Ep 160: Box CEO Aaron Levie on Silicon Valley's AI Problem, Evidence for AI Job Creation & What Doomers and Accelerationists Get Wrong

Speaker A: You're telling me this thing is gonna destroy the planet and you're the ones making it and you're not stopping. The actual creators of this stuff are scared shitless about it. Imagine if like Larry David made AI models. Like that's kind of what we're dealing with right now.

Speaker B: So where are we in the AI, uh, wave?

Speaker A: Companies going from 0 to 100 million or 500 million in revenue in one to two years. This has never happened before in software. Anything that was slow in a process. Because of that, we should be able to go and have agents go and automate. If you believe in job destruction, like in any kind of very large scale way, you basically don't believe that markets ADAP going to become the number one topic in the next presidential election. I think we're in for some very messy few years on this.

Speaker B: My friend Aaron Levy is the founder and CEO of Box. He's been a leader in the technology world for over 20 years. Over 100,000 businesses use his products. He's a champion for AI, a voice of optimism against a lot of the doomers. He's also a really funny guy. I think you'll enjoy hearing from Aaron what's going on in the AI, ah, world. Welcome to American Optimist. Really excited to have our friend Aaron Levy from Box here with us. Aaron, thanks for joining.

Speaker A: Thank you. Good to be here.

Speaker B: It's nice to see you out here in California. Don't tell them I'm visiting for too long.

Speaker A: Yeah, exactly. Is that like a tax issue or.

Speaker B: No, I come here a lot. Probably, probably about a quarter of my time actually.

Speaker A: Uh, we would love to see you more.

Speaker B: Yeah. Uh, it is the center of the AI world. Although you probably, you did see Sironic in Texas last week.

Speaker A: Did you see that? I saw that. Uh, I mean, obviously, um, I think Austin's having a moment and obviously broadly Texas. So, um, it turns out if you can build like real things, uh, it helps your state. So congratulations on that.

Speaker B: Thank you. Thank you. You guys are still ahead of software. Let's dive right in. We're right in the middle of the AI wave. I think it's the most exciting time to build. You know what I'm seeing and I saw a bunch of them here in the last week. Of course it feels like every two to three months is like a year now in terms of build time. It's like something's accelerated and so business itself's going faster. Right. Are you seeing this too?

Speaker A: Yeah. So, uh, I mean we have good, nice, um, comparison checkpoints so if you go back 20 years ago, you were like, okay, we have an idea, we want to get it into the market and you might be heads down for a year or two before the thing actually goes live or seen by anybody. That's obviously shrunk into a matter of months at this point. Um, if you take that type of time compression, um, I think that's changing everything about feature development. New, uh, product launches, getting products to customers obviously have covered companies going from 0 to 100 million or 500 million in revenue in one to two years. This has never happened before in software, um, or technology broadly. So, um, definitely an amazing moment. And it's all driven effectively. The underpinning is model progress. So, um, the downstream, uh, effect of model progress is everything else just moves way faster and we're definitely seeing that.

Speaker B: Speaking of how long it takes to get into production, I've seen some examples now where an agent will get feedback from the customer and then like go and suggest what to fix, put in their pr, get it approved and fix it. Which is kind of crazy. It's as real time almost.

Speaker A: Yeah. I still wonder if that's more of like a startup phenomenon, but the idea of like this automated software development lifecycle is pretty compelling. So you know, the moment from bug, uh, to then live production fix of that, you could be, you know, you could compress something that used to take days or weeks into a matter of minutes. Um, you know, the main, the main thing is what is the kind of human in the review step that's needed in that. But it's a very, it's very compelling.

Speaker B: As someone from your generation, I think we should be reviewing it in person.

Speaker A: Yes, exactly. We need to keep people in this. What are we talking about? We have so much more taste. Um, uh, but I think that also applies to many other analogous things. So imagine, uh, a security incident. You obviously would want to respond to it as quickly as possible, so that'll become entirely agentic. Um, any kind of production issue with your systems, you'd want to be agentic because you could take the, the downtime incident from maybe, you know, 10 minutes or 30 minutes or 60 minutes to a matter of seconds.

Speaker B: So where are we in the AI wave and where do models go from here? I want to get into the Kimmy China debate stuff. But before we go there, like, you know, when you talk to the guys around Anthropic, like Dario, at least a few months ago was saying these things are going to double every few months. And I can already see the next Two years where it's going. And this is one of those things that's very unintuitive even to me because it's like, you know, if you double eight times in two years, that's, that's 256 times. Like what does that even mean for like where they are now, where they're going the next six months? What is, what is, what do you see for your business?

