough?

Starting point is 00:23:23 And one of the core breakthroughs for this company that's been growing really fast of the last year is if you're Patrick and you want a commercial use video to promote this podcast or positive sum or one of your companies, is their generated video of you speaking by yourself in a room good enough that you're comfortable using it. And until this past year, until Hey Jen broke through that minimum viable quality barrier, it just wasn't for like any player. Input ease matters. So for Hey Jen, it's like you give them two minutes of video. That's consumer quality. It could be you and me on a webcam not recording something perfectly and you'll still get good enough out. If you had 30 cameras in a studio in front of a green screen that are all super high-deaf professional cameras, this problem is a lot easier. it is a little bit multidimensional, but trying to define within a company what's good enough quality as the customer experiences it is like this really important new thing.

Starting point is 00:24:20 It's a moving target as you expand scope of use cases. If you go look at self-driving, like the scoping of minimum viable quality for a particular use case also seems really important. I don't know if they're allowed to call it full self-driving or not at this point, but stay in my lane. They reached a long time ago. Now you have minimum viable quality as proof. to safety regulators in the city of San Francisco for a taxi with no driver. We just got to this year. You can use this concept across a lot of different machine learning products and then it feels like quite different than scope for other types of companies. It brings to mind that classic idea of the uncanny valley in a lot of this stuff. Talk through where the valleys are still the most uncannier, the hardest to cross. It is crazy how so quickly some things just seem to

Starting point is 00:25:08 have crossed it and just seemed to work fine. We were on a road trip recently, and I had one of these audio models, like narrate our journey from my kids, like teach us about what we're driving through as we're driving through it. And I had this thought of like, holy crap, like, it's good enough. This is fine. This is better than having a human guide in the car because it's more knowledgeable and we can tune it how we want and have it explain things a certain length. And the valley is crossed, so to speak. So I'm curious where you see the most interesting valleys that are left.

Starting point is 00:25:37 Some of them have been crossed. Some of them haven't. What hasn't yet been crossed, do you think? Oh, I think the world is full of valleys. It's all valleys from here. We've crossed three of them in little ways. Even in what should have been, what is considered to be like one of the very first domains, like writing, you're somebody who cares a lot about quality of content. One way to think of the uncanny valleys and how to cross them faster, what the foundational capabilities offer you.

Starting point is 00:26:05 And increasingly, the labs are all working on reasoning that should continue. to improve writing output generation. It's a super multidimensional problem. And for example, if you look at a writing product, whether or not I use something for writing is going to depend on, oh, well, man, I'm so OCD about what I will actually publish. If you generate something for me, and it's going to take me just as much time to edit it into what I want,

Starting point is 00:26:32 then I'm not going to do this again. Versus part of the new product thinking for these AI companies is going to be, how can I make the cost of managing errors or going from 80% quality to acceptable quality, really, really low cost for the end user? Another way to think of this would just be there's increasing interest in just how to do things that are really cheap for models to create a better end user experience. And so in more and more model systems, they generate a bunch of candidates. and they have some sort of verification or ranking against those candidates to get to an experienced

Starting point is 00:27:13 outcome for the end user that's better. And it's like cross the valley that way. A really important valley, just to make sure I answer your question directly, is, well, how much code can you generate from a natural language specification? Not much today. There are a bunch of different benchmarks out there, but like if you look at sweep-inch or something, it's like, oh, well, all of a sudden we got to 13, 15, even 20 percent. That's not good enough.

Starting point is 00:27:34 We have some software engineering interns right now. If they generated stuff that was good enough, 20% of the time, they're fired. Go fix that. The question is, well, also a classical problem in computer science and math, but like, is something more easily, more cheaply verified than solved? I think that's going to be true for lots of classes of model problems where people are designing verification and designing ranking because this is, I think, one of the most recent progressions and code generation. But if I can, like the Google team did, generate a million candidates

Starting point is 00:28:06 and then come back with a reasonable view of the top three, the end user experience is much better. If I have seven interns, look 20% right, but then I know which of the best two intern answers I should look at, that's better. And so we're going to have foundational improvements. We're going to have company-specific improvements, but they're going to be a combination of different systems approaches and research approaches to this. So I am very optimistic that lots of valleys, lots of people pushing for valleys to be crossed really fast as experienced by us or other end users.

