Sierra CEO Bret Taylor on why the AI bubble feels like the dotcom boom

24:01So I'm excited for it because I think like all technology trends, we're in the bottom rung of Maslow's hierarchy of needs. And it's very hard to see self-realization or whatever the top thing is, but I think we're going to get there relatively quickly. And our hope at Sierra is we can help every company in the world navigate that. Step one is make an amazing customer service experience for your customers that make them feel respected and valued and truly personalized. Step two is set your company up for whatever the future holds. What does conversational commerce mean? What does it mean when people are doing their consumer research on open AI rather than search engines?

24:42What does it mean when, you know, I'm sure many of you have done that. When you get like a lab result, I just upload into ChatGPT immediately before I talk to my doctor. I don't know how he feels about that, but I promise you 100 % of his other patients are doing too. Like the whole world's changing. And so a lot of what we think about us here is like, how do we set up every company in the world to be successful in that new world? Technically, though, are you developing your own models? What is the actual tech secret sauce here that you have? Is it models or what is it? We do a lot of fine tuning.

25:13We don't pre-train any models. I think most applied AI companies shouldn't. It's a very fast depreciating asset and probably would not produce a meaningful return for your shareholders. But it is quite complex. For any given message to one of the agents on our platform, probably 20 plus inference calls just to generate one response, just to give you a sense of the complexity. Lots of different models under the hood. There's not one provider or even one parameter count, which is a measure of sort of the complexity of these models. I think that's the way for an applied AI company like Sierra where the world is going, because it's almost like saying, what's the right way to store my data?

25:56And for the technologists in the room, there's like a trillion different databases and data storage systems from Snowflake and Databricks to traditional transactional databases. And we've gotten to the point now where like a modern technologist would know, hey, for this use case, this is the correct choice. That's where we're going with, I think, in the applied AI space, not artificial general intelligence, but the applied AI space is these models are truly pieces of infrastructure. And, you know, sometimes you want something that is really fast and sometimes you want something that's really cheap and sometimes you want something that's really high quality.

26:30And this price, performance, latency choice, there's really an option everywhere in that matrix for whatever you want for your business. And I think it will end up like the database market. It will be practitioners of building these agents and other things. They're not going to be the researchers who know how to pre-train a model. And my intuition for what it's worth is even fine-tuning will wane over time just as the context windows and quality of rules adherence improves in these models. But there will be what it means to build an application on these models will be like, hey, do you know how to use a database?

27:09not doing how to write a database. And those are two very different skill sets today. And I think that's kind of where the applied AI market is going. I think we saw with the release of GPT-5 that the models, they're still getting better, but the step changes are not as dramatic as they used to be. Maybe that will change. The space moves faster. I don't totally agree with you on this one, but finish your question because I'm rudely interrupting. You're on the board of OpenAI, nor should you. But I guess what I'm saying is, do you agree with the thesis that the models themselves are kind of commodifying?

27:42They're becoming like, I mean, you talked about it as infrastructure. But I guess what I'm getting at is, what are the second order effects if that is true? If the models are really just becoming like anything plug and play? Yes, they have certain attributes that are better. But they're not dramatically step function changing like they used to. Well, the reason I was disagreeing wasn't being an open AI homer, which I am, by the way. So happy to play that role. Actually, more just saying, I think it really depends on the task. If anyone was using GPT-40 or 4.1 for coding, and then swapped in GPT-5 for coding afterwards, you saw a dramatic improvement in performance.

28:23So through the lens of that task, it was very much a step change in performance. And so for people who are using this for coding agents, I think through that lens of that use case, what you said was definitely not true. Absolutely a step change in performance. I planned one of our vacations on ChatGPT earlier this year, and I think I was using 4.0 to do it. In my guess, if I used GPT-5 for that same trip planning, it would have been like, yeah, okay, it's like whatever, slightly better. I had a great vacation, so maybe I just didn't have high enough standards. Maybe it would have gotten a lot better.

