E44: How Benchmark Invests in AI with Eric Vishria and Quilter Founder Sergiy Nesterenko

0:02Welcome back to Turpentine VC, a podcast where we discuss the art and science of building successful venture firms, VC to VC. In this episode, we're focusing on Benchmark's investment philosophy in the AI era with general partner Eric Vishria and Sergey Nestorenko, the founder and CEO of Quilter. Sitting in for me as host is AI Scout and the Cognitive Revolution podcast host, Nathan LeBenz. Up ahead, they cover the questions that Benchmark asks to determine whether a new company is solving an enduring or temporary problem, and the reasons why Benchmark hasn't invested in a foundational model to date.

0:37Eric Vishria, general partner at Benchmark, and Sergey Nestorenko, founder and CEO at Quilter. Welcome to the podcast. Thank you for having us. Yeah, excited for this conversation. Today, I think kind of a little bit of a broader conversation on just kind of where this whole AI thing is going, big picture, and how you guys are thinking about both investing in companies that will hopefully stand the test of time and also try to build a company that will be able to successfully ride the wave and not be crashed over by the wave, which is definitely a challenging dance for many founders right now.

1:13Eric, maybe you want to start off with just giving a little bit of background as to what it was that got you guys interested in Quilter specifically and how that fits into your broader philosophy and portfolio? Sure. Yeah. Well, super, super excited to be here. Thank you again for having me and having us on the show. So obviously, there's no shortage of AI startups out there. So each of the benchmark partners is probably meeting four or five of them a week. And so you're kind of perpetually meeting new ones. And a lot of them sound really interesting. They're really interesting ideas. And I would actually even say the revenue traction on a lot of the companies is tremendous.

1:58It's just unlike anything that we've seen just in terms of the speed of revenue traction and everything else. But obviously, in the kind of timeline that we're investing as early stage investors, you're looking on a five or 10 year horizon, really, of the company maturing into a big business and wanting to be one of these exceptional kind of companies that get created in our industry a few times a decade. And so a big part of what we're looking at is which of these companies is not just riding the wave and not just on a sugar high, as I would describe it, of either revenue or traction or developers or hype investor dollars, but is actually building, has an opportunity to build a really, really big business.

2:47And obviously, that's really hard to figure out at the early stages of the company, but that's what we're paid to do. So that's what we try to do. and one of the frameworks that I use right often is just like, well, what's changing in the world that is going to enable this company to be really, really large and in the case of, which has nothing to do with company actually, it's just like what's happening in the world that is this and of course there's AI, which is like, okay, that's a wave but there's actually, you need something more, you need something more substantive and specific for a company to be able to be created and be durable around that.

3:24And I think in the case of Quilter, there were a couple of things that really caught my eye, which was first off, obviously electronics are permeating every aspect of our life, right? Like things that were like very, very simple not that long ago, like a, I don't know, a light switch now have electronics in them. They have circuit boards in them in ways that just weren't true before. And so you have a rapid expansion of the number of companies, the number of products that are electronics and have circuit boards in them. And that is a first very, very fundamental thing that's happening in the world and seems bad.

4:14And I didn't really understand this until I met Sergey, But like the process of actually designing or laying out these circuit boards, place and route, I guess, as it is called in the parlance, is a is like an incredibly manual and time consuming and slow process. And one of the things I think we know just industry wide is the faster you're able to iterate on products, the better and better they get. And so, you know, that kind of friction in the process seems bad in a world where electronics are permeating every kind of everything that we interact with. So that was kind of interesting. But then the real thing where it started to get, you know, where the story started to get really compelling to me was, you know, particularly, I don't know, six months ago or so, everybody was talking about co-pilots.

5:11Like co-pilot this, co-pilot that. There's a new co-pilot for lawyers. There's a co-pilot for doctors. There's a co-pilot for developers, which is, you know, in each of those spaces, there's like 15 companies that are that have emerged that are doing some form of co-pilots. And Yusuri's kind of provocative statement at the time was like, this isn't a co-pilot problem. It's just not for co-pilots. like you don't want a human assisted AI to do printed circuit board layout. And I was like, oh, okay, that's different than anything that I've heard in the last few months. Like that's just a different view.