Speaker A: Yeah, um, uh, so maybe we'll start first with model progress. So model progress appears to be effectively not slowing down at all in some domains, even accelerating. Uh, and so every prediction, like last year around this time there were these kind of debates of like, have we hit a wall? Uh, and I think we've just completely blown through that. And a mix of that is, uh, the companies are continuing to improve their pre training, uh, steps, uh, the kind of data that they get into that and then the post training steps as, and then kind of everything in between. So we just keep unlocking, I think new breakthroughs of how to improve these model capabilities, uh, which is incredible. Um, I don't know if you guys invested in these, but there's a whole ecosystem of new startups that are doing data for the labs. Um, and so if you kind of go back to Dario's originally kind of scaling laws concept, which is just more compute, more data, um, you'll just see progress effectively follow uh, that line and we are just getting an exponential amount of really, really high quality data and compute keeps coming online to let you go train these models. So you could just assume kind of a complete exponential curve on model progress. And we're seeing this across all of the labs. Um, obviously anthropic, really touching the current frontier with Fable GPT 5.6 being just kind of neck and neck in a bunch of domains. Um, obviously Grok having really good showing, Meta coming back, um, Gemini doing quite well. So you've got at a minimum of five leading us lab that at a minimum within a very relatively narrow band are going to all be following each other at the frontier, which is fantastic. Um, and what's cool about that narrow band is you get kind of a narrow band of intelligence difference but a pretty wide cost variance. So you have different players saying okay, we're going to be the low cost provider, other players saying we're going to be the kind of very purpose built provider for high end workflows. That's effectively how things are shaping out in the us Then of course you've got the sort of open weights models, a couple in the us, uh, Nvidia, Thinking Machines and a couple others. And then obviously the sort of China story, the biggest I think update for everybody in the past three to six months was I think the debate would have been probably if you had just gone back one or two years and we were to take the top 10 AI researchers and say how far behind is China? I think general consensus would have been six months or a year, kind of at a minimum, just based on either access to chips or um, kind of the methods they were using. Now it's very clear you have something like a Kimmy K3 moment, which is really, uh, kind of. If you look at uh, artificial intelligence analysis and their overview of AI models, it's considered to be basically the third best model kind of on a composite sketch of.

Speaker B: It's like a near copy of Fable.

Speaker A: Basically it's able to do kind of near Fable esque. Uh, and there's debates. People say it's far less efficient on tokens. Um, it's got obviously going to have maybe less judgment and taste somewhere in there, but let's just say it's a very good model kind of regardless. And even on some things that are more subjective, like this front end design test that people do, um, in uh, LM arena, uh, it actually was the top model in the world. So on some dimensions is actually spiking as the best. So no matter what kind of, you can look at this and say there's an open source ecosystem in China that is near neck and neck, let's say at least still maybe kind of one to three months behind, but not a year or two years behind and then a really, really strong frontier, uh, ecosystem in the US all basically racing ahead. So that's kind of model progress. And then um, what's going to be interesting and Dario, I'm stealing this from Dario a little bit. Um, like a year and a half ago, he kind of framed this I think quite well on some podcasts where he said we as consumers are probably going to start to experience this model progress kind of almost less and less. Like we're going to go to ChatGPT, we're going to go to Claude.

Speaker B: We're not trying to solve hard math and physics problems day to day.

Speaker A: Exactly. So like, we're not, we're, we, we are not uh, you know, giving some, you know, incredibly unsolved math problem to uh, you know, to a model. We're saying, you know, hey, where's like a good place to, you know, that's open for uh, for food tonight?

Speaker B: This must Be like for your experience where you feel underappreciated or your intelligence.

Speaker A: Yeah, exactly.

Speaker B: Why is it far?

Speaker A: Absolute people asking me really hard questions all the time about strategy. Uh, so, so what's, what's going to be kind of fascinating is you're going to see this kind of bimodal thing, which is consumers probably are already at the point of like, like, you know, I'm talking to an AI almost, you know, five times a day about some personal thing, you know, you know, a kid thing, a family thing or you know, somewhere we're trying to, you know, maybe go on a trip and like it's already super intelligent. Like, I don't, I don't think I need that much more.

Speaker B: You don't need Einstein versus the physics professor down the street to talk about your pizza order.

Speaker A: Exactly. Like we've, we've actually solved that. In fact you might actually like, there's probably like a, uh, a little bit of a horsesh. Einstein would probably get the wrong pizza. So there are some upper limits of how much intelligence you want to apply to different problems. Um, and so now I'd say consumers are just like, we kind of know what the consumer thing is going to look like. Obviously when you add robots and stuff you're going to need that level of intelligence. But for consumer kind of end user products, we're in a pretty stable position. The flip now is on the enterprise side and the enterprise actually right now we're nowhere near the level of intelligence needed for the vast majority of enterprise tasks. And that's because if you're doing an insanely hard due diligence problem on a uh, on an M and A deal, you're doing, you know, there's just a lot of compute and a lot of intelligence. You want to apply to that.

Speaker B: You're bringing a lot of data together

Speaker A: from a lot, you have mass amounts of data. You want to have sort of the entire history of like every M and A deal, you know, in the world that's ever happened. Um, you want to be able to understand every single legal framework that you know, on the planet that might be, you know, kind of have implications for this. So there's just a lot of knowledge that needs to be, um, you know, uh, uh, you know, packed into these models for those types of use cases. And that's more like basic knowledge work. Now you go into life sciences, you go into, you know, material sciences, you go into healthcare, um, you go into um, anything dealing with kind of manufacturing. And we, we still have massive sort of leaps to Go.

