How Sierra Is Pulling Ahead in the AI Race | Co-founder Bret Taylor

23:49Bret Taylor:And you've got enough demand for all these tokens. And I think the Moore's Law characteristics of, you know, these models will drive token costs down because there's every incentive to do so. and also scientific breakthroughs that facilitate it, just like the transistor in the early days of Moore's law. The interesting thing is the way I see the models playing out is a little bit more of a heterogeneous set of models for applied AI applications. I think if you look at the pursuit of AGI, you end up using a lot of tokens because you do both that training, but also inference time with these reasoning models, you're using a ton of tokens, but producing remarkable amounts of intelligence.

24:37Bret Taylor:In the more applied AI world, you're seeing this with codecs and modern versions of Claude, where they're actually thinking for long periods of time, consuming a lot of tokens, but writing extremely high quality software. So, you know, it's not like the actual, you know, iterative, like each iteration of using codecs, like the cost hasn't gone down, it's probably gone up, but you're actually building software you couldn't do before. So it's quite valuable. And then the other end of the market, you might be using, let's say you're making an LLM augmented thing to detect credit card fraud. Well, you can't, you don't want to use a reasoning model for that.

25:10Bret Taylor:It's too expensive and high frequency for that. So you might, that you probably want, you care about latency and other things. If you're doing voice AI, you care a ton about latency. And so my sense is we're going to end up with a, we use the term internally at Sierra, a constellation of models where you have different price performance, which is both latency and throughput, and different applications are sensitive to both quality. And you're going to trade off all those different levers for your application. And if there's an analog in the pre-LM world, there's probably databases. You know, like I don't think there's a database you use for everything, right?

25:46Bret Taylor:If you want a trillion data points and you want to do analytics on it, you'll use one columnar data store type thing. If you want a transactional database for credit card transactions, you'll use something different. And if you want to, you know, eventually consistent cash, you'll use a different thing. And engineers over the past three decades of building cloud applications have pretty much a shared set of rules of thumb to figure out which data storage technology you use for different applications. I think that's the way models will play out over time. And I think we'll sort of grow up in, you know, five years and there'll be a lot of different things that we call large language models today, but sort of in that heritage of model, probably with new advancements, just like the reasoning models.

26:33Bret Taylor:And there'll be a lot used in concert to power an agent. And I think that'll be a really healthy progress in the ecosystem. Earlier, you mentioned, hey, like a market correction might be good here. Let's like play fake economists for a second. Like we don't know. I'm curious if you think that will come from, I don't know, like models slowing down progress in air quotes. Um, if that will come from all of a sudden public SAS comps are down 50%. And so people start getting jittery. Like, I'm curious if you're like, as you read the tea leaves, obviously you have to think about some of this because you have to capitalize the business as a byproduct of what you see coming down the pike?

27:15How are you indexing the risks here?

27:19Bret Taylor:I think it probably will happen. I think, as you said, the juxtaposition of multiples of software stocks being lower than they have been and the premium on AI investments, which are largely private, creates this awkwardness, which is like, what happens when these companies go public? Do the valuations settle down? And so, you know, you could end up where a world where, you know, as some of some, I'll quote unquote, AI companies, and that definition is an interesting one, just because I'm not sure exactly what it is. Like clearly open AI is one. Right. There's everyone else is striving to be one.

27:59You mean it doesn't count if you put dot AI at the end of your, that doesn't count?

28:02Bret Taylor:It could be that you end up where a few of these nominally AI companies, you know, enter public markets and don't get the reception and that trickles down to private markets, late stage and then early stage. Maybe the more likely one is macro, though, you know, when interest rates went up because of inflation, you know, after the pandemic, that corrected a lot of the software market. And so, you know, it does feel like these macro forces are greater than any sort of insular tech specific thing. Um, and you know, we'll generally make, you know, uh, just change the way people invest. I'm not sure what, which will cause it, but it does feel like, um, I don't think it's unhealthy by the way.

28:47Bret Taylor:I mean, just, just say like, I'm a capitalist. So like, I think it's really healthy that there's lots of companies going after these problems because it enables the market to kind of like do more exploration of ideas. Um, so it's, But my sense is it just can't go on forever just because you can't have 20 companies going after the same space. There's just too much sort of saturation in the market. But I think there will be some form of correction there and then some form of consolidation. And I think it will be fine. I think we'll survive it and it will be fine. And actually, the economic impact of AI will be viewed probably with as much optimism as anyone has right now.

29:29Bret Taylor:I it's truly changing. We can feel it deeply in software engineering and customer service. It's completely transformed those industries. And for those of us in the middle of it, you can just see the capabilities of the models because you you feel it, you know, you feel like the difference between using codex today compared to three months ago and you see what it can do differently. And it's easier if you're in the middle of one of those industries to extrapolate to all the others. but I just, I do think it will have a huge, huge impact. I'm sure you read that article, I guess it was yesterday that came out.