5:53And so it's like, well, why? And it's like, well, like, you know, if, if, if a copilot or an auto layout does 90 % of the job, then by definition the other 10 % of these layouts and, you know, You can look at any circuit board and see what we're talking about, which is like there are all these thin lines that are connecting one component to another component. And they're all over the circuit board. And so you can just see, as you look at it, you're like, wow, a human laid that out. And that seems really tough. And if I had to draw another line, how am I going to draw the line and not mess up everything else that's already laid out?

6:26They're pretty dense. And so it's just not a human assist problem. It's a problem that should be AI and done. AI should just do it, like do it entirely. And in order to get there, we have to, you know, have a like a clean interface where we kind of get the design and architecture of it. And then we have a thing that can be, you know, spit out and go to manufacturing. And so, but those interfaces exist in the industry. And so we sit in the middle and we kind of like take on that, that aspect of it. And we just like lay it out. And he's like, we have to obviously start with simpler boards. We can't do a super, super complicated, you know, iPhone A14 to start.

7:10Like we got to start with something simpler and like work our way up, but we can work our way up and just be AI and done. And so I thought that was a very, that was the first like really provocative, like just makes you think like, oh, wait, there's a whole bunch of things that we just are like doing with humans and like Copilot's Vogue. But like is Copilot the right answer? Like in a whole bunch of cases, like Copilot's not the right answer. Like it's just not like we should just like let AI do it, which is cool. And so that was a different perspective. So that was the first thing. And then the second thing, which really kind of got me thinking, was, okay, everyone's talking about large language models.

7:47And everyone's like LLMs and Gen AI and blah, blah, blah. And he's like, yeah, we're not using any of that. That's not what we're using. He's like, oh, okay, well, that's different too. And I'll come back around in a second. But that was, you know, his perspective was like, hey, you know, that's not the right way to, that's not the right AI for this problem. Like we tried all of these other things, but like this is, that's not the right AI for this problem. The right AI, and Sergey can describe it much better than me, is like, is we design a game, we tell the game what the optimization problem is.

8:21And do we want to optimize for price? Do we want to optimize for performance? Do we want to have a really dense board that's expensive to build? but like super power efficient, or do we want to have a really cheap board that's like bigger and has more space on it? And if we actually like give the user these controls, then the AI will be able to give them a range of solutions and they can kind of pick where they want to be on that, on that solution set. And I was like, Oh shit, that's cool. Like that's a, that's a very provocative thing. And so those were kind of the very specific things like, which is like there's electronics permeating everything.

8:56there was a view that this is a AI and done problem, not a co-pilot problem, which was different. And there was a view of the right kind of AI to use for this particular problem, which isn't, again, the kind of thing that everyone's talking about. And as you kind of dug into it, it seemed like really cogent. And I know nothing about printed circuit boards and know very little about, you know, obviously specific technology. So like, you're kind of like thinking about it and then you're trying to validate it and then you're talking to people and going after it but the biggest thing that just to kind of zoom out for a second it for any entrepreneur and anyone kind of thinking through these things there's a chris dixon wrote this short post several years ago called the idea maze and if and if you haven't read it you should read it because it'll take you like 30 seconds but it'll stick with you because it's a it's a short but very powerful pose.

9:49And I'll summarize it, you know, not as eloquently as Chris said it, but there's, you know, what you want to do when you're talking to an entrepreneur is like you want someone who's been like rolling around in this idea for a long, long time, like they've been in the idea maze and they've kind of gone down a path and they've hit a dead end and then they've backed up and gone down a different path and hit a different dead end and then backed up and kind and try to figure their way through the maze. And as you're talking to them and asking them questions, what you keep getting is you keep getting like, yeah, I thought about that, but here's why that doesn't work.

10:26So here's why I think this could work. And like, and you're just like, and what you realize is like, maybe they have a path through or maybe they don't, but they've been exploring and living and rolling around in this idea and turning it over and turning it this way and that way. And so they just, there's so much depth on the problem. There's so much depth on the potential solution that you kind of are like, okay, this person, if there's going to be someone who's going to figure it out the way through the maze, like this person has a really good shot at it. And obviously with Sergey, his experience at SpaceX designing a bunch of boards, going through that, and then playing with and trying to figure out solutions and applications of AI to solve this problem.