Starting point is 00:28:42 What have been the biggest mistakes that you've seen application companies make that we haven't talked about so far in this fast-changing landscape? I know I just slammed this idea of the GPT wrapper narrative at the beginning, but there is a seat of truth in it. maybe I'll backtrack a little bit. State truism here, which is like, well, if something is really easy for you to build, unless your distribution is totally unique and defensible, you're probably not going to be able to capture rent or economic value on that for a long time.

Starting point is 00:29:15 And so we still see plenty of entrepreneurs get into initial traction with something that amounts to weeks to months of work in a handful of prompt templates, essentially. but the idea that could that be an entry wedge to get customers to engage with you so you can serve them more deeply, totally. We're not expecting that companies build billions of dollars of enterprise value with a software project that is three months old. But if they think that is the long-term answer, we disagree. Because if they have done it in that period of time and it is that simple, competition is coming. There's a lot of people who recognize the opportunity. And so I'm somewhat surprised at how short-term people are in terms of approaching the strategy piece of this.

Starting point is 00:30:03 What problems do you hope the next generation of frontier models solves? What do you most hope for as leaps forward in the next generation? This is more optimism grounded in a lot of really smart people at the labs and large players, having confidence this is true. but if you could have models that were better calibrated on whether or not they are correct, they can tell when they have a good answer or not. They become much, much more useful. Hallucination management is the blocker for many, many use cases in the enterprise,

Starting point is 00:30:44 and there's a lot of confidence that the next generation of models improves against this. The most obvious answer is just going to be a multi-step reasoning. It is a core mission of what the labs are working on in terms of solving that in a more general way. So the things that I'm most hopeful for, multi-step reasoning and self-improvement, the things I am confident on that have commercial value, hallucination management. Let's say you get everything, like we get the best possible version of GPD-5. Where are you like going to run towards in terms of seeking new company opportunities, like what spaces will get unlocked or use cases will get unlocked in your mind by like the perfect

Starting point is 00:31:26 version of GPT5 that would get you excited to invest behind? I would say, I mean, the perfect version of GPD5 is a lot. So, you know, root in the way you just described. Bring it out. But, well, of all of those things are true, I think we're in a really different zone. If we just narrowed the scope of it to models that were more verifiable, more calibrated, better hallucination management, these are things that block the enterprise from adoption. There is, I actually think people are asking this question for good reason, but there's so much enthusiasm and so much CAPEX spend against AI model development right now. The idea that we're going to hit an air pocket because adoption of these tools and the enterprise where there's

Starting point is 00:32:06 real economic value is lagging, I think that is real. But if you talk to companies like, oh, there's the things that are just the enterprise overall, like change management. But when you ask them, like, what is the problem from a risk perspective, the adoption of AI in large, foundation model-based applications internally or externally built in large financials, traditionally a huge spender in technology is marginal. It is extremely low. And it's because from a risk compliance reliability perspective, the bar is really high. We haven't hit minimum viable quality.

Starting point is 00:32:39 And so part of that is all the clever things we're talking about in terms of entrepreneurs designing systems and product scope such that the experienced quality by the user is good enough. But part of it is just we need a little bit more from the models. but it's coming. We need lower hallucination from models. Maybe say anything more you've learned about who is overrepresented and underrepresented in the buyers of AI products. Like that financials point is really interesting. Who are the biggest buyers of this stuff already? And is there any other group like financials that you see as big laggards that will become addressable when the quality gets higher?

Starting point is 00:33:12 This AI thing is making me personally rethink a number of my assumptions that were long held as an investor, but we may be surprised. I think we're going to get some leapfrog effect as we get over this minimum viable quality bar in different domains, because the areas where we're just talking about healthcare operations, I had not been super enthusiastic about selling healthcare IT over the last decade. I looked at it often on health care is a quarter of the American economy. I mean, globally, it's very important to every single human being. It is not particularly efficient. It is especially inefficient in the United States.

Starting point is 00:33:52 It is hard not to want to work on, except then you look a little closer at the companies that actually work in this space. You're like, ah, the incentives are a mess. It's super slow. The ecosystem is extremely complicated. There's regulatory capture in every zone. But there's so much to unlock there because the number of people employed and the number of hours wasted that in the administration of health care in the United States that

Starting point is 00:34:18 causes a good deal of why our health care is so expensive is really high. If you ask me about another area that I wasn't excited about in terms of investing in, but generally, please still call me, especially now, but might be, as the models get better, is government services. These are all huge parts of the economy that are incredibly inefficient where people are doing analysis and moving data around the workflows you can picture in a way that make a ton of sense to apply these model capabilities against creatively. I guess maybe my fundamental optimism comes from this place as well, if we make it a hundred times cheaper, no matter how complicated that ecosystem is, I think we can sell it. That's one thing that we've begun to see. And the leapfrog effect is

Starting point is 00:35:02 the types of businesses and the types of functions that had the most inefficiency can be most ripe for selling new solutions. What's your best guess as to what the market structure will look like for the foundation model providers in like three to five years. What do you think the healthiest version of maybe don't predict it? What do you hope it looks like? We work closely with all the large providers. And that means take our portfolio over to see the foundation model companies, make sure they have research connectivity, co-invest with them.