30:05It was like, do you remember the name of the article?

30:08Bret Taylor:I don't, but I did read it. Okay, yeah, it was like, how would I characterize it? I would say like sounding the alarm bell of the change that is to come through the lens of coding as the like models are improving. Basically, like the author was arguing that people, no matter how bullish you may be, you're still underestimating how quickly this rate of change is. And as most like primarily evidenced by what's happening in coding today with how good the models are getting where, you know, like two years ago, you could barely write, I think like, you know, it could barely write a line of code. And now it's like writing like thousands of lines of code autonomously well, like really well, like full code bases and then checking that code and it's really good.

30:56And, you know, I think the point of the article was like, hey, for people outside of the valley that don't really aren't living and breathing it like we are, like you have no idea, basically. And I kind of viewed it as like there was definitely a dystopian lens to it. Would you agree with that? Like it was somewhat just hard to tell.

31:16Bret Taylor:It sort of ended optimistic, but it was. Yeah, I sort of said I agree. It had a kind of anxious dystopian lens to it. I agree. Yeah. Like the world's going to change and we're not ready for it. And, you know, it's interesting, like, I think every, what's crazy is like every three to six months, I feel like we, the pendulum swings from like, the models aren't improving to nobody's going to have a job anymore. Like, it just keeps going back and forth between those two things. Like it's overrated and nobody cares about AI anymore to like, we're all going to sit on the beach and have Mai Tais one day.

31:54Yeah. And that's like a pretty, like, I think if you don't come from, I think if you come from our world, it's still like a pretty heavy feeling. I can't imagine basically the rest of the world, how they feel just reading these things from the outside in. And Sam at OpenAI also talks about this in, and like Dario, like these guys are also talking about this, like this, um, uh, like almost change management of society that is coming. And, um, you know, in some ways, like there, I think everybody's right in that, like, we are like, things are going to be like drastically different, but in other ways, I'm like, you know, the pendulum has definitely swung into like, we're all going to be on the beach in Mai Tais.

32:41I'm curious, like, I like, where do you follow? Yeah.

32:44Bret Taylor:Like I'm not in camp Mai Tai, but I also, not in the other camp either, I find, I don't, I question the premise of the two extremes. I'll give you my perspective. So first, part of the reason I think you have the juxtaposition of the models aren't changing, and oh my God, everything has changed, is how much intelligence you need for the task. So if you're using ChatGPT to plan your vacation, and you use GPT-40 and use GPT-5.3. My guess is the experience won't be that different because planning a vacation on Chachibiki doesn't require a huge amount of intelligence to do. It was already, you had to reach sort of sufficient quality, you know, a year or two years ago that actually, like everything is sort of feels like incremental on top of it.

33:33Bret Taylor:In contrast, if you're using the model to write a Rust module that is going to do something low level systems software with like high sensitivity around both correctness and latency, the models were like woefully, like they just could not do it two years ago. And all of a sudden, over the past three months, they could. I think what's ended up happening is if you end up, I don't think it's really possible to do, but if you're just like, take every task you do in your life and sort of order them by how much intelligence is required to do them, we're sort of moving down that line. And if you're testing something that we'd already, like the horizon had already passed, you're like, nothing's changed.

34:12Bret Taylor:And that's actually people's lived experience with ChatGPT. A lot of it is these sort of simple, and they're useful, right? But we already solved that. So now it's a commodity. And then if you're trying to do drug discovery or software engineering, you're seeing it change on a daily basis. That's why I think there's a juxtaposition. And put it another way, one interesting sort of thought exercise is we've reached sufficient intelligence for a lot of tasks. Put it out of the way, if we paused innovation and just absorbed the intelligence of all the existing models, my guess is there's still trillions of dollars of economic value we haven't realized yet, which is interesting unto itself.

34:49Bret Taylor:I think it's really interesting. Then you go to like the what's going to happen. I think a couple of things are true. I'm an optimist in all of this. The reason why I don't, I'm not in camp Mai Tai is I just don't believe that humans will stop doing things. I think we've, uh, through automation, like eliminated a lot of jobs from the jobs of a bank teller to a lot of agricultural jobs, which have, you know, I think it's like what 5 % of our, uh, jobs are in agriculture today compared to like 95%, you know, a few hundred years ago. And so just because we automate jobs, I think we often lack the imagination to imagine what comes on the other side of it.

35:30Bret Taylor:And I firmly believe that. I just think that we are, I believe this technology is fundamentally a tool, even though it's a really impressive and super intelligent tool, but it's fundamentally a tool. And we will create an economy and identity around it. And I don't believe that because it took something we're doing now, it's taking away our identity and we want to go sit on a beach. I just don't believe that as a human being. The other thing, though, to temper the excitement is software engineering is both getting the most attention right now because the research labs know to create self-improvement for AGI, an automated AI researcher is a prerequisite.