11:12first realizing the problem trying to pursue a whole bunch of different ways to solve the problem having the realization that it really is a ai and done kind of problem having the realization that of the right kind of ai techniques to use for it those were just like evidence to me that that like he's like lived in this in this maze for a long time and um and so that that was probably like the the macro thing that gets you excited so it's this area of that and we can talk about like so many of these other things, but that, that was just like a very, um, you know, it just ends up being like very exciting when you find someone, um, who has, who has kind of approached it like that.

11:53Cool. I appreciate the backstory. Um, I kind of want to maybe circle back in a minute to this like disruptive approach. I mean, this seems like a pretty classic, almost textbook example of a disruptive solution in that it seems like it's coming in at kind of the low end of the market. It's serving people. If I recall from the other episode that we posted that it was striking to hear actually that SpaceX has an internal board team, but they couldn't serve you. And it's like, man, talk about a market that's got to be very broadly underserved if an internal specialist team can't even serve the other team at SpaceX.

12:39That's pretty wild. Hey, we'll continue our interview in a moment after a word from our sponsors. So it seems like there's potentially a very kind of textbook pattern here of starting at the bottom of this market, massively expanding the bottom of the market. I wonder how often that's something that you're seeing across the portfolio or specifically trying to do. But maybe before coming back to that, could I try a little bit of this idea maze stuff? I have a couple of ideas that I'm wondering maybe why they didn't work or why they wouldn't work. Would you be game for a couple of possibly harebrained ideas that you can shoot down, Sergey?

13:17Yeah, if I can, happy to. Okay. These are potentially quite novice ideas, but I guess for starters, what's the data landscape in this space, right? You know, typical deep learning approach is predicated on a lot of data. Is there any like open source data set out there that somebody could go like tap into any significant scale of like published boards that could be used in that way or was sort of just lack of data a forcing function to make you go in another route AI wise? Yeah, that's a good question. And probably one of the most common questions that I get about like how Quilter is using it for this problem.

14:04I think what most people tend to think is like, you know, how does Copilot work? You take all of GitHub and all of Stack Overflow, feed it into an LLM, and it makes good predictions based on the average human behavior. And obviously, we don't do that. So we don't do that for two reasons, right? One that you mentioned is, there actually just isn't that much data, right? So if you look at all of GitHub, not that many open source boards, there's a few sources here and there, but not a lot. The best ones are locked away behind companies. They're in Apple's repositories and Google's repositories and SpaceX's repositories.

14:36But there's another reason that's even more compelling to me why that's the wrong approach, which is fundamentally people are not good at designing boards. Just full stop. right so if it's a process that takes you know for a complicated board two three four months you are going to make a whole bunch of like margin on margin on margin decisions that make your resulting board like much bigger than it otherwise could have been use more layers than it could have be more expensive than it could have and if you just use data to do supervised learning and try to predict boards you're probably going to get roughly that level of performance of like a high level human designer even if you could get all of that data but with reinforcement learning you have the opportunity to go significantly better than humans, right?

15:19So this is the most famous example I always come back to is DeepMind playing Go, right, the AlphaGo problem. They actually started first by training on human data, and then creating agents that are based on human expert moves. And they got to a grandmaster level with that. But the best system today starts with no human data at all, it just learns how to play the game, you know, and it determines whether it wins or loses. And then that is what gets you to far, far, far superhuman levels nobody can match. So did you also start with like literally zero human data as input? Or did you have some and especially if you started with none, like, how do you get over the sparse reward problem?

16:01This seems to come up all the time. And I'm reminded also of the eureka I used to say about AI, I know Eureka moments, meaning like at least from the generalist systems, you wouldn't see them doing like legitimately new stuff better than human. With the Eureka project, actually, I started to have to say precious few Eureka moments because now there's at least like some examples that are starting to pop up. In that case, I'm sure you're well aware they use GPT-4 to write the reward function for the robot hand as it was like learning to do all these tasks where in the beginning, you know, it's. success is so fleeting or even non-existent that it's really hard to even score what it's doing.