Starting point is 00:35:35 And we are also first round investors in Mistral. I'm very loathe to make predictions here because AI is just really hard. Anybody who like really says that they know what is coming more than six months from now that doesn't need a large lab is likely. And even so is I think the large labs were quite surprised by the success of open source over the last year. And so when you say like what do I hope for, I want a million flowers to bloom because the last mile to humans with more joy and play and productivity in every corner of the economy and every part of the globe, that last mile is really long and there's a lot of them. And so it's very hard for me to imagine a single company or two or three walking that mile,

Starting point is 00:36:26 essentially. And so what I would like to see is you have these amazing assets, you have multiple options, you have open source. And there is enough choice and competition and robustness at this layer that you can have companies built with true economic value on top and in partnership with them. I think that will happen. The alternative point of view is that the race for development gets increasingly expensive. I mean, that's going to happen anyway, but like that it becomes harder and harder for people to keep up. But an interesting dynamic here is the number of people who know how to train large models is increasing. And the second-comer discount is very high.

Starting point is 00:37:10 as that expertise is known, as you get to use modern hardware, as you get to not train all of the intermediate models and just try to be at the state of the art with the state of the art techniques. It is a competitive market. The commitment of Zuck to that market makes it also much more interesting. I still think that enterprises have some challenge believing that Facebook is going to be a great long-term partner to them. But the more they work with the ecosystem of deployment and inference and consulting partners, the more credible it is. And a rich and competitive ecosystem of models means that there's, you know, not total rent collection at that layer and lots of opportunity for many companies to walk that last mile.

Starting point is 00:37:57 What's your model thinking about Mistral versus Anthropic and Open AI? How should we think about the difference between them? I know one's open source, but like maybe say a bit more about, yeah, how you think about it, maybe even why you invested early on. Going back to the description of the ecosystem that I think is great for innovation and great for the end user, it is one of faster, more democratized progress and economic value at multiple layers in the stack. And I think that is Mistral's view of the world.

Starting point is 00:38:26 Arthur and Guillem and Timothy are amazing researchers with a bent toward both state-of-the-art performance, but also a view on efficiency. And again, going back to blindly believing, some dominant narrative in the AI landscape is very dangerous. There was a narrative. Maybe it still exists. I don't know. That efficiency doesn't matter. Model efficiency doesn't matter. It doesn't matter how expensive it is to run or how big it is. You just want the best performance. And nowhere in the history of computing has efficiency not mattered because somebody is paying for it in the end. And so maybe the model company can absorb it for a very long time if they've got like a great

Starting point is 00:39:09 other business model like chaty pt, Facebook, Google. But in a vacuum, more efficiency is better for latency and cost and scalability reasons. And so one big potential change in the ecosystem is whether or not people are going to start using a mix of distilled models, models of different scales, putting them in compound AI systems, and getting to performance and efficiency. And so efficiency was not a huge consideration amongst research labs and application companies until relatively recently. But I'm sure you've heard this term towards intelligence too cheap to meter. And I will make you a claim that there's no such thing is cheap enough because we've been working a compute for a long time as an industry.

Starting point is 00:40:04 And it is not so cheap that nobody cares because we just keep doing more with it. There's no company that isn't like, oh, if compute resources were free, similarly, we'll look at this and be like, if intelligence resources were free, we wouldn't do more. But they're not free. I do think efficiency, especially if we start getting to the next generation scale of models, even limits to data center size and power, thinking about using the resources that we have well, both in training and in inference for every particular application problem, is just going to become a much bigger consideration. You see it now. A year and a half ago, I'd say focus on data quality wasn't considered the highest status part of research. People

Starting point is 00:40:48 care a lot about it now because we run up against some limit, which is, well, we took the internet data and improved quality is going to lead to improved model quality. And so I just think the dimensions on which people optimize change over time. I'm very loathe to make any prediction, but I think it's going to be a rich ecosystem, and I'm hopeful of that. We've been told that software is like the ultimate business model for lots of reasons. And if you look at some of the growth for these firms like Open AIs, revenue numbers are incredible. For how young the company is, it's like this huge revenue business. Maybe the only thing most staggering is like their OPEX, like an incredible amount of cost in these businesses. So far, business models that are,