36:11Bret Taylor:So as a consequence, all the smartest research labs are specifically working on this. So it can write the next series of models. That's right. And if you saw the OpenAI post on 5.3, there was a lot of openness about sort of using the model to help build the model, which I thought was really interesting and what's happening in all the labs right now. So first, just because software engineering is getting great, the idea that that's completely general is not, it may be true, but it's not obviously true. There's a lot, how much this generalizes the G in AGI is the hard part. And it certainly generalizes, but how broadly is an interesting question.

36:50Bret Taylor:The second thing is like which parts of the economy can basically absorb intelligence completely fluidly. Software engineering is definitely one of them. It's a purely digital profession, right? You write code, you compile code, you produce binaries. And so the entire process is digital and you can test it digitally. Contrast that to a drug discovery. it might work until you need a wet lab and then you need a wet lab and all of a sudden you need robots okay then let's say you did robots in the wet lab then you want to bring it to market well you need a clinical trial and that clinical trial even if you have great ideas on how to do it better it's still a process that is regulated by the government like there's no amount of like intelligence you can just like pour onto that process and make it run faster and so the interesting thing I see is like clearly finance and software engineering are professions where you can take intelligence and sort of absorb it into the profession quite efficiently, you know, because a lot of the job is essentially operating in the world of information.

37:52Bret Taylor:And then you say, okay, like, what are all the other parts of the economy? And not only are we absurdly focused on coding agents just because of its importance to the AGI labs, but also it's not obvious to me that every other profession is as, it's as easy to take a information oriented digital agent and transform the profession. So I think it's gonna have a really big impact. But I think the, the takeoff in the AI labs might be faster than the takeoff in society to some degree, just because different, different professions will be sort of differently impacted by it. And it's useful. I actually have started doing this, which is like, I'll walk around and just think about like, okay, if we had like super intelligence by any measure, you know, what would happen to that, that flower shop right there?

38:46Bret Taylor:You know, like what would change about it? And it's really interesting to simulate that because I do think it sort of shows both the short-term opportunities and like the complexity of, you know, how this technology gets rolled out around the globe. Super interesting. And then, you know, a lot of what the labs are saying right now is, you know, we're not going to be hiring engineers in the future. Like, there's this big take from the heads of the labs that they're going to reduce their own headcount. You know, Sierra seems to be growing. Like, there seems to be a lot of people that you're hiring.

39:18Like, you know, it seems like you're still hiring a bunch of software engineers and salespeople. Like, it seems to me from the outside in that you're building a high growth company and staffing it with the resources that I would have seen for any company growing of this size.

39:36Bret Taylor:Square that. I think it's a correct critique. I always laugh. If you look at Anthropic and OpenAI, they don't look that different than a high growth company that proceeded it, I would argue. Sam and Dario might disagree, but they're not fundamentally different. It's hard to know, though, because on the other hand, OpenAI's revenue scale is truly unprecedented, there's no way to A, B test what you would have needed, you know, preceding it. So maybe that's an unfair comparison. I think one of the things just to state it is these coding agents have only gotten to this level of quality literally over the past few months.

40:17Bret Taylor:So these companies have been built over the past years, not months. And so one of the, I think the, perhaps an interesting question to ask is with the current technology, what shape would you want your company to be maybe uh as a proportion of your revenue or users or whatever the right you know uh proportionality might be and i think one could argue that if you project out the skill set you may want a different shape in the future so one reason may be that actually we just hadn't gotten to the point where the kind of new shape of the company were really possible, but for the smart leaders of these labs, they're projecting forward.

40:58Bret Taylor:And it might be right. Going back to my, you know, bank teller example, though, you know, when the banks automated distributing cash with an ATM, they decided to keep their branches and keep employees in those branches, but have them do higher value things. And so we live in a, well, hopefully in a free market that's very competitive. And so clearly the people who are software engineers won't be doing that. They're already not doing the same things as they were a year ago. Like they're operating, you know, Claude and Codex, not typing. And just nine months ago, they were in Cursor. And like nine months before that, they were in VS Code.

41:37Bret Taylor:So like every, you know, it's incredible how rapidly it changes. But will you want the same number of people? I don't know. The way I'd answer that question is, does having more people enable you to gain more market share than your competitors? And if the answer is yes, I think the answer will be yes, because you don't just like recoup efficiencies in your business and pass them on to shareholders, because every one of your competitors has access to the exact same technology. And so what's going to happen is the second order of fact is, if you assume every company in a market has access to this technology, which is truly democratizing, you know, what does every company do with those like that new higher leverage operating model.