Speaker B: First of all, we're still trying to cure all these diseases. But you know, it is interesting. I do think it's mostly head of doctors now in a lot of diagnosis areas, which is kind of crazy already, you know.

Speaker A: Yeah, yeah. So, so I think for again the, the kind of common things you would go into, you know a doctor for AI plus doctor and, and people Deb like is the AI doing most of the work is the doctor's judgment on the AI but you can still eke out kind of positive gains by having doctor plus AI. Um, but now I think the real frontier is sort of this diffusion of AI in enterprise workloads that actually are really driving meaningful change in the real world that would actually affect our lives. And what would affect our lives it would be if life sciences companies can have breakthroughs on a new cure for an unsolved problem. Uh, we would advance our lives. If you could uh, have automation breakthroughs in the manufacturing process, um, we would have more breakthroughs if we had better uh, software and AI in all of our kind of common ways of interacting with the world. So that's kind of the upgrade cycle that needs to happen. And the only kind of caveat is that's going to probably take like 10 to 20 years. Um, not just because of model progress but because of the rollout of Interesting.

Speaker B: Because I'm seeing a lot of things. For example, one big thing is can we bring manufacturing back to the US and there's all these really hard AI problems to do. What the manufacturing engineers in Shenzhen do and a lot of my friends think we might be able to do a lot of that in the next couple years. I guess that's a question. Are there going to be things along the way that we're solving that that are helping us or you think a lot of it's further out?

Speaker A: I actually think this is, and this is the, where this is the spot where you would, you would start to veer between the kind of um, either like hyper accelerationists with AI and what's funny is actually the accelerationists and the doomers sort of share roughly the same timeline because they, they both think that AI is the intelligence is the only thing that matters. And then I'm more on a pragmatist sort of timeline which is like intelligence is, is a super critical component but then diffusion actually is your biggest rate limiter. And diffusion is like a human has to get the intelligence from the model plus probably their data and then they have to go interact with the real world and do something in the Real world and then bring back whatever the feedback loop is back to the model and the data. And that's actually your ultimate rate limiter, which is just like the speed at which humans can actually make sense of what the models are telling them to go do. The ability to go do an actual study on humans with a new kind of biomedical research, the ability to actually get the permits to build the thing that the AI helped you accelerate building. That's all of the kind of limitations of the real world. Right. So maybe, I mean, this would be an argument for like, become the state that sort of lets you actually do all these things.

Speaker B: We're actually thinking of putting the permitting into, like, AI land and having the rules be cleared by humans, but then letting the AI go really fast. Yeah, because you have to go fast 100%.

Speaker A: So think about just even the paperwork kind of processes that slow down the real world. Uh, if you can automate all of that, obviously, then you could shave things like a permitting process for maybe months or years, in some cases to a matter of days or weeks.

Speaker B: Um, sure.

Speaker A: There's always this other bottleneck that something pops up to because there's an escalation that somebody has to go look at, and then they have to talk to somebody else. So I think that the really bull case on AI is if you think about it as everything that is sort of information constrained in a process, you should just be able to automate away the reading of something, the processing of something, the typing of something, anything that was slow in a process. Because of that, we should be able to go and have agents go and automate. And then the only limiters should be like, atoms and human kind of coordination.

Speaker B: So you're on my team. That bureaucracy is the last great challenge for our civilization.

Speaker A: I do, I am. But maybe I overweight that as that will just be like an impenetrably hard problem that will be around forever.

Speaker B: Well, we can't be too mean to them because they fund one side. But stepping back. So you're seeing all this data in all these enterprises. And so you're saying the same problem with diffusion is just like people, things like, what does that show up as in most of the companies that you're looking at?

Speaker A: I think what it shows up as is, uh, you go into an enterprise and they want to go and automate their client onboarding process, or they want to go and automate their clinical research process. Uh, um, and they have to. First it's like, well, do you just deploy? Everybody has their little AI agent in their own personal workspace that they interact with. Um, and that's an interesting strategic question. Or do you deploy something that's sort of centralized, that is like the new machine in the workflow. Um, when we deployed manufacturing plants, we weren't like, oh, let's just give better tools to every single person on the manufacturing line. You step back and you said, no, let's actually figure out what is the new whole manufacturing process going to look like. Most companies are still in that kind of former process, which is, okay, we're going to enable everybody with tools and that will accelerate each individual, which is

Speaker B: incredibly powerful, or even each department with its own separate areas, but they're not working together 100%.

Speaker A: And so now you have this issue which is, well, every single person now has been accelerated in their own kind of domain, but that probably didn't like reinvent the process for a world of agents. So then you have somebody wake up and they say, oh well, what we probably need to do is make sure that we are reinventing this whole process end to end. The client onboarding process, the clinical life sciences process, um, the M and A deal review process, which means we need more of a central system that can both have access to the right data to work with, which is its own very hard problem, and have the right sort of input and outputs for the humans to interact with that workflow.