Starting point is 00:41:28 don't look like the traditional zero marginal cost software business model. Say a bit about margins and the basic stuff and how you think that will evolve, given that so far we've seen these things are freaking expensive to run. I mean, we just had a conversation about how efficiency matters will matter more and more. I am not particularly concerned about opening eyes OPEX. The winds behind them, not that they like shouldn't care, but I am not concerned that this is like an existential thing for them where like chat GPT can never be a good business. It depends on whether or not chat GPT is like a differentiated product in market, but the wins behind them look like the rest of

Starting point is 00:42:09 computing investing in efficiency against these workloads. And so we're super early. Right now, we're in this crazy immature part of the infrastructure cycle. So let me like say a little bit more about that. Look at the history of where'd you get a server for, let's say, a client server a web application. You had it on-prem and then you co-located it. Somebody gave you like power and real estate and cooling and you put your server there. And then you have hosted renting server and the data center from somebody and maybe they're even running an application for you. And then you had virtualization. I'm renting an instance. And the kiddies, I graduate from computer science programs now, if they're not thinking about AI, serverless. I don't want to know that

Starting point is 00:42:58 there's hardware behind that. Click button, Amazon, Google, infinite capacity. In AI, we're like step one or two. You are building and deploying your own clusters in data centers. You're in three-year reservation mode. You're managing your own depreciation cycle. And you're like, oh, no, our B-100s from Nvidia going to totally destroy the amortization schedule on this? I don't know. That's a really hard environment to innovate in. And so power to the labs in every company inside, outside our portfolio that thinks about it because they're training on large resources in this super immature ecosystem. Think about the amount of compute that we have accessible in our phones when like a huge ecosystem is focused on the efficiency of getting more and more power into our

Starting point is 00:43:46 hands here or in any other machine. Could we use as an ecosystem more competition to invidia, yes, this stuff will get cheaper if there are multiple players. I mean, Nvidia is innovating pretty fast, but pushing Nvidia and then eventually challenging a 90% margin product. But overall, I expect the industry to get much, much better at hardware utilization and at efficiency in a bunch of different ways at every layer of the stack. And so when I think about the OPEX of Open AI as an example, but any other application company, The optimizations that a company itself can do, then in infrastructure management, I'm on the board of this company called Base 10. You want to do inference on GPUs of different types that is serverless and not write all of your own scheduling and GPU failure management and whatever else.

Starting point is 00:44:38 You can do that. And then there are people working on everything from memory bandwidth to chips to systems. The entire ecosystem has innovation that we expect to as an industry reap the benefits of over the next five or 10 years. There's no reason to me that won't happen. I can't predict how quickly the cost curve comes down for the same level of compute. And I think it will still feel like you always need more. Think about progression of PCs. You always want a latest processor.

Starting point is 00:45:08 But I am not worried that the end margin structure of any application company is bad or any worse in this era. It's not clear to me that that's a structural problem versus the software era. That's probably the best answer I've gotten to that question and I've asked a lot of people, which makes me wonder like, what else do you want to tell us about infrastructure that you've learned? I'll claim one thing that's maybe interesting to a broad investing audience like you have, there are a lot of people thinking about why is there not a legitimate challenger to Nvidia and what are all of the structural advantages that Nvidia has? and broadly, basic analysis of the landscape would land you at, okay, AMD is working on this

Starting point is 00:45:55 and their software stack has not been competitive to date, but maybe they're getting closer, software and hardware, but weakness in the software stack. Then you have a series of proprietary chip efforts, sometimes from acquired companies, at Microsoft, Amazon, Google, rumored now opening eye. And then you have upstarts who are trying to do. either systems with a particular DGX boxes, more like a mini data center than a chip, but either like a full system or a chip. And so you have rock and cerebrus and etched and Maddox and whatever else.

