The Twenty Minute VCAaron Levie: Everyone is Wrong; We'll Have More Developers in 5 Years
EVERY SPOKEN WORD
60 min read · 11,577 words- 0:00 – 1:51
Intro
- ALAaron Levie
Yeah, I'd still be probably loading up on all of the frontier rounds. These numbers could continue to get much larger
- HSHarry Stebbings
Now, I think Aaron Levie is one of the luminaries on AI pervading into enterprise, and he did a viral tweet the other night, and I said, "Dude, we've got to do a show on this." So this is specifically on how AI will impact the biggest enterprise in the world, how agents will be introduced into the largest enterprises, and we couldn't have anyone better than Aaron, founder and CEO of Box, one of the public companies of the last decade. This is an incredible discussion.
- ALAaron Levie
What we are in is a commercial and economic race. We haven't removed humans from the loop. We've just changed where they enter the loop. Everybody is so myopic about this. I want to just, like, shake the industry. There are going to be more lawyers in the next five years than we have today. The workflow needs to be redesigned for agents, not for people. The budget of tokens will have to move out of IT spend and into regular OPEX spend.
- HSHarry Stebbings
Is your job harder than ever?
- ALAaron Levie
Yes. If you're in software or infrastructure or building agents, it's a year of complete unrelenting execution.
- HSHarry Stebbings
Ready to go? [upbeat rock music] Aaron, dude, it's so lovely to have you on the show. You know that we have Rory on every week, and he's just like, "Aaron is the greatest." I'm not going to do his accent because I suck at them, but he's like the greatest-
- ALAaron Levie
Wait, you can't, you can't easily do a, an Irish accent?
- HSHarry Stebbings
Well, you know, I, I can't really do the Irish accent so well.
- ALAaron Levie
Okay. [laughs]
- HSHarry Stebbings
But, uh, that's fine. Exactly. Uh, but you basically, I'm sure, bought Rory one of his houses, so no wonder he's grateful. Uh, but [laughs]
- ALAaron Levie
We, we... Rory, uh, we got more out of Rory than he got out of us, so, uh, uh, I'll, I'll take that.
- HSHarry Stebbings
An exceptional man. But I wanted to start, and we were just chatting. I was running around the park listening to the Dwarkesh and Jensen episode, and I was like, "I don't think Jensen came out very well."
- 1:51 – 5:12
Why the Experts are DEAD WRONG About the US-China AI Race
- HSHarry Stebbings
Do you agree with me that Jensen didn't come out very well from that episode?
- ALAaron Levie
Um, I, I think this is, like, the greatest Rorschach test of all time of, of where, uh, where, where, where somebody is mentally on AI. Um, I, if... Uh, so I happened to see a bunch of the tweets before I watched it, and so I was a little bit obviously biased in advance. But if I hadn't seen any of the commentary and I had just watched it, um, I would've been very confused by the, by the commentary post, uh, you know, post, uh, interview. Uh, a-and, and, um, to be clear, I kind of jumped to the more, uh, salacious part of, of, you know, China and, and that topic. But I'm, I'm almost probably 80% with Jensen. Um, and, um, and I, I, my, my sort of, uh, kind of way of thinking through the logic actually works much closer to, uh, to Jensen. Um, you know, the, the idea that we're in some kind of, you know, kind of race, existential race where a month or two of advantage is gonna, you know, change the total outcome of, of AI progress and, and what everybody does between us and China, I, I just don't agree with. I think what we are in is a commercial and economic race. Obviously, with safety, you know, built into that, there's no question. Um, and I think we actually have a lot more power globally, uh, if it's our technology stack that's powering AI. And so I, I, I kind of am more in the camp of, of Jensen on, on hi- uh, on his lines of logic. And, you know, Dwarkesh kind of oversimplified a few, a few components. You know, he said, "Well, you know, with Mythos, if, if we get early access to that, then we can go and upgrade all of our systems." And, you know, with, with, again, great respect to Dwar- Dwarkesh, it's like, it's like upgrading software is a multi-year effort, so unless they somehow keep Mythos, you know, closed for the next decade, there's not, like, some magical moment where you can just secure everything. This is an ongoing, endless, you know, till the end of time, you're always in this sort of leapfrogging, you know, between the defensive side and the offensive side. Um, and so I just don't think these things are as binary, and so I, I actually more am inclined to, uh, to, to Jensen's view of that. And then Jensen had a really key point that was... didn't go viral yet, so maybe you could kick it off. But he had this little small vignette, uh, about 90 seconds into the whole conversation, where he said, you know, "We're gonna do ourselves a disservice if we scare people out of engineering, if we scare people out of radiology, if we scare people out of healthcare because they think all these jobs are gonna get eliminated with AI. That is not helping us. Uh, that is, that is, uh, it's doing a disservice to the next generation. It's doing a disservice to society as a whole." Like, we, we don't yet know any way to use AI in a capacity other than augmenting our work, where we still eventually have to go and review the work in some, in some form. Maybe you don't have to review the tiny little parts of it anymore. You can review a, a bigger, you know, part of the, of, of the, of the, you know, work product that happens. But, um, but we haven't removed humans from the loop. We've just changed where they enter the loop. Um, and, uh, and I think that, that Jensen has a more pragmatic view of, of the technology. We should be, you know, very thoughtful about how we make these systems safe, but I much more land in Jensen's camp, uh, on the overall kind of contours of the debate.