42:19Bret Taylor:And that's what's exciting about it. And this is why I find a lot of people's projections about jobs to be simplistic, because let's just take the mobile phone market in the United States. So you have Verizon, T-Mobile, AT &T, all fighting for the same, how many mobile subscribers are in the US? I don't know, 300 million? I don't know what it is. It's a fixed pie. So like, you know, just assume all of them have access to the same AI technology to improve their business. Well, then the question is like, okay, they're not going to pass that on as cost savings. I mean, they can, but if one of them lowers prices, the other one will have to lower prices.

42:56Bret Taylor:If one of them finds a new interesting channel for customer acquisition, the other person will. And so I think the key in all of this is you in a competitive market with a technology like this, everyone's going to absorb the impacts of the technology and then compete. It's a little bit like, to some degree, if we went back to 1995 and I were trying to sell you the prospect of making a website, and I said, if you make this website, you'll be able to do X and Y and Z. And if I ended that way, then none of your competitors will do that. It would be a lie. In fact, for the benefit of hindsight, the correct sales pit for a website is you should build a website, X, Y, and Z.

43:34Bret Taylor:Because if you don't, all of your competitors will. And here's what's going to happen to your business, which will be not have access to these new digital channels and search and all the other demand generation, all these things. And all of a sudden, it becomes an imperative. So that's kind of how I view AI is it's sort of a strategic opportunity, but it's actually more of an imperative. If every software company in the world can produce software at a marginal cost that's like much lower than you can, you're at a disadvantage. And I think because everyone is going to do it, it's going to play out in the second order of factors where all the job creation will be, but also the interesting competition.

44:14Bret Taylor:And it's fascinating. And I think it's very hard to imagine just because, you know, if people knew people would be doing it already. It's like you kind of need this to wash over the market and then, you know, have these companies compete. and I think it will surprise us what comes of that competition. When you think about capitalizing the business, like I'm really curious, like is the last valuation public? Is that out? Have you guys? The 10 billion. 10 billion, okay. And so we know your ARR roughly. We know about when you did the deal. So it was like 100-ish million at 10 billion. Okay, those revenue multiples do not resemble the public market comps.

44:56It's high. It's expensive. Why? You can probably raise even higher, but let's just assume that's a pretty high clip for you to raise at. Why even capitalize the business so much right now? Why raise at a higher mark? I'm just curious how you think about the strategy of what capital and valuations means as an edge to your business?

45:27Bret Taylor:There's a lot to the why, because you're probably saying why I did it and why maybe our investors chose to do it. Actually, less the latter. Yeah, why we did it. Yeah, why you did it. We aspire at CIRA to be the platform of choice for every company in the world, in particular, the largest companies of the world, when they're thinking about using an AI agent for their customer experience. So, you know, whether it's, you know, healthcare companies like Blue Shield of California and Cigna or revenue cycle management companies like R1 or telecommunications companies like DirecTV, SiriusXM, or the banks that we work with or the fintechs that we work with or the insurance companies we work with, we want to be the default.

46:09We want to be the company that you

46:12Bret Taylor:can partner with that will enable you to go live, get success, and we want to be your first phone call. And I think that requires a lot. One is, I think part of the reason we've grown so much is we have by far the best sort of customer base. And so when people are saying, they look around and say, I'd like to use a partner that companies I respect use, we're in that. It also requires scale. So we opened an office in London, we opened an office in Singapore, we opened an office in Tokyo, you know, you need to be present in these markets to work with the clients who are present in those markets.

46:47Bret Taylor:And so I just don't think, you know, being so lean, you know, focusing so much on that is really the right way to win in this market. We want to, we use the term face in the place, like we want to be next to where our customers are. We work with, you know, one of the large Spanish banks, we need to be present in Madrid to partner with them. You know, we work with one of the Southeast Asians telcos. And you don't want to have to wake up in the middle of the night to have that phone call. You need to be present there. And then similarly, I want to make sure we're scaling our product as well. I think I always admire sort of what the rippling guys talk about this, but like your pace of innovation is actually almost important than your product, right?

47:29Bret Taylor:Because in a world where AI is changing so rapidly, your roadmap matters a lot, right? Because you could have the state of the art today and 12 months from now have something that looks pretty archaic. And, you know, our goal is to have like the fastest pace of innovation as well. So we want to capitalize the investment so we have the fastest pace of innovation. We can grow our customer base the fastest and be present in all these markets. And we just need to grow up pretty quickly to do that. And it's in part because if you are one of the largest banks in the world, you want to know your partner is going to exist in 10 years.

48:00Bret Taylor:You know, you want to know that you're not the first, you know, structurally important bank to go live on this platform. Or, you know, second or third is probably five and not first. So that's and that just requires capital and growth and maturation. And that's what we're investing in. That makes sense. And then on the valuation piece, like, do you have a philosophy on like, just take the highest valuation so you can have the least amount of dilution? Like, I'm curious, like, you know, the let's take the market correction example. Like, let's imagine that it happens this year and it happens meaningfully.