Starting point is 00:46:31 I would love to see most of the industry that isn't a giant holder of Nvidia, but even some of those would love to see more innovation at this layer. People understand the power of Kuda and the optimization and effort that's been going into that for more than a decade and a half. One of the core challenges that people don't recognize is we had a portfolio company who used a different chip, not Nvidia GPUs, to try and train a large scale models. These are X large lab researchers who know what they are doing. And knowing is saying that the like engineering and infrastructure management work that they did was perfect, but it exposed me to an understanding of what is making this problem so hard for people

Starting point is 00:47:23 to attack. If you don't have people running large-scale training workloads at multi-thousand node cluster size on your chips, you do not know if they are going to work. And so this isn't something that you can do as easily, okay, we're going to. to do test in simulation, test in FPGA, and then like we have as much confidence as we can that chip will tape out, chip fits the spec, it is useful. The types of failures that can happen in large-scale training, which is a huge driver of consumption that a lot of people are going after, it seems to, at least with some of the new contenders, only come out at scale. And so that's a really difficult risk to mitigate where I'm like, I don't know, can you do that in SIM?

Starting point is 00:48:12 because it's a new problem. I think that's kind of interesting because then if you look at, for example, Google and their chip efforts, those work, actually, because Google has the workloads where they're training on their own chips at scale. And I would ask you,

Starting point is 00:48:29 do we know that the other chip efforts are working? You either need the patient zero or that patient zero to be yourself because it is very hard thing to answer without that workload. Doing it physically is really expensive because then you need to go deploy 10,000. Maybe that's just an interesting anecdote where we've been thinking a lot about,

Starting point is 00:48:48 is there opportunity here and what are the barriers? And love and respect for Jensen and a video, spending time with him as well. I think the barriers are pretty high. Back to like the thousand flowers blooming. It's cool to imagine all of the smart people that are economically and just curiosity motivated to tackle all of this all at once. It makes you, of course, wonder about like past technology. bubbles where lots of money has been spent trying a million things, and then we find our way through

Starting point is 00:49:17 and money's lost and money's made. Where with your like bubble hat on, do you have a bubble hat? Like, what does your bubble hat tell you? Talk to us through that specific lens. Oh my goodness. Of course I have a bubble hat. We have skin in the game. I'm a large owner in my own fund and this is the main thing. If the promise I made to LPs was multiple, we've got to go make good on that. I think most people should not be training general foundation models. Maybe that's obvious. That's a way to lose a lot of money. I would also say, as somebody who wants, I do think that there will be multiple players at the chips and systems layer and chip systems networking, the new data center layer. The opportunity is too large. Capitalism works. It will happen. It's just barriers are high.

Starting point is 00:50:06 But I don't think that the market is not infinitely deep for another LLM player that isn't particularly differentiated, but is just a handful of great researchers working on something that feels very general or feels directly in the path of the existing large players. And so I would say companies keep getting funded that are of this shape, because people look at and are envious of the value creation at Open AI and Anthropic and others. It's the great game. It's really interesting, of course. Like, you and I are interested in that too.

Starting point is 00:50:45 But when the entry price is now more than half a billion dollars to do pre-training of something interesting or at least get into the game of a general language model, that's a really easy way to lose money. Can I ask you the question I ask Martin Cassato, which I quite like to his answer to, which is, what do you think that 2030s look like in the world as a result of all of this, intentionally creating like a little buffer between now and that? And we're always so focused on like the next three years. But your kids are young, my kids are young.

Starting point is 00:51:18 What's our kids like teenage years going to be like? I think they're going to be a lot more intelligent and creative. So 2030 is really hard because I can't assign a probability to it. But there is a piece of me that says real probability of abundance. I don't know if we're all sitting on a beach partying in Europe or something. But I say that like a little bit glibly because the ability of society to ingest that much change in productivity is not great. There will be winners and losers and the job displacement dynamic, I think, is real. I also think it's probably not quite as fast as people are worried about.

Starting point is 00:51:59 So I'm splitting my mind between, okay, model progression happens. very, very fast, and we solve a bunch of really important problems and we have abundance. If I take a step back from that in terms of just optimism of timeline, I'd say the things that I can picture very clearly is the ability to learn and express a capability that would have been a very skilled thing before is going to just be completely changed. Your expectations around creativity of entertainment. Everybody has desktop Pixar. Everybody has the world's knowledge in the form that is most useful to them immediately. Andre is working on this educational tutor thing. Somebody can become an expert in something at an incredibly accelerated rate because of that.