- HSHarry Stebbings
Oh, Jesus, dude, you didn't leave me... Uh, you didn't... Uh, it was quite a lot to unpack.
- ALAaron Levie
I didn't have, I had a coffee. I just wanted to have a coffee.
- HSHarry Stebbings
Okay.
- ALAaron Levie
I'm sorry.
- HSHarry Stebbings
Uh, okay, first,
- 5:12 – 9:27
Everyone is Wrong About Labour Markets: You Will Not Lose Your Job
- HSHarry Stebbings
a disservice, uh, by discouraging people to go into categories like radiology or engineering.
- ALAaron Levie
Yeah.
- HSHarry Stebbings
Do you think you will have more engineers at Box in five years' time?
- ALAaron Levie
Uh, we will and, and I think the part that everybody misses is that the-- everybody is so myopic about this. I wanna just like shake the industry. We are s- we are so myopic and, and self-interested and, and we think that the entire industry is the tech industry. And when you go around the country or world, and you go and talk to a tractor company and a bank and a pharma company, and you ask them, "Do you think you have enough engineers to go and automate what is gonna happen in your industry going forward?" They absolutely, unequivocally, universally always say no. And so what, what the breakthroughs of CloudCode or Codex or others are, are doing is it's making it so those companies now can actually do the same kind of engineering that Silicon Valley's been able to do. And so we are myopic because we think that, that tech is the only use of, uh, of engineers. And tech is only, I don't know what the ri-right number is, eight, 10, 12, 15% of, of GDP in the economy. What happens when 85% of the economy now gets access to engineering like tech has always had? That is what will happen that-- And, and so yes, maybe if you're graduating, you know, name your computer science school today, you don't go immediately to Google. You go to literally John Deere or Caterpillar or Eli Lilly, but the skills that you have are gonna be just as relevant in just a different domain. You're not gonna be building a little app with little buttons. You're gonna be automating pharmaceutical research. You're going to be doing AI for the future of, of, of, you know, farming and industrial equipment. So we're just too myopic about, about how this works. And, um, and, and, you know, uh, uh, you can already start to see this sort of playing out. There was a really funny FT article, which is lawyers are being inundated by all of these kind of AI responses that they're now getting from their clients saying, "Hey, can you review this contract," or, "Can you review this memo," or, "Can you look at this, this case?" Well, guess what happens when everybody thinks that they're a lawyer? Do you know what the ultimate constraint is? The ultimate constraint is the actual number of lawyers that, that actually are able to go and review all of this stuff being produced. So like, I would take the other side. I'd rather-- I-- Like, like there are gonna be more lawyers in the next five years than we have today because we've made it easy to generate legal content, but it has not gotten any easier to actually get any of that approved by any court system, or file a patent, or any of the things that law actually ends up relating to. So these are, again, this is where I, I, I just differ from the rest of the industry.
- HSHarry Stebbings
Do you really think so? With the greatest of respect, we are seeing the eradication of kind of lower ranking legal positions.
- ALAaron Levie
And, and that, that is a different issue, which is how do you do the next generation of mentorship and apprenticeship when AI does automate the maybe, um, traditional tasks that those workers are doing? Big question, a big question facing