16:44So I guess I'm curious, how did you get over the sparse reward problem, especially given how little data you had to go on at the beginning? Yeah, totally. So in general, the sparse reward problem kind of broadly stated is you're kind of a good reward. So winning the game of creating a circuit board is so rare that as you randomly explore, you never find it. Right. And that's an issue because if you never get a signal that you've won a game, how do you ever learn anything? Right. So the nice thing is that like people have this problem, right? Like people in general have broken up the problem of circuit board design into many different steps and have basically come up with heuristics along the way.

17:25So what quilter can do in terms of not running into the ultimate sparse reward problem of like design an entire board. And at the end, only if you get a yes from all the physics simulators that this board is going to work, do you get a one and otherwise you get a zero, right? What we do is just break it up into problems, right? So we take the first part of the problem, which is placing the components. You know, we can make sure that at that point it's manufacturable, the components don't collide, you know, things of that nature. And then compute some heuristics that basically indicate how likely we are to succeed at the next phase, which is the routing phase.

17:58And so on and so forth. similar things for actually meeting all of the physics constraints, right? Like you can simplify the problem to an extent to compute basic fast, dense rewards that correlate to your ultimate sparse reward. And that's what we have to do for now. Now, I will say that like long-term, the dream for this is definitely like a single sparse reward, you know, just a single yes or no, this board will work or not. And when we're really uprooting the typical patterns that humans use and are trying to do much better than them, that's what we'll have to do. But for now, since we're not as good as humans at layout, especially on more complicated boards, we can still use the same heuristics that they would use to just at least automate similar to what they would have done.

18:43Yeah, interesting. It sounds like, if I understand your comment correctly, the answer to this might be yes, but I was wondering if there's an analogy between the reinforcement learning processes that are used on the language models and these sort of evaluator systems, which I understand are like not models, right? They're just sort of either simulators or like checkers that are deterministic. But, you know, people worry a lot about in the context of language models, the idea that the human reward signal is not fully reliable, right? Like we're kind of inconsistent. We're sometimes mistaken. We, you know, they, it's been observed in many language models that there's a certain like sycophancy tendency where it seems to try to tell you what you want to hear versus the truth in some cases, because maybe that's what got higher reward in the training process.

19:33So it sounds like there is kind of a similar problem here where the checkers are, I guess I could imagine that they might be like purely physics simulators and could be like rock solid, or I could imagine that they're on a foundation of sort of a bunch of heuristics, which might in some subtle ways kind of also lead the process a bit astray. I, yeah, one thing, and obviously Sergey can talk about the specifics, but one thing that I've, this is an analogy that is imperfect in about 50 ways, but it's kind of supposed to, like, one of the things that I've been thinking about to try to articulate the Gen AI limitations, and particularly like the LLM and even stability models, like limitations versus versus some of these other techniques and what humans are actually good at is like, I think humans are very good at extrapolation.

20:32So like coming up with novel things and developing and pushing creativity. And if this, then that, like we can also do this and you can like take it and extrapolate. But because of the way these are trained and everything else, it feels like, and of course who knows but it feels like the a lot of the gen ai stuff is good as interpolation which is within the bounds of things we already know like what are other points like in that space right and it kind of makes sense if you think about it like it's they're interpolating like they're interpolating they're they're figuring out like what the next word is or what the next image could look like based on the training set that is like and pushed in a bunch of dimensions and humongous and everything else, but it's still like bound.

21:20It's bound by all the stuff that humans have prior created. And so like this is part of why I think the AI scare stuff is so misguided and not really because it's like, hey, if they're interpolating, like, okay, that seems fine. And it feels like humans will continue to be good at extrapolation, which is developing new novel techniques to, to do like various things. And like, that's a, that's a very important and valuable thing that we'll be able to continue to do. That was a very macro answer to your, your question. But like that, but I think it's a, I don't know, maybe, maybe it's a useful thing.

22:05Curious what you guys think. I have a ton of thoughts about that, but I think our subscribers have heard them in other contexts. It's, I kind of feel like all this stuff is converging. So that's why I was trying to make the connection between basically how much do we trust the reward signal in the context of reinforcement learning for language models. It's like, yeah, it certainly helped a lot, but not that much, I'd say, is kind of the consensus answer. I wonder what the situation is in the context of circuit boards. And then we could maybe also speculate about a similar question when it comes to self-play.