48:33the counter argument to raising whatever 100 at 10 billion is i'm just making up numbers but just for the story's sake um you you just have a proverbial gun to your head of a really big target that you have to hit at a valuation that was quite frothy at that given moment i'm so like i'm more just like philosophically curious about how you think about it um we definitely

48:57Bret Taylor:don't choose the highest valuation in fact in all three of our rounds we uh had higher valuations available that we didn't take. So factually, we definitely don't do that. We have a dilution that we care about, but within that range, we're choosing a partner and that's how we think about it. But it is a generous valuation. The way I think about it is actually quite simple, which is how much capital do we need to grow into the next milestone, which might end in entering the public markets. But between that, it would be to have a revenue scale and growth rate to capitalize the business in the next round in a way that all of the existing cap table is happy with that outcome.

49:42Bret Taylor:And so the way I think about it is, you know, if you take the$10 billion valuation scale, what does it mean to sort of fill in that valuation? What revenue scale and growth rate? How much capital do we need to achieve that? That's all it is. And I think the, so we're very much focused on like our business plan, which is like, how are we going to invest in, uh, you know, the products and, uh, you know, our go-to-market teams to grow, uh, grow our revenue. And it's a sort of a spreadsheet, to be honest with you, that's how we think about it and the unknowns. Uh, but we just put a lot of, you know, air bands around it are, you know, the demand environment and competition.

50:21Bret Taylor:Um, though, you know, when you're a company like ours, you just assume you're going to, you know, remain the best product on the market, which is not something you're entitled to, but like, that's why you exist. And then some of the things like there's a lot of unknowns around we've talked about, like, well, how many software engineers will we need to use? I don't know. I have no idea. But with usually what these things is you try to plan somewhat conservatively, because the last thing you want to do is under capitalize your business and then be in an environment where you have to raise money on terms that aren't favorable, which is, you know, happened a lot, but you don't want that to happen to you.

50:53Bret Taylor:So we tend to sort of, you know, you put a bunch of padding in on different assumptions and run some contingency plans and all those other things and just make sure that you have enough capital to get to where you want to go to. And by the way, the reason why, you know, 100 does some, but most companies aren't going 50 % quarter over quarter either. You know, so that's the, that's the dynamic. And so I'm not sure it's right or wrong, but I would invest. No judgment passed. I'm more just curious. one of uh i i incubated a company here with mamoon um who's uh on the board with me and um we came out of stealth and um there was a lot of investor demand and you know we like our things are working let's put it that way and i really wanted to bet on call it like flipping over one more card or two because there was like giant multi-million deals dollar deals that i was like pretty confident we're going to get done and you know like i could kind of calibrate accordingly and um i don't think you'd mind me sharing this uh ilia here at kp like he had heard about like the you know kind of that there was um a lot of demand and uh he was like you know there's an old kleiner law like i guess there's kleiner laws and he was like one of the kleiner laws is um uh when the appetizers are passed take an appetizer and that stuck with me you know like uh that stuck with me so maybe there's just something to it like when that's the thing is macro matters a lot like there are a lot of companies and whenever interest rates are growing up that maybe could have been healthier than they were but just didn't have capital available as well um so i that's the thing that is always like you know whether i mean i my first company the bank banking crisis happened in the middle and so you know it turned out fine but it just it changes markets and so like you know right now there's just a lot of volatility in the world.

52:42Bret Taylor:And so, yeah, when the appetizers are served, eat. That's right.

52:50You guys do paid POCs. Is that right? Maybe the question behind the question is that every CIO, every large company, maybe it's not the same voracious demand as a year and a half, two years ago, where their boards are telling their CEOs that they have to go use AI, but they are looking for wins. And then that like trickles all the way down. Right. And then what do they do? They like go online or they start asking their peers. And then they usually do like a bottoms up market first. Like what are the good markets? Okay. Like what are the projects that we should start in? Then they'll usually start in coding.

53:27They probably already doing something with Microsoft. Then they'll make their way to, you know, whatever windsurf, and then they'll go straight to Claude or Codex or whatever, right? Then they'll probably look at legal, you know,

53:39Bret Taylor:and then they'll probably look at support, right? So like, there's like a few buckets. Yeah. And you're in one of those buckets. Okay. And the good news is like, you're in one of those buckets. And so you're like, you have more demand than supply in some ways. The bad news is, I don't know, but like, I'm just parroting this back to you. There's probably a lot of tire kicking. Like there's probably like, it's not that your software can't scale. It's like you have a forward deployed engineering model where I suspect the reason you do it is because the underlying technology is nascent. The LLMs are nascent.

54:17Sierra is early and you want to, it's prerogative on your, it's your prerogative to basically give these large enterprises, is a bear hug, a Brett Taylor sized bear hug to make sure that like they are successful. And so you want to invest a bunch of resources in making sure that they're successful. And in order to do that, like you need great people that can partner with them. And so like maybe not to put words in your mouth, but I'm curious, like, do you then have to raise the bar of like when you are supply, not demand constrained, who you choose to work with and how those engagements go? Because that's a very caviar, but interesting problem right now.