28:57I think that we're getting to the point for a lot of tasks. We've reached sufficient intelligence for those tasks. So when new models come out, if you're measuring it relative for planning my vacation, you'd be like, gosh, I don't see a huge change in the quality of this model. If you're trying to discover a new therapy and you're doing drug discovery, or you're trying to autonomously write a complex piece of software, or you're trying to do a complex asynchronous agentic task, your perspective on how big of a step change may change. So my intuition, but it's just one person's intuition, is that perception of how big of a step change these models are will increasingly be a function of how complex of a problem you're trying to solve with them.

29:46And if you think about what it means to build artificial general intelligence, we need some more improvements, right? There was a really interesting thread on X from an open AI researcher who gave it a math paper, and it actually had a relatively novel approach for a type of math I don't understand. So that's the limit of what I can say about that. But it was really interesting. It was really creative. It really had that sort of almost like AlphaGo moment of like, wow, that's interesting. It's sort of novel, new mathematics. certainly we want to get to the point of uh developing new ai research finding new drug therapies uh proving some of the unproven uh math problems in the world we have some work to do we haven't gotten to that point but my guess is for you know what motivated your comment probably for a big bunch of tasks the models have sort of gotten to the point of of sufficiency and i think so going back to your question which is so what does it mean i think you know open ai is a mission driven company.

30:43Our mission is to ensure that artificial general intelligence benefits humanity. And we want to work towards beneficial AGI. And we're not there yet. You know, we need to continue to do that research and development. There are parts of it that already are super intelligent, but there are a lot that aren't. And that's really what we're working on. Does it mean for different tasks that Sierra solves or you do in your personal life, we need those, you know, really powerful models? Maybe not. And I think that will just result in an ecosystem of models and what they're used for. But what's exciting just around here in San Francisco, we're not done yet.

31:19We want to create AGI and that's really exciting. And I think despite the perception of these models slowing down, I'm not sure, I don't really subscribe to it. And you can see in some of the true research breakthroughs, the math Olympiad results. I mean, these are really meaningful new changes that weren't possible with previous models. And I think they're dang exciting. We need to take another quick break. We'll be right back.

31:53Support for the show comes from Rippling. If you're a business owner, here's the truth. SaaS promised to make work easier, but now the average company is buried by hundreds of apps that silo your teams, slow you down, and simply don't work together. That's not SaaS. That's SAD. Software as a disservice. That's why you need Rippling. Rippling is the unified platform for global HR, payroll, IT, and finance. They've helped millions replace their mess of cobbled-together tools with one system designed to give leaders clarity, speed, and control. By uniting your employees, teams, and departments in one system, Rippling removes the bottlenecks, busywork, and silos your software created.

32:36Automated, perfectly in sync, and seriously simple to use, Rippling gives your company one source of truth for your people, their data, and everything they touch. With Rippling, you can run your entire HR, IT, and finance operations as one. Or pick and choose the products that best fill the gaps in your software stack. And right now, you can get six months free when you go to rippling.com slash decoder. Learn more at rippling.com slash decoder. That's rippling.com slash decoder for six months free. Terms and conditions apply.

33:20And we're back.

33:25I'm glad you brought up AGI. I am increasingly of the opinion that no one knows what AGI means. But I think as the chairman of OpenAI, it actually really matters what you think OpenAI means. And I would love to know what AGI means. And I would love to know what you think it means and if that has changed at all for you, especially in the last year or so. Do you have a sense of like, this is AGI. When we have achieved this, we have hit AGI. First, I'll answer the last question, which has it changed. Yes, it has changed for me. I think we are already at what I would have defined AGI at three years ago.

34:01Actually, by the way, I think we're at what almost anyone in the world would have defined AGI as three years ago already. There's this thing called the Turing test, which is, I think, actually, I don't know what the original one was in the paper, but the way it was taught to me in computer science was having a conversation with an AI and having it being basically indistinguishable from human. And we've been past that for years. and we passed this. This was like the big thing in AI for long. We passed that and just like, yeah, forget the Torian test. That was a dumb idea made by the smartest computer scientists of all time.