Speaker B: So my latest bias, which is super obnoxious to big companies, is that it's so hard to re engineer how some of these things work based on what's possible today, that they're just going to get crushed by new companies started by younger people who are maybe not even younger necessarily, but people who are just doing things in the new way, freshly designed, the processes, the whole idea. These AI services or these companies just coming in like building competitors. I think, I think there's going to be a lot of new big companies right now.

Speaker A: Yeah. So, uh, I am 100% bullish on that thesis. Um, with one sort of caveat which is, um, I tend to now think that what happens is you end up with more of an ecosystem which is still like on average the big incumbents eventually find a way to kind of

Speaker B: break through, especially the capitally intensive ones

Speaker A: like the guys that own or regulatory control ones. Is jp, am I, am I that dependent on the speed of getting a wire transfer done that I'm going to change my entire bank because of one bank has agents and the other is a little bit slower? Maybe not. There's probably more. I kind of just want to make sure that I am on the board

Speaker B: of Erebor with Palmer.

Speaker A: Okay, fine. Okay, well then you're going really fast. But Jamie Dimon's probably not going to

Speaker B: lose his job in the near term.

Speaker A: I think that's the point. You could imagine, um, um, I'll grant you that that could be a half a trillion dollar company and that would be great. But like, does JP Morgan go out of business? You know, not sure. So. So, um, and we saw this before. Like, like, basically I had, you know, I was very, you know, had a very big thesis on like digital disruption in like the early 2010s. And we talked about this a lot back then, which is like, okay, you're gonna have like a digital version of every single existing incumbent company. So in the early 2010s you had like Robinhood, Instacart, Doordash, et cetera, all Airbnb, all these incredible digital disruptors. They all became fantastically large businesses and yet the incumbents, like they still exist in every one of those categories.

Speaker B: Shop and hotels are still around and

Speaker A: doing some, they're serving some purpose. So I think actually both will be true. Um, but I love the idea of the kind of agent first law firm, bank, uh, life sciences company, um, uh, kind of uh, uh, new financial services firm. So this is going to totally happen and I think it's going to present a huge opportunity for all new startups. What's cool is in AI right now for the diffusion of AI, you actually need, I don't know, there's probably five different layers of the stack and they're all kind of working at the moment. So you have the agent first company in that domain. You have the company that provides agents for all the other companies in that domain. You have the infrastructure providers for the agents in all of the domains. You have these sort of post training kind of agent model, uh, uh, training companies for all the domains. And then you have data and infrastructure and everything else. So it just turns out there's opportunity kind of across the stack for all of these players. If you're a cognition, you're going to both arm the new disruptor of, let's say system integrator companies that are doing software development and you're going to arm the existing system integrators. Your job is to make sure both end up growing fantastically well. And I think there's going to be opportunity for both of those plays.

Speaker B: It's a very positive sum view of the world right now.

Speaker A: I'm not seeing a lot of zero sum actually. And across the board, even with jobs. I think actually this is a technology that largely ends up being positive sum.

Speaker B: I love it. Well, it is called American Optimist. And um, I'm on the same page as you. I do think longer term, especially with robotics, there's some disruptions that are very scary. Although I think it's very positive overall long term.

Speaker A: Yeah. And I think the thing that you end up using robots for are going to be a lot of additive things that we just don't do today. You're just going to end up getting an abundance of a whole bunch of areas of work done that just were never possible. The thing, um, some of the top use cases we see from customers with our AI agent and their data are actually things that they never deployed people at before. So it's. I want to go and have an agent read every contract in our company because I want to pull out insights from these contracts that will tell me like what customers can I go and upsell better or where is there a new sales opportunity? It's an entirely new set of work that's being done. Nobody read those contracts previously. So agents haven't replaced anybody's work. And as a result of what the agents are doing, the company now actually is driving more revenue, they're driving more opportunity. That's going to happen in so many more places than we realize.

Speaker B: 100%. I think it's a really fun thing. If you think about you're talking to someone from 100 years ago and you describe these as robot servants or something and then it would be like absolutely comical to them like how much these robot servants are doing for you for like basic tasks. Yes. You're seeing this across the board that, you know, one of the new ones is, by the way, if you're in a marketing department and you're like creating material for a customer, you now need to like fine tune it if you're doing your job for every customer explicitly for their use cases. Right. So it's like you're making the robot service been like a week to get ready for this one customer email.