54:57Bret Taylor:Yeah, I think you're right. So just talking about the kicking times, we have a name for it. We call it AI tourism. And so there are companies often, as you said, due to board and CEO pressure where you're trying to show AI momentum, but without necessarily like a business mandate. And that's where you have, there was that MIT study that got sent around like everywhere about failed AI projects. I think a lot of that is AI tourism where people sort of start ostensibly POCs, but they're essentially, they're like POCs marching into the void. Like they never were, there was no path to production for any of them, but they were, you know, showing motion or trying to learn about the technology.

55:43Bret Taylor:and so we with our proof of concept uh more often than not you know we're essentially always paid and the reason for that is we want companies that are serious about doing this we don't want you know the companies that are doing air tourism um and uh but it's modest it's a little bit more just like hey like you're you're for real you know you actually want to to deploy this project at the end. Going to your point on the bear hug, you know, more often than not, our product can fix itself and it's pretty easy to use. So the forward deployed part of it is different than just being, having a seat at the table with our clients.

56:24Bret Taylor:Sometimes it's change management. We work with one large scale medical device company that had 40 call centers and they're consolidating into to one with AI. Some of it's technical, right? There's 40 different stacks and that's where sort of a, maybe a forward deployed motion helps, but they also need help with that transformation as well. And so, you know, part of the benefit we provide at Sierra is we can, whether you're small or big, by the way, it's not really a size thing, but the hard part of AI is like, how do you successfully deploy it? And some of it's human, some of it's technical. And what we want to do is actually be accountable for that outcome.

57:01Bret Taylor:It's why we have outcomes-based pricing. It's why we have, we call it agent development, that sort of forward deployed team. But it's not, you know, we can, we have a lot of clients who's like, hey, I want to do this all myself. And that's great. And we'll just say, here's the product, go to town. We have some who really need a lot of help, but our best are sort of, you know, co-innovation, you know, we're there helping you consult, if you will, on sort of like the right way to do it. And I think it's the way a lot of companies want to work in AI because they want to know best practices. You know, they want to know, hey, I noticed, you know, you went live with this healthcare insurance company in two months.

57:38Bret Taylor:And that's really impressive. That's almost unbelievable. How did that tell me how to do that? That's actually kind of the thing that we can do at CIRA that's really unique is we can kind of come in and say, we have experience with the largest banks. We have experience with the largest healthcare companies. We have experience with the world's largest telcos. And if you want to move faster, we can help you do that. And that's not always technology. It's often a lot of other things. And so that's really been our sweet spot. So, and it does though, it is a higher touch model, if you will, and it's not all engineering, but what we want to be as a partner, not a vendor.

58:13Bret Taylor:That's how we want to show up with our clients. Yeah, that makes sense. Can I ask like personally, like one of the things that I was thinking about this morning in anticipation of this conversation, I'm like, you know, given your pedigree, both like with what you're doing in open AI and everything else that we know, and obviously what you're doing today at Sierra, you're probably getting hit up on like, it's gotta be like five to 10 cool things a day. Like, I don't know how else to say it. Like there's, and some of them are customer things, which I suspect you prioritize. Then some of them are like, should I go talk to Jubin today?

58:47You know, on grit and whatever, do that. Right. And then some of them, I said, yes, Yeah, you did. Twice now. Thank you. Some of them are, you know, probably like the other founder dinners and stuff. Right. Some of them are like, go give a talk somewhere. Like the list goes on. Some of them are like fly you to some awesome place and go skiing with investors. Whatever it is. Right. Like it starts to add up. Like, uh, uh, have you had to develop the no motion? Like, uh, it's a, again, it's kind of a caviar problem. It's the same problem as like too much demand at Sierra. Right. But like, there's a lot of demand on your time and you have a wife and, and kids at home.

59:27Like, like, uh, tell me about the time management piece. I'm very curious about that.

59:32Bret Taylor:Uh, first I work a lot. Um, I, I love to work. Um, so I mean, I guess all the time. I mean, I, I think you know this as a founder, you know, if you're in bed and your eyes are closed, you're still thinking about work. I mean, at least I am. But I enjoy it. It doesn't mean it's always easy, but I just love what I do. And I love Siri and I love OpenAI. So I just really enjoy it. And I do put in a lot of effort into it and a lot of time into it. And I don't know, it's fun. And I, but also like my, my wife likes, like we're like, we're, this is sort of, she knew she was married and like, we love talking about it.