Starting point is 00:52:51 There are a bunch of areas that should market structures be willing, buying processes, be willing magnitude of improvement being enough, become a lot more efficient. So I don't think I have a clear version of the future, I can picture a lot of really amazing individual experiences. Is there any commercial or product thing that you're surprised we haven't seen yet that you're like got your eyes peeled for and it hasn't happened and you're surprised it hasn't happened? I'll describe two opportunities that I'm excited about and one of them I know why it doesn't exist yet, even though it's so big. The vast majority of the time, incubating companies is a terrible idea because my foreign partner, Jerry, used to say there's no such thing as inception. People

Starting point is 00:53:34 have to live the problem and the ideas and go after and make it their own. The entrepreneurs that we want to act, they have the agency to go do it. But one of the places where I think this is not true is where essentially the context and capability to build a particular type of company rarely come together. This is one of the things that is so powerful about Harvey. How many times do you get somebody who is going to be entrepreneur on Ford thinking about workflow, that they are a lawyer working 90 hours a week at a good firm and are still like, can I make Chatsyipit do my job? And then his roommate is a great researcher. It's uncommon because you ask your average researcher at a lab, what does the legal workflow look like? That's not part of their life experience.

Starting point is 00:54:18 And so one of the really interesting opportunities that we think about and actually do support incubation around is places where the context, of like what people are doing that could be automated or improved is so far away from the engineering and research community that understands the capability at the state of the art today. So an example would be there are many billions of dollars spent a year on configuration, monitoring, and maintenance of enterprise software systems, ERP, HR, CRM, workflow systems like service now. That gets done.

Starting point is 00:54:57 by large consulting shops in the U.S., in India. And I think if you asked, if we look at the capabilities that are coming in software generation, in code generation, I think some of these tasks are pretty tenable, or you could at least create a services org that has a very different margin structure than today's orgs. Better margin structure, better SLA, better end-user experience,

Starting point is 00:55:24 maybe change the industry structure of software because this configuration and maintenance and depth of embedding in a particular enterprise system is the thing that makes it so sticky. The fact that you configured SAP 20 years ago, it's just all your business process you couldn't possibly get out of it. If you can replicate in a different system, that's a very high mountain to climb, but pieces of that could change the industry of software, enterprise software, and that's really interesting. But if you think about all of that work, that type of, of software engineering work, that's considered very low status work in Silicon Valley engineering

Starting point is 00:56:02 circles. What's happening over there? Who's writing configurations of this platform? It's so far away. But if you think about the economic value and what technology is actually super important to the largest companies in the world, that they feel like they can do nothing about or they can't improve, that's kind of it. That's, to me, the type of under-exploited opportunity where the domain is so far away from the technologists, even that's still software. Never mind healthcare operations or government administration or call center operations or something like that. And so I think there in Lars a little bit of the opportunity. Another one that we're just really excited about in terms of potential value for society. And clearly the data set is not generated and owned at the large labs today is material science as a domain

Starting point is 00:56:52 for foundation models. But if you ask me, like, why doesn't that exist yet? Well, I think you need to have the right talent, you have the right thesis on data, hopefully some cleverness around efficiency of collection of that data. Would you get excited by some sort of pitch around we're going to create the next Accenture, but built on top of,

Starting point is 00:57:12 basically the question is like, would you get excited about a services business that just was built around using and deploying all this technology to make the offering better and faster? and higher fidelity or service business is just too lame of a business model for you? The answer is yes, but what would get me jumping out of my seat excited would just be somebody with the relentless ambition to figure out if we can make the margins as good as a software business. And let's think less about margins, more about scalability, because you can go like,

Starting point is 00:57:44 okay, fine, why are services business? It's lame. Well, because you're fighting on human capital quality and volume. You can only scale so fast. It only gets valued X multiple in the public markets eventually. It is only so profitable. And hey, like to some degree, because you can't really tell, it gets a brand game in the end, go with a big consulting firm. Those are some of the reasons people don't like services businesses. And so for a story I could believe around scalability and ambition of scalability and all the things that come out of that, like how much of it is a technology business? Yes. The answer's high enough, yes. That's true when you look at some of the most interesting application layer companies today, they may not frame themselves this way or even be

Starting point is 00:58:30 consumed this way. And that might be the difference. What we're talking about is consumed as a services experience. But they're replacing things that were services before. Is there anything that scares you a lot? The things that are interesting to be afraid of, bio-risk. runaway model that is optimizing paperclip production, China weapons system or something, I think we should go figure out if those risks are real, but I actually think they're like a huge distraction from the near-term abuses. And so for you and I, investors in technology, technologists themselves, the ability to adapt to new tech is a built skill. But it will not be true across the entire world that people adapt very quickly. And so there are some incredibly basic

Starting point is 00:59:28 abuses where I would be excited about solutions. And it was very easy to say misinformation and fraud, but I think those are quite large. I'm more oriented toward, wow, we have like a lot of abuse that's happening already with older technologies that just gets so much more amplified, gets much cheaper with leverage of this generation of models and improving capabilities. We should address that pretty immediately and in a sophisticated way. Is there any debate in this field that you find really interesting personally? This is a little bit in the weeds, but I think the debate about how you improve multi-step reasoning in a general way is open question amongst the labs.