Speaker A: 100%. So, so this is the funny thing. Uh, and Bill Gates and Andy Grove had this dynamic um, of like, you know, the CPU would get better and then Windows would just kind of use all the CPU again. Um, and so there was just this uh, like, I think it was like Andy Grove giveth and Bill Gates taketh or something. So, um, uh, that same thing is kind of happening for like work which is uh, you would assume if you just like, you know, didn't have any intuition on this and just like kind of looked at a basic model, you would say, okay, okay, this company currently produces 20 marketing assets per week, or this company reviews 100 contracts per week. Now agents come in and they do both those tasks. You'd be like, well the people that were doing those tasks before clearly can't. There's no work for them to do because the agent just did it. Well, what happens is the economy starts to say that those two tasks no longer have much value. Like we end up raising the bar of then what that job is. So to your point, like you don't just now do like, oh, we're going to just like, like make five ads in this marketing campaign. You make like 5,000 ads because you tune it for every single geography, every single market segment, every single industry you're going into. And the reason you do that is because as a customer you no longer expect that just like the basic ad campaign or the basic targeting is going to work anymore. So all of our expectations have just risen as a result of AI and we see this in sales as an example. So, so one, um, would have thought that AI would help every sales rep prepare for every single meeting and instantly make sure that you've got the, you're super tuned into the meeting. And that means that that's going to reduce all the jobs. Because what about all the work that we were doing previously? What about the solutions engineer that built kind of showed a demo or uh, ended up making marketing collateral? Well guess what, the new thing is you go to that customer and you actually build a working demo prototype of the product that you're actually pitching. And so now you wouldn't just like as a customer, you wouldn't buy software if you didn't see like a working prototype of your environment with that tool. So what we did was AI agents made the previous task really efficient to the point where the previous task is kind of like not differentiating. So then we go and deploy our time at the new, much higher version of that task because the market now expects that. I think that's going to happen in so many more areas than I think what people are anticipating. So if you believe in job destruction in any kind of very large scale way, you basically don't believe that markets adapt and you basically don't believe that some other company will emerge and say, I'm going to just do this way better. And the way I'm going to do it way better is by adding people time again into the process with really, really high judgment and they're going to go and find a way to use these tools to differentiate my product or my service or my marketing better than the competition. Which then raises the bar for everybody else.

Speaker B: So obviously agree with you. A lot of people right now are afraid to talk about this in public. There's a lot of doomers, a lot of people maybe who don't have the economic imagination and understanding that you do. Uh, the polling is getting a lot worse on AI. Uh, bad bowling, bad holding both sides right now. New York state just imposed a data center moratorium. Yeah, there's red states where they're weighing bands. I haven't done it yet. We're fighting hard. But it's, but it's, but so people are really negative. They're really worried. Like, like, I mean, are we just like really unpopular, nerdy people they don't like and therefore they're coming for us. It doesn't matter what we say or like, is there, is there a better.

Speaker A: That's probably true no matter what.

Speaker B: Yes.

Speaker A: Like, like in all cases that's obviously true.

Speaker B: Obviously we have different politics, but we're both part of the whole evil tech, tech bro regime.

Speaker A: It turns out it's a very bipartisan problem.

Speaker B: Like everyone dislikes us. Um, what should we be doing and saying? What's the right messaging? Like, what do you want tech people to be doing about this?

Speaker A: Well, and I think we probably also agree on this. This is sort of a self inflicted problem, uh, from the industry. And um, it's just like super obvious. We have been scaring ourselves about this technology for the past decades, obviously. But even in the past kind of four years since the ChatGPT moment, we've been scared of the next model at every single turn. And we have shared those fears with the rest of the world. I'm saying we very politely as an industry, um, I think both of us, we've taken the other side of this conversation. Um, but the industry has kind of let out a very kind of neurotic, anxious kind of feeling about this technology.

Speaker B: It's very anti Semitic.

Speaker A: And um, uh, there is an interesting correlation, uh, between the participants. But um, it's very like imagine if Larry David made AI models. That's kind of what we're dealing with right now. Um, so, uh, the issue is we've sort of scared ourselves about the AI and we've told everybody those fears. And so then a lot of people are sitting around being like, wait a second, the actual creators of this stuff are scared shitless about it. Maybe we shouldn't be that into it either. Why is this all happening? And it's a very kind of weird cognitive dissonance, which is you're telling me this thing is going to destroy the planet and the world and you're the ones making it and you're not stopping. So now I'm just supposed to be very restless about this whole kind of set of facts. Um, that's kind of the industry. And when you have people like, uh, and this is like. I think everybody sort of attributes this to maybe one company or group, but it's a pretty broad thing. It goes back to Geoff Hinton. We'll tell people that all radiologists are going to lose their jobs or whatever. Um, and it's like, well, that just didn't happen. It actually just turned out that we just threw more compute at the radiology problem. And you still need people that are in the kind of human review loop on that. We have a history as an industry with this particular segment of technology

Speaker B: of

Speaker A: sort of not using our imagination about what happens next and then kind of publishing all of our fears about it, thinking that maybe if we do that, that will kind of create some kind of broad societal progress. And I think there's actually an argument that it will. I'm actually very favorable to the idea that these are open discussions, but we're getting exactly what one should expect, which is very low popularity, um, about this technology. So then that's going to have a wide range of implications. It's going to mean data centers don't get built out. It's going to mean that sort of lockdown AI models. Um, it's going to mean potential major new kind of tax regimes, depending on the administration that's in charge. Um, and I think a lot of it will have sort of been exactly what you should expect to kind of play out. Um, and I don't know, unfortunately, I don't know exactly how to change it because the way to change it is just everybody experiencing kind of universally positive outcomes from this technology and us putting a lot more emphasis on those while still mitigating all of the risks that we're afraid of. There are real risks with AI, there's no question. But that has dominated the sort of examples and the conversations that have come out of the industry, I think more than anything. Um, and so it's unfortunately a major messaging problem. Uh, uh, I think it's going to become the number one topic in the next presidential election. It'll be such an easy thing to create, uh, a populist type of conversation around. So I think we're in for some very messy, uh, few years on this.