1:00:16Bret Taylor:It's great. So I, I do put in a lot of time. I do try to prioritize my time. I try to spend basically time on our product and technology and time with our clients. And I try to do as little other things as possible. And the idea being that I think I love spending time with our, our customers, our partners. um a i learn a lot um you know i always joke you know if i'm talking to a banking ceo or a banking cio they've forgotten more about banking than i will ever know um i know a bit about ai but to actually have the combination of sierra and that that client produce something great is like the the two of us together and that requires a relationship and like deeply listening you know truly being a partner and you know when i was talking about that sphere and the surface of the sphere versus the center of the sphere.

1:01:06Bret Taylor:Like I try to live, live on the surface and really spend time with our customers. And then I try to spend all the rest of my time on product and engineering. And I'm fortunate enough to have both an amazing co-founder and Clay Bivore and also just an amazing team. So I have the luxury of spending time on, you know, product and engineering and spending time with clients and more or less keep my calendar to those things and the occasional podcast that makes sense yeah um you use the word fun uh to describe like how you feel um people ask me all the time like uh how am i enjoying you know like building a company and doing doing all this stuff um i like you mentioned rippling earlier like i like parker and mckinnis's metaphor which is like it feels like playing my favorite sport that's how i feel i love the way parker talks about like i know parker like not that well but like every time he talks i'm like yeah he's one totally 100 we're in the same like he's a sicko for the game yeah complete sicko for the game i love that guy he's great and i like the way that he uses that metaphor because i feel the same way but i would not describe even if i'm playing my same my favorite sport as fun the way that i would describe it is like i feel deeply fulfilled and the the reason i don't say fun is like a different version of fun for me is like, I don't know, going and playing golf or hanging out with friends or doing other things.

1:02:28And those days are like long gone at this point. But I would never, I never long for doing that version of fun, if that makes sense, because I can't imagine doing anything else. It is what I want to be doing. And it is all consuming because it is extremely fulfilling. But because it's also all consuming, I wouldn't necessarily describe it as fun because sometimes at two in the morning when my eyes are closed, I would like to actually be asleep. You know what

1:02:54Bret Taylor:I mean? You're right. Fun is too simplistic a word for it. I actually agree with every word that you said. And I feel almost exactly like no notes. Like that is also how I feel. But there are moments of fun. I mean, that's the interesting thing about sport. Like, you know, yeah, I'm a huge Niners fan. And, you know, it was fun when we were winning games when we shouldn't have. And it was really painful to lose against the Seahawks in the divisional round. But, and that wasn't a fun moment. That was a tough moment. I can't even imagine if you're Christian McCaffrey and you've been like burning the kettle at both ends, you get to that point, probably didn't feel fun, but he did probably feel fulfilled.

1:03:30Bret Taylor:And I think the NFL honors honored that. Like, yeah, I think he deservedly won comeback player of the year. And, uh, and so kind of similar, you know, where I think the interesting thing about to be an entrepreneur is you do feel every bump in the road. I think you, I generally, I hate to lose more than I like to win. I think that is something that's pretty typical of most entrepreneurs that I know, which makes the lows feel particularly low. You know, if you lose a deal or you, you know, something goes wrong, your candidate says no, whatever it might be like it, it haunts you. Which is why, you know, the, I think it's hard to be not, I think the intensity in that are like two strands of DNA that are really intertwined.

1:04:13Bret Taylor:But you also have those moments of true fulfillment, but also fun that go along with it. And one of the things I would say, I'm a huge believer in having a co-founder. It's not sort of a philosophical thing on solo founders not working. I don't mean that at all. I think those generally speaking, I don't agree with rules, like there's rules of thumb or lack first principles thinking in my general view. I think it's just hard to be a founder by yourself and like clay and I do everything together um, like internally we're clay and brett. There's not like one of us separately and like having that partnership makes it more fun because in those moments where you're about to go too low The other person lifts you up and when you have those moments of great things happening You have someone you can share it with and it's actually why I like i'm a huge believer in marriage I'm a huge believer in co-founders like life is hard It's really great to do it with someone that you care about.

1:05:07Bret Taylor:And what's nice about a spouse or a co-founder is there's something sort of unconditional about it because you're literally kind of in it together. Like it's either going to work or it's not going to work. And so for me, I would say like clay makes it fun, but you're right. It's more fulfilling than fun. That's like, it's complicated. Yeah. I, um, one of my other observations that I've had in this same vein, um, that you just kind of made me think about with the co-founders thing. I have two co-founders and both in them, but also in my team, the trait that I have come to value almost more than anything that has surprised me is default optimism.

1:05:48Bret Taylor:You can't handle the pessimism. I can't handle it. Yeah. And with them, it's like true in spades. Like they just, they've never met a problem that they don't think they can solve. Yeah. And there is I think like we all feel this way across like the three of us. And I think it permeates through the company. Like I think when you are a default optimist, it starts to imbue a sense of inevitability into the organization. And I think that's really healthy because it builds momentum towards an end state that is promising. And momentum does to me feel like the oxygen of at least Roadrunner. Maybe you feel the same way at Sierra.