Starting point is 01:00:19 And this is like a very fundamental question of where does the next step function of intelligence come from beyond just scaling? Because on both a compute and a data side, that is more difficult than it was in the last generation. People have varied points of view on this in the different labs. What are the sides of it? How would you describe the sides of that debate or the different opinions? Labs and upstarts.

Starting point is 01:00:42 I will describe one directional effort that people are investing in, which is, I just had Oriole Vignals from DeepMind on our podcast to talk about this. But does it make sense to invest in math and computer science as a domain where it feels closer to pure logic, you are doing multi-step reasoning and in terms of improving model capabilities specifically against this domain and using generated data in this domain as new training data for the model. Lots of people are doing this. There's a contingent of folks that say that is not a general approach. Human reasoning is much richer than that. And so looking at code or looking at math, translating problems into lien and verifying them, that's to some degree a dead end because we cannot

Starting point is 01:01:38 do this with the rest of think about the qualitative reasoning that you do on an investment. There's no proof. That's one open question of, are there more general answers to improving multi-step reasoning that are not in a particular domain? Because some people will look at code and mass and extension of R.L. in games and is that the right path forward? I think lots of people are interested in it. It's like much less clear consensus in their research efforts than more scale, more data, human expert data, and better quality of data, improvement of the quality and the pre-training

Starting point is 01:02:16 as well. Do you think there are any important, I'll use the term crossroads that we're at as an industry in AI today, or we have to go one direction versus the other? I think a point of change Now, that's kind of interesting, is at the scale of clusters required for the next generation of training, these are for training essentially like single use gigantic constructions of CAPEX. And they're not easily extensible to the next generation. Obviously, you can use those for training smaller models, experimentation, inference, whatever. But the idea that we are just going to build larger and larger data centers consuming more and more power in order to improve model performance feels like it has some limits.

Starting point is 01:03:11 And it probably goes back to the discussion we were having before of efficiency has always mattered. And maybe there is increasing investment in some sort of modularity or that. that efficiency of training matters in a way it hasn't in the past. Because that is when you start talking about the multi-billion dollar training run that gets you to state of the art for some period of time, it's a hard cap-ex investment. If you could commission a research paper, like a crazy well-done, in-depth research paper on any topic in AI, what would you pick? Three domains that I think are interesting.

Starting point is 01:03:53 I can't choose one. But one is the first. The problem is nobody's going to write this research paper because it's all like proprietary knowledge. That's okay. You have a special magic wand. The comparison of generalization of multi-step reasoning, I think is an interesting area. A open question is we have all of these benchmarks, exams that humans in different domains take, MMLU or a legal exam, a medical exam, et cetera. We are surpassing the benchmarks in a bunch of different ways where it begins to,

Starting point is 01:04:25 open the question of how do we do evaluation of these models as they are super intelligent, not in any super philosophical sense of the word, but only in the sense that they are better than our human experts at a particular task. And so human eval becomes hard. How do we understand progress as we go from there? Something like ARC, would that satisfy what you mean? That's one version of it. I think another example of an idea is, have you very ever heard of the term centaur play? If you play chess. Oh, yeah, yeah, sure, sure, yeah. A lot of people in AI don't believe in this idea anyway. One theory of it would be Patrick with a model can come up with better answers than Patrick or the model. That will be the bar for human eval for the next set of

Starting point is 01:05:14 models, but that's a version of it. But it's an interesting question. How do we continue to evaluate models as models progress? Understanding the different theories on reasoning traces leading to model self-improvement. What is it in your personal experience that has led to you caring so much and willing to work so hard in this field? This is the most important change to happen in technology in our lifetimes. And so if you believe the impact is very large, not in like a religious AGI way, but if you just think the opportunity for productivity and abundance is very large and you can create new

Starting point is 01:05:58 economic platform players. And the landscape is so open because your opinion actually really matters now versus if you were investing in SaaS like you and I were in like 2015, there were incremental discoveries and understandings about companies and technology that did matter. I'm not trivializing that, but it was at a very different scale. If software 1.0 was about human engineers writing explicit instructions and source code. And that was decades. And then Andre Carpathie wrote a essay about software 2.0 being like replicating and finding in search space a specific behavior from a data set and a neural network and then training the program from the data set. This is like quite a lot of work to do from scratch. I genuinely think of this era as