34:33And so we just keep on moving the goalposts because we have exceeded our own collective expectations about what this technology can do so many times that what we had intuitively thought of as AGI, like we've lapped it like four or five times. So the way I think about it now, and I may change again, is that in the domain of digital technology and ideas, are these models exceeding human intelligence or at human intelligence in almost all domains? I say that in the digital domain of ideas because I think it's one thing to invent new types of math, which I think a lot of people would put in the domain of AGI and superintelligence.

35:21Interacting with the physical world is a whole different thing. And I think that's a separate problem that's unrelated to intelligence per se. Just being concrete about it, you can invent a new therapy, but a clinical trial is a completely independent process. So I think the intelligence part of it is really what I was trying to find a measure of. And the other part of the G in AGI is generalization. And so one of the things that I don't believe, but, you know, talk to a lot of researchers, that's what's interesting about AI is like some of the smartest people don't agree on all these things. But, you know, if you make something that is really, really good at math, how good it will be at a lot of other things.

35:59And you'll talk to some researchers who think, well, math is sort of the basis of reasoning and it will be great. A lot of things you talk to other people, like, I don't know, will it generalize to different parts of biology and other things like that? So I am more on the camp of, I think, as long as the model isn't, you know, trained for something like the math Olympiad and as a byproduct of the model, it will generalize. But I think we'll have to see, you know, and I think we'll see how it generalizes to other forms of science and things like that. But more and more, just looking at the rate of progress and seeing some of the results coming out of OpenAI and the other research labs, I'm more optimistic.

36:35And I'm looking forward to the first true scientific breakthrough coming from an AGI. And I think it will happen in the next couple of years. I don't know, but it feels like it will. It's a better question for the researchers at OpenAI than me. But certainly, if you start to see some of these early results, it certainly feels possible. Why are people like your old boss, Mark Zuckerberg, now talking about super intelligence? What is the difference here? I mean, this is like a thing when you're out here in Silicon Valley and SF, like now people are saying super intelligence. And it's like, is it because we, like you said, everyone is kind of like, well, we did it, shrug, like Turing test.

37:14Rebrand. Yeah, rebrand. What is the difference? I don't really understand, to be honest. Super intelligence, I think, literally just means that it is more intelligent than humans. And so, you know, I guess if there's a subtle distinction is if you made something that was generally intelligent and functioned as well as you and me, is that like no lackluster? No offense, Alex, by the way, I think it'd be great if we made it. You're sufficiently intelligent for me. And so I think it's a higher bar, which is truly exceptional. There's a few reasons from a research and safety standpoint it's useful to talk about superintelligence because it's a reminder that if the models exceed your own capacity to reason, how do you monitor it?

38:03How do you make it safe? And you really have to use technology to monitor the technology if it exceeds your own capacity to do so. There's lots of precedent in non-AI things. You have lots of things in an airplane or a car or monitoring for things you can't understand or operating too fast. But that is a really important area of research. So I think it's useful to talk about. There's sort of the public relations part of it that I don't really have an opinion on or care to think about. But it is useful when you think about safety. And there's a real question of like, how do you know that it's aligned if you can't understand it?

38:38And how important is it that a human being understand it versus a supervisor AI that we made to understand it? And there's a lot of both technical and philosophical questions and all of that that I think are really important to answer as we develop. Yeah. I was at a recent dinner with Sam, Sam Altman. This dinner got a lot of headlines because Sam said that he thinks we're in an AI bubble. And his exact quote was, someone is going to lose a phenomenal amount of money. We don't know who, and a lot of people are going to make a phenomenal amount of money. I'm wondering... It's like the old marketing quote.

39:12Yeah. It's been 50%. Only 50 % of my marketing is useful. I just don't know which 50. Right. Do you agree with that? And if so, why? Oh, absolutely. Yes. I've given this analogy before, so I apologize, Alex, if you've heard it. But I think there's a lot of parallels to the internet bubble. If you look at the internet bubble, a lot of people think about the flops like pets.com and Webvan. Through the lens of the past 30 years, though, we've now gotten most of the largest companies in the world, including Amazon and Google, two of the largest companies in the world. But then you look at how much of Microsoft's market cap is from cloud and others, and you start to look and you say, actually, if you look at the GDP of the world, how much has actually been created or influenced by the existence of the internet?