Speaker B: Yeah, I'm hoping the positive impact is so much more obvious by 2028 that it's hopefully more popular by that election, because I feel like it will have created a lot more jobs. And I think you're going to get some disinflation from it by then. But I guess that's. That's a question.

Speaker A: We'll see. Yeah. And, you know, I think the tricky thing is going to be, and the one thing I'm kind of a little bit pragmatic about is like, I do think the consumer surplus, uh, that comes from it might be a little bit kind of hard to pin down because, like, again, I think on the consumer front for AI, most people that maybe aren't in tech are going to be like, yeah, I talk to my phone now and gives me answers, and that'll be like, awesome. And they would say, I don't want you to ban that. But they're not going to be as sort of sensitive to like, oh, in this part of the life sciences workflow, it let the lab go and run a million experiments instead of a thousand. And that led to the cancer treatment that you now have. That's kind of, um, a very hazy, amorphous type of thing to kind of get around. So I think that's going to be the one dilemma is because I actually think this is an enterprise technology mostly. Um, where is intelligence most valued in kind of enterprises coordinating and building things. Coordinating and building things. And those things are not probably classically what voters sort of think about or get excited by. So I do think you're going to have a slight dilemma where it'll be easier to look at the negatives of AI just by virtue of you can kind of get your arms around those things than the amorphous, positive things. So that's going to be a dilemma

Speaker B: for the industry, like the problem with capitalism in general in a way where it's probably easier to attack. You said you created 13 new jobs in New York Times, your company for AI.

Speaker A: What are those? Yeah, so, um, 13, uh, probably job families.

Speaker B: Yeah, Types of jobs.

Speaker A: Types of jobs. And, you know, they represent, you know, they'll eventually represent hundreds and hundreds of jobs. But, um, uh, you know, some of them are kind of all flavors of one giant category. But effectively, uh, there's a few immediate jobs that AI have generated for us. One is just the deployment of AI agents within our company. So we have these new kind of AI Automation engineers. And that role is just, how do you deploy, um, agents within our world?

Speaker B: It's like a sysadmin sort of thing for AI, which is different.

Speaker A: Yeah, Basically it's kind of like what the future of the IT role will likely be, which is, it's a, like, you know, you will spend some time implementing salesforce and workday, et cetera, but you're going to spend a lot of time implementing agents for the workflows of the company. It's a highly technical role. You have to understand data, you have to be able to build data pipelines. You often have to actually do real software engineering. So this is going to be this fantastic opportunity for a whole segment of engineers.

Speaker B: One of my favorite teams is working on that to help support that role, actually, while we're talking about it.

Speaker A: Okay, cool. This is, uh, actually part of the kind of antidote to, well, where are all the engineer jobs going to go in the future? It's like, well, actually, now everybody's going to have these engineers. Every kind of company, like, right. Like, it used to be that mostly it was only software companies that hired engineers. Well, if you have agents in every form of knowledge work, then every type of industry and every kind of company will need engineers as well. And like, I don't think we've done the math on this yet. But, but think about, like, you know, the, you know, Silicon Valley as a, as a location, but think about more as an industry was, you know, kind of the dominant hire of engineers previously. Computer scientists. So the Googles, the Metas, um, you know, our various startups were the ones that hired all the engineering. So if you were an automotive company, if you were a life sciences company, if you were a law firm, if you were a bank, you still tried to hire engineers, but certainly the talent pool was less abundant. Now all of a sudden, agents let those engineers become much more deliver, uh, much more, uh, from a volume of execution standpoint, in all of those industries,

Speaker B: everyone needs more engineers.

Speaker A: Everyone needs more engineers. And this is like classic Jevons paradox, which is we lowered the cost of engineering engineering, which means that everybody wants to do more of it. And now finally people can. And so you're going to see this huge surplus in demand for engineers across the economy in industries that we've never seen. Like, I've talked to, I don't know, three or four law firms in the past maybe month or month and a half. Okay. Like, this is like, it's worth really processing this. Like three or four law firms. The best month and Month and a half about literally building, you know, kind of custom agents or even custom models for their law firms. Uh, like this is, is, this, is, this is an audience that five or ten years ago we were having a hard time selling them SaaS, you know, technology in their law firm. Now they're talking about training AI models,

Speaker B: building their own small models to reflect how they work.

Speaker A: So, so think about how many engineers are going to now need to exist across all of these types of companies to make any of that be real or, or work. This is like going to be a total boon for, uh, if you're technical.

Speaker B: Let's go back briefly to the origins of Box. Just so people know. You've been working on this now for 18. No, no, 21 years. Yeah, 21 years, is that right?