Starting point is 01:06:43 a new era of software. We are manipulating models of different kinds where a bunch of the work has been done for us. The foundation models are so capable. And if you can build intelligent systems to do increasingly useful work, how could you not want to work on leverage these superpowers in every domain? Maybe do it better than humans in areas that really matter health care and science, but also in a daily work and taking the operational toil out of every industry. And so it's just really interesting. It's an ambitious era. I think that's super. exciting. From a personality perspective, we had some founders doing references on us in investment context recently. They walked away convinced that we'd go to work for them. And they're kind of like,

Starting point is 01:07:31 oh, people are surprised. It's like why you're trying so hard when you don't have to. And I'm like, oh, anybody in technology, like above subsistence level, what does have to mean? It's like the funniest observation to me. I guess my reaction to that is, of course I have to. What else am I going to do? let the world pass me by, be number four. And maybe they mean it like financially. Oh, like, doesn't money really matter? I'm like, no, but it's really fun to be right, go compete, work with great people. The discovery for me is maybe this resonates with you, the motivation of working with

Starting point is 01:08:03 the most interesting and motivated and high leverage people in the world and helping them be a little bit more successful and being right about that over a long period of time. Of course you have to. Hey, man. This has been so much fun. I think you know my traditional closing question. What is the kindest thing that anyone's ever done for you? There's too many to answer here.

Starting point is 01:08:25 I'm very grateful for the number of people that have given me opportunity is not an acceptable answer. I'm going to name three people. The first is, I mean, you have my husband Pat on the podcast. He's my favorite person in the world. I'm like, oh, thanks for asking me to marry you. Like, that's great, right? I just appreciate that.

Starting point is 01:08:44 Got very lucky. And then I'd say two people that I will be eternally grateful to for different reasons. One is Anil Bustery, who used to be a partner leader at Greylock and was the founder of Workday and Amazing Technologist where I think in Anil's view, he was like rescuing me from finance at Goldman. At one point, like, I was at Morgan Stanley and you like belong back in technology. And I'm like, I know. Like, I'm a technologist. I'm just here learning about the business side for a year.

Starting point is 01:09:11 And I'm like, I'm going to go work at Stripe or something. But the thing that Anil did for me was just, say, working together for a small amount of time, like having complete conviction. And I did too, but you have potential. We don't know what we're going to do with you. Just come work here. And so it was like raw goods bet of an opportunity where I'm like, I haven't done anything interesting in the technology world.

Starting point is 01:09:35 I am a kid from Wisconsin who is working in banking who happened to work on your IPO. and it just takes like a certain conviction in your own judge of talent for somebody to be like, oh, come work here, when on paper, who cares about your profile? I'll forever be very appreciative to Anil for that. My former partner Ashim Chana is legendary enterprise, in particular security ambassador. But one of the things that I seek to emulate in Ashim is his ability to work with earlier career investors really comes from a place of amazing personal confidence. One more sentence about this.

Starting point is 01:10:15 If you could be productive for, I think shame was actually like a little bit skeptical of me when I came in, which is totally okay, right? Trust is earned. But he became a really big sponsor for me. The attribute I tried to emulate is give your people as much and perhaps a little too much opportunity if they prove they can take it.

Starting point is 01:10:37 When I say confidence, he was incredibly. incredibly open with his network. He was incredibly meritocratic where I don't mean to be cynical about this, but there are many people who would have been like, I'm investing in enterprise security and you as a 24-year-old, really young-looking girl from no particular interesting accomplished startup background, yes, we're going to go do all of this company building work together. And I'm going to take your judgments very seriously, very quickly. I'm going to open my world to you. I think that's an amazing thing to do for somebody. And in terms of giving the people that I work with opportunity, may we try to be like that. What a wonderful trio of closing thoughts.

Starting point is 01:11:21 Sarah, thank you so much for your time. Thanks, Patrick. If you enjoy this episode, check out join colossus.com. There you'll find every episode of this podcast complete with transcripts, show notes, and resources to keep learning. You can also sign up for our newsletter, Colossus Weekly, where we condense episodes to the big ideas, quotations, and more, as well as share the best content we find on the internet every week.

Invest Like the Best with Patrick O'Shaughnessy - Sarah Guo - The Power of Conviction - [Invest Like the Best, EP.383]

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.