40:03And one could argue that all the people in 1999 were kind of right. It was as impactful pretty much by every measure. And even things like Webvan, there's now, you know, as the internet became more distributed, there's really healthy businesses like Instacart and DoorDash and others that are sort of built that are now that the smartphone and the scale of the internet has sort of matured. So even some of the specific ideas were actually not that bad, but maybe a little early. But if you look at the internet, so like if you, you know, were an Amazon shareholder from their IPO to now, you're looking pretty good.

40:38If you're a Webvan shareholder, you might feel a little differently. And so both exist at the same time. And so I think right now you have just, I think the AI, modern large language models and modern AI are absolutely going to have a huge impact on the economy. If you just look at software engineering and customer service by themselves, I mean, they're just, we haven't seen a world with which we've reached a sufficient number of software engineers. And we probably will with coding agents, you know, just because we've taken something scarce and we're making it more plentiful. and what is the market for developing software?

41:13I don't know. I don't even know how to measure that because every company in the world is now a software company to some degree. So as a consequence, I think just for me, it almost has to be that there's gonna be huge winners in this and because of the amount of economic opportunity, you just end up with a ton of investors and some companies will fail and some will succeed. If you look at the people who built out fiber in the early days of the internet. A lot of them went bankrupt, but that fiber ended up getting used, you know, just by the next person, the private equity firm or whatever that bought it.

41:47So I think it is both true that AI will transform the economy. And I think it will, like the internet, create the economic, like huge amounts of economic value in the future. And I think we're also in a bubble and a lot of people lose a lot of money. I think both are absolutely true at the same time. And there's a lot of historical precedent for both of those being true at the same time. Does it worry you at all that the bubble could be in the sector of AI you're in, in the enterprise? I mean, there was that MIT report that everyone's been talking about where a lot of spend is not seeing results.

42:17I know you have a different pricing model that's more geared towards success. But I don't know. It seems like the bubble could be all the enterprises that have rushed in and are spending a ton of money on stuff that's not working. What happens when that reverses? I'll decouple whether I worry about it and that study, which I disagree with. because I do worry about it, but I don't worry about that study. So I'll decouple the two. So I'll end with the study because it's like more optimistic than me worrying about my existential issues around my business. But I'll start with that. Yeah. I mean, it's weird.

42:48So there's this like story that goes on around me about rewriting Google Maps and it's like mostly true and a little embellished, like many great stories are. And it's interesting to me because that particular thing is people like to tell the story because they're like, oh, wow, one person wrote a lot of software over a weekend. And now if you've used like codex or cloud code, you're like, yeah, I can just have an AI agent do that a weekend. So it's like the thing that's sort of like so exciting that was like actually part of my own personal identity is now literally an AI agent, maybe not quite yet.

43:22I wrote some pretty good code, but like, you know, probably in a couple of years, like, yeah, an AI agent could totally do that. And so it's going to go from, wow, that was impressive to, wow, people did that, you know, like over the next couple of years. So there's the business thing, which is what is the software market of the future? I think it's a really good question because if you pull the thread and we reach plateaus, like, you know, self-driving cars, we're really excited. It took a long time. So, you know, even smart people can be wrong on these things or are too over optimistic, but with agents doing software engineering, we're taking the scarcest resource and one of the highest paying jobs, scarcest resource.

44:02And we're literally making AI agents that do that. So what will that do? I have a lot of people say, should I study computer science in school? And I have a bunch of opinions. I think the answer is yes, but honestly, no one really knows, right? And are we going to reach a world where generating software and generating is not the hardest part of software as most software people know, but generating software will largely become a commodity. Maybe a lot of people think that. What does that do to the software market. My hypothesis is actually, it doesn't change a lot. I don't think when you buy, going back to my ERP, I don't know why ERP systems are on my mind this morning.