Speaker A: 25. I mean, I like the 18 better, but yes, it's actually.

Speaker B: And, uh, and you actually, you dropped out USC to start the company. Where did it come from originally?

Speaker A: Um, the company.

Speaker B: Yeah. Why.

Speaker A: Yeah.

Speaker B: What do you do? And why are you still. Why are you still running as 21?

Speaker A: You're like the statute of limitations on running a company.

Speaker B: Yeah. Tell us what you're. You're obviously passionate about it.

Speaker A: So we started Box, uh, in. Basically launched it in 2005. And the original idea was, was actually like so dead simple. It's like, kind of funny to talk about, which is we just wanted access to our files from anywhere. Uh, and this was right at a moment where the cost of storage was coming down, Internet was getting faster, browsers were getting better, people kind of working from mobile devices a little bit more. This was like a time of blackberries and everything. So we just were in college and we said, hey, there should be a better way to access files from anywhere. I was doing an internship at the time. It was very obvious how hard it was to just work on data. So we launched Box, um, and it wasn't like an overnight success, but hundreds of people signed up in a matter of months and we're like, wow, this is a real thing. And we had done lots of different projects prior to Box, so had tried, I don't know, a dozen different startups or half a dozen startups before Box. This was the one that actually worked and people were signing up for. So we dropped out of college, um, uh, as a result of raising money from Mark Cuban, fellow Texan. Um, and so, uh, we raised, um, capital from Mark Cuban, dropped out of college, uh, and then eventually, about a year after dropping out, pivoted to the enterprise market market. And so for the past kind of 19 maybe years or so, the whole focus of the company has been enterprise. And the idea is again, still pretty simple. Over that time period, um, enterprises create an insane amount of data. Mostly it's unstructured data, which means it's research files, contracts, marketing assets, financial documents. It's all of that kind of data. It's very messy, it's hard to share. Some of these files are massive. You have a lot of it, so you have to kind of organize it and keep it secure. So we built a platform that lets enterprises do that, that, um, uh, and so we now have about 120,000 customers, um, about one point, kind of $2 billion revenue, uh, run rate. Um, and the whole idea is be the best platform to help enterprises manage all this information. And then the really big breakthrough, uh, is how do we make enterprises, uh, really able to now tap into the value of all that information. So, uh, how do you start to ask questions of this data, process it in new ways, automate workflows around it, it use, um, agents to understand all of our documents and be able to just ask it lots of questions. Um, and this is why we're so excited about AI. And my passion for it is AI needs data. Most of that data is going to come from unstructured data sources. It's going to come from your documents, it's going to come from all of this information. And so, um, for us, it's sort of solved this ongoing existential challenge, which is you have all this data, you don't know what to do with it, you don't know what's inside of all of it. Now, for the first time, LLMs basically are really good at that. So our vision is really how do we let you now tap into all of this information in your enterprise? And so all these use cases around, what if I had a knowledge base of every decision that my company made, um, every research project that led to some important outcome, every contract that we're working with, every marketing asset that will help create the next one or a better sales pitch, all that information becomes this valuable set of insights that again, agents need just as much as people, and we can all tap into the same data source to go and automate these workflows.

Speaker B: So are you helping people understand the processes that exist on top of their data and then how to use AI for that?

Speaker A: Yeah. So this is why we're kind of very close to maybe the diffusion kind of dynamic, which is you have the data. First of all, most companies still don't even have their data in modern environments. So most of the data exists in fragmented legacy systems. Um, uh, and so you have to first get your data into a modern platform, then you have to make sure it's organized well. Then you have to make sure the access controls are set up to let agents work with. And then you have to actually have a whole ecosystem of how does the agent actually get that data and automate workflows and where should the human be in that workflow and how do you actually automate that process? That's the work that we do and we just see it. It's a real amount of work. It's going to take people, it's going to take FDEs. In many cases, it's going to take system integrators. So this is the journey that I think the whole industry is on.

Speaker B: Well, it feels like Box is right in the right position now.

Speaker A: We've been waiting for this moment. Yeah. So we're pretty excited. Excited.

Speaker B: I love it. Well, well, well. Good luck. And, and, and, and you've been very optimistic about other things. I know you're investing as well. Sometimes.

Speaker A: Yeah.

Speaker B: What are your. Some of your favorite things you're seeing as an angel, like how should be thinking about the possibilities there?

Speaker A: Yeah, I think the, um, you know, you get these moments every maybe 15, 20 years, uh, where, you know, I think, I think as we've seen, you need a fundamental kind of market shift for new startups to emerge. Like something has to click about the market that changes, you know, for the incumbents versus the, versus the insurgents.

Speaker B: Big new possibilities.

Speaker A: Yeah, like you can't, like, it's hard to have like complete stasis and then like new startups just emerge.

Speaker B: You can't do Uber in 2018.