22:47There, it seems like the sort of narrow problem, not just one domain, but in general, narrow problems are better suited to self-play for now. But we are also starting to see some of these self-play techniques be applied to language models. And that's where I'm also like, I don't know if it's going to stay in the bounds of what humans have given it for all that much longer. But let's take it piece by piece. So Sergey, let's start with the, you know, how much do you trust the reward signal in the context of your problem? Yeah, so one of the nice things about working on a hard physics problem, right, like humans are not in the loop.

23:29Like humans are actually bad at looking at a board and judging if it's going to work or not. That's where like 80 % of boards that are built are faulty in some way. And it's because humans are just not good at that task fundamentally. But all of the core physics is computable, right? Like we care about laws of the Maxwell equations and laws of thermodynamics. And we've had convergent techniques to solve those for over 100 years. So in our case, like if you actually use the oracle, right, like actually numerically solving the Maxwell equations for every set of possible considerations and problems, and you're kind of being careful to make sure that you're convergent and approximate everything correctly, it's completely trustworthy, like much more trustworthy than the human result by far.

24:12Only problem with that is speed, right? So if it takes 20 minutes to compute a single, you know, a physics solution, and you need to do millions of those as you fine tune on your model, like that's problematic. So what you can do is you can make approximate models that are just conservative. And so at that point, maybe you're not doing like the best possible arrangement that physics could allow, but you're still doing one that works, still doing one that competes or beats humans. And it's still definitely going to work because you've been careful about the physics approximations you've used. So I think that's one of the luxuries that we have is like, we don't have to negotiate with humans, right?

24:46We just have to make sure that it's manufacturable and that it's going to work. And physics tells us the answer and physics is unambiguous about that. Gotcha. So, and these were tools that existed already in the industry that you're able to build on top of. Yeah. When the system is going about the process of designing a board, how iterative and sort of tree searchy is that process? Because I think folks, again, will know at least the basics of AlphaGo, right? And these are like hyper parameters, ultimately, that you can sort of turn up or down at runtime, right? But how deep are you going to allow a system like AlphaGo to search the space of possible moves is a huge factor in how well it's going to do.

25:35If you just make it do a single raw guess, it's not going to do super well. If you allow it to map out a bunch of possibilities and then get scores on all those possibilities, then it can sometimes land on superhuman results. So give us a sense for kind of what that search iteration and scoring loop looks like in your context. Yeah, I think actually this is somewhere where a user has meaningful choice. So for a lot of designs, like you just want it fast, right? Like you're maybe primarily focused on making sure that your schematic is okay. You don't care if the board is rather sparse, you know, big, not very dense, maybe it's a little more expensive to manufacture, so on and so forth.

26:18And in that case, like you don't really want a whole lot of search. You just want the first answer that's going to faithfully implement the schematic on the board. And so in that case, the ideal thing is to return an answer within a few minutes to an hour. On the other hand, suppose you're looking at a board that's going to have 100 million units produced, like an iPhone leather board or something like that. And if you save a cent or a dollar on every board, that's a lot of value. In that case, you might let it search for a month and explore in all sorts of directions because that's how long it would have taken a human to do it anyway for a single design.

26:51Never mind for the billions and billions and billions you could explore with this kind of system. So this isn't like a lever that we have in the tool today. we kind of just we basically treat overnight as the constraint for us right now right so the idea is that at the end of the day you finish your schematic you uploaded it and either immediately or sometime within the next 12 hours you get a result and that morning you know the next morning you can look at it and see if it's if it's up to your standards but in the future i see that being a lever and it's for you to decide what is more important to you yeah that makes sense what is the compute, like what are the size of the models?

27:28What does the compute look like and how much of it is on the inference side versus on the scoring side in the physics simulations? Yeah, those are all really good questions. So like, obviously, we're not dealing with the kind of compute that LLMs deal with, right? Like we're not using 10 ,000 GPUs to train across a trillion tokens and whatever crazy numbers are being used nowadays. You know, this is a relatively focused problem. And so, you know, we can deal with much smaller models, we can deal with, you know, much faster convergence times, much less compute effectively. So the right now, the vast majority of the computer is going into kind of the, you know, like actually playing the game.