44:39But you know, if you buy an ERP system, you're not buying like the bits and bytes that did it, you're buying the fact that a lot of companies have their ledger on it, and that you can close your books every quarter on it and it's reliable and they have like patched their servers so that you know that your cloud-based ERP is not going to have a security vulnerability and they have these compliance certifications and all these other things that aren't particularly exciting, but they're kind of the boring but important part of enterprise software. If you could write your own ERP system as a major CPG company, is that a good idea?

45:15I'm not totally convinced it is. I always like to say software is like a lawn, you have to tend to it. And so if you build it, you bought it, right, you have to own it and maintain it and deal with all the, you know, there's a new accounting standard that comes out. And all of a sudden, you're, you have to like do that yourself. So I think it will change the way we write software. Do I think it will completely upend the landscape of the existence of an enterprise software market? I don't totally believe that might be wrong. it's really new we're just in a really new world because we're taking something scarce and making it plentiful i i have thought about this movie a lot recently which is uh and i recommend people watch it which is a hidden figures it's a great disney movie about the putting people on the moon but in particularly the women who did the math calculations to do it and their job title was calculator i'm a calculator i didn't know until that movie and i watched with my kids that that was a job title.

46:11And there's an interesting one of the women in there, they're putting in an IBM computer, which is like the size of a living room. And she, you know, sort of a savvy way learns how to use punch cards to so that she doesn't, you know, like basically for job security. We're all kind of going through these moments right now. Like I'm a calculator, basically. And that story of me with Google Maps is like a story of a calculator, right? And so but I think the second and third order effects are a little fuzzy. I believe the enterprise software market will change towards from software to agents. But I believe companies want to buy solutions to their problems, not build software.

46:48So I believe the market will continue to exist. That study, I don't know the basis for the data in that study. I think it was problematic because it conflated people building their own software with AI and buying solutions off the shelf. And I think those are two very different types of AI adoption. We have basically a 100 % success rate with our customers doing a proof of concept and going live with our platform. I think it's because we're not selling AI, we're selling customer experience. And you just like turn it on and it works. There's an amazing company called Harvey somewhere here in San Francisco, actually don't know where their HQ is, really great company.

47:26I've talked to so many law firms who've deployed Harvey for a lot of their legal analysis, and they're all successful. But it's not an AI project. Like Harvey uses AI, that's the only reason the company can exist. but they're doing it because you want a better antitrust review process. And that's what they're buying from it. I think we're just in the early days of AI, which there isn't a wonderful vendor for every problem that you have in your business yet. So as a consequence, you have to wait or you have to build it yourself. And so I don't know what the right answer is. I'm not, you know, and when you're trying to build these things yourself, it's just a, just a glacial amount of complexity.

48:02Uh, and, and what you end up having is a lot of these, I call it AI tourism. You know, like a lot of people doing, you know, performative AI projects, but to actually go to that last mile is quite difficult. I think the solution will be applied AI companies. I think, you know, if you want to build a better customer experience, buy CIRA. If you want a legal AI agent, buy Harvey. And I think we need to go through every use cave from supply chain optimization to accounting to maybe an auditor for your accounting department. All of those will be AI agents, but I think there's a company in every single one of those domains.

48:37And I think that is the correct way to package and purchase AI software. And I think that study is a measure of all the people who are trying to, which is just lack of maturity in the market. There's just not a solution in every space yet. But there's a few VCs in the room, hopefully, thanks to all of you. In a couple of years, there will be. And I think that will be the new, what was software as a service, those new agent companies will be that next frontier of business solutions for enterprises. All right, Brad, we have to leave it there. Thank you so much. Thanks for having me. All right.

49:36and subscribe wherever you get your podcasts. Decoder is a production of The Verge and is part of the Vox Media Podcast Network. Our producers are Kate Cox and Nick Stat. Our editor is Ursa Wright. The Decoder music is by Breakmaster Cylinder.

49:57Refresh your good vibes with Tic Tac. Mit den richtigen Vibes wird Oh man! Zu

50:16Jetzt probieren! Die verschiedenen Tic Tac Sorten. Von minzig frisch bis fruchtig süß. Tic Tac Refresh your good vibes