1:06:27It just feels really critical to continue to stack wins. And now I've even employees, I have started to crave that value of of optimism. The venture version of this of this world is the expression that pessimists sound smart and optimists make money and applied to startups. Like it's not about making money, but like there's something to that. And anyway, like maybe one of my most unexpected things that I look for now is that. I think it's, I agree with you.

1:07:01Bret Taylor:And it's hard because it's sort of uncool as a founder to, you know, show up and be like, I'm freaking out, you know, because as a founder, you're, you have to essentially manage a lot of stakeholders. Like you have to, whether it's your employees, your customers, your investors, your, the public writ large and their opinion of your, your firm, there's a, it's part of why there's sort of that distasteful kind of like hustle culture and performative stuff on social media is people need to act like everything's great. And it's so obviously bullshit, but people do it anyway. But it's partly because you're trying to convince yourself and try to convince everyone else.

1:07:40Bret Taylor:The thing I'd say is I agree with the optimism, but I also think you have to be careful not to bleed into inauthentic optimism. You know, So we have a philosophy at CIRA related to our value of craftsmanship, which is basically the idea is we fail as a team. So in engineering, there's this idea of a root cause analysis when a system goes down. And the spirit of an RCA, a good RCA, is you don't blame people. Because the whole point of a well-engineered system is it should be impossible for a human operator to accidentally trip over the wire and have the whole thing go down. So for example, if you push new code and it failed in production, it's not the person who wrote the bad code or even the person who pushes.

1:08:25Bret Taylor:Like, why wasn't there a test? You know, if it took the whole service down, why didn't we roll it out to a set of canary servers first and measure the metrics? And you say, OK, like, what is the system that could have been in place that would have prevented that thing from happening? We try to do that with all parts of our culture. If we lose a deal, if we lose a candidate, if anything goes wrong, we'll do a dispassionate, blame-free root cause analysis. We call it lessons learned. Across the company. Yeah, we do it in front of the company every week. And the idea being that it's a way of making a failure collective because it makes it not pessimistic, but it's like, here's how we're going to keep this from happening again.

1:09:11Bret Taylor:And I think it's important because for us, when I brought up the flip side of optimism is storytelling. And, you know, you want to have people who can solve a problem, but you also don't want people, the collective delusion that everything's okay when it's not because it might actually impact your decision making. So it's interesting because you sort of want people who are like optimistic, but like with a real solid sense of reality. and sometimes you have people are optimistic and they're like hype men, you know, and like everything's great when it's not. And then there's the toxic people. They're like, everything sucks.

1:09:48Bret Taylor:We're going to lose. Blah, blah, blah. Like get those people. Like I don't have the emotional capacity for that. And so our tactic is losing collectively. And I think it's really important. What I mentioned going back to the beginning and success as a thousand fathers, failure as an orphan. How can we make failure have a thousand fathers? And like, how can we make it so that when something goes wrong, there's a cultural, it's like cultural antibodies come out and say, we're going to make sure that we never make that same mistake again. And I, that's our mechanism. It's like Andy Grove, only the paranoid survive.

1:10:22Bret Taylor:It's like our way of operationalizing paranoia. Yeah. I think that's very well said. And I totally agree with it. I think on the optimism thing, like it's very easy to be a default optimist when things are going well. Yeah. You know, like the other slice of this is like almost I view it as a responsibility to inject tension and pressure into the system when things are good and then be supportive and optimistic when things are tough. uh and i think when things are tough i really value like yeah no shit like this bug just took down our system you know what i mean like it's pretty easy to be down on ourselves and i think we all collectively agree that shouldn't have happened and i wish it didn't happen but i think that in those moments of like real vulnerability uh that's that's the moment where i like crave the optimism.

1:11:17A hundred percent. You want the person who's like, let's go fix this. Not,

1:11:22Bret Taylor:oh my God. Yeah. A hundred percent. A hundred percent. I appreciate you doing this. Um, like, can we do it more? Like, I would love to do this more. I'm right down the street. Are you? Yeah. Where are you? Uh, we're right over on like, uh, right over on second and Howard. Oh, good. Yeah. All right. Yeah. We could do like a bi-annually or something. That sounds great. There's so much change. Like, uh, we could do it. Six months is like a year. I was going to say, We could do it weekly and we'd run out of things to talk about. Yeah, absolutely. Good to see you. Thank you. Are you hiring? Any roles that you want to shout out?

1:11:53Bret Taylor:We're hiring in all functions, engineering, products, our agent development team, which is a mix of sort of both engineering and some consulting roles, sales. We're also hiring in the UK, Europe, Singapore, Japan. So if you are talented, please reach out. All right. Thank you. Thank you. That's it for now. If you liked the episode, please leave us a review or go back into the archives where we've done more than 200 episodes with some fantastic folks. This podcast is a Kleiner Perkins production, and I'm Juven. Thanks for listening.