Speaker A: Yeah, yeah, literally. Or, or, or 2004. Because like, you know, you need, it's like, it's like this really kind of like, fine balance, which is like you need a new technology that the incumbents don't want to adapt to that produces a new business model. And that's like the moment. So we had one in, uh, you know, the, the kind of mid to late 90s. We had another one in kind of the late 2000 and like late 2000s, early 2010s with kind of cloud mobile SaaS. And then we, we actually had kind of a dark period for about, you know, 10, 15 years. Where, where if you. A very brief dark period. And um. Wait, what? Why?

Speaker B: What? No, it's true. It was not, it was not quite as good a time to start something that's going to be hyper growth.

Speaker A: Yes.

Speaker B: During those years. Yes.

Speaker A: Is there something that I like missing? There's still $3 trillion new possibilities.

Speaker B: We saw a lot of us built multi billion dollar companies. There weren't as many like super giant things. There's no right in the circle.

Speaker A: There's no question that, that every one of those years there was a $10 billion company being produced. But there was a lull for probably seven years where you didn't exactly know where was the market entry point, what was the thing.

Speaker B: It was harder to build really big things.

Speaker A: It was hard because it was all the incumbents had saturated a lot of the markets, et cetera. So enter 2022, 2023. We start to understand now, wait a second. These models are going to be really powerful. They're going to improve exponentially. The only way they're going to be useful is if they get applied to real problems. Uh, the labs are going to do a really good job at building the actual capabilities of the models. They're going to get better and better. But then there's this layer that's sort of needed between the model's capabilities and the ultimate customer's workflow. And that's kind of this bridge layer. And that's where effectively, uh, probably trillions of dollars of market cap will eventually emerge. And you were one of the earliest pioneers of that with Palantir. Like how do you bridge, you know, this technology progress with the real world environment? There needs to be some kind of layer between these two things. And that, that's going to produce a variety of really interesting opportunities in that layer. So there's going to be, you know, legal opportunities, finance, there's going to be marketing, there's going to be hr. There's going to be like everything in that layer will be able to be built out. And then there's going to be various levels of infrastructure that also created that power, that layer as well. So that's, that's kind of the space that's very exciting. Um, you know, in general for my personal, uh, kind, uh, uh, interests, uh, from an investing standpoint, first of all, because we do a lot in sort of knowledge work, I can't do large kind of groups, uh, of investing. But the stuff I am able to do are usually things that are like, okay, we're going to push the frontier of coding, we're going to push the frontier of cybersecurity, we're going to push the frontier of more applied intelligence, um, and the training of these models. Um, and so I Think there's a lot of stuff around those domains that are very exciting. Uh, and. And that that's just going to create a tremendous amount of opportunity.

Speaker B: And last question. We started the podcast to push back on a lot of cynicism and pessimism you're seeing around our country. And obviously the doomers are making a lot of noise right now. We've spoken at length about the economic theory of why they're wrong. But what gives you confidence about, uh, 20 years from now, America's going to be in a great place with AI?

Speaker A: Ah, yeah. I mean, uh, I think the spirit of. I mean, you know, our country is just insanely entrepreneurial, um, and you kind of travel everywhere and, uh, it's just like. There's nothing. There's just no comparison to our ability to organize capital and talent and ideas to launch into anything that people want to be able to create. Um, it's obviously the greatest country in the world on a bunch of dimensions, but this dimension especially, we're just very unique. Um, I think there's always these sort of things that we have to keep improving on. I think that you need to keep pulling in great talent from around the world to help us with these ideas. You need regulatory frameworks that support this innovation, and we need to not curtail that. Um, you want a certain degree of optimism, uh, at the very top of the country at all times that can point into these futuristic directions, which does make me a little bit nervous about the next election cycle, because I think there's a chance of. We actually don't capture all of the exciting breakthroughs that are happening. When I look at Bernie Sanders and talk about data centers. So you actually want people to die of cancer. Right? That's actually what we're trying to do, is we don't want to have this amazing progress, uh, that will totally add 10 or 15 or 20 years to people's lives over the next 50 years, um, and absolutely kind of impact preventable diseases from people. That is the choice at the end of the day, between progress and acceleration, um, and not. And there's lots of stuff that is going to be messy along the way. I lean more towards forms of social safety nets that ensure that anybody that doesn't make it through these kinds of transitions, we're doing our best to make sure that we're supporting them while also making sure that there's no cap on the upside of the opportunity for those that are able to go and chase it. That, to me, is the kind of barbell effect you want to be able to go and create. Um, and I, uh, very, uh. Um. Uh, self. Confidently think I could design, like, the perfect system that would do that. Nobody wants to.

Speaker B: Well, you'd have my vote if I got to vote in Democratic primaries here.

Speaker A: Yes. But I think we've merged many of our ideas together. We could come up with a good platform.

Speaker B: So if they let us run things from out here, I think it'd be a lot better.

Speaker A: Um. Yeah. And all of a sudden, 97% of people disagree with that, but, like, the combination of these two ideas would be. Yeah.

Speaker B: Well, I appreciate your optimism for the.

Speaker A: Thanks for having me on. Appreciate it.

Speaker B: Appreciate it.

Speaker A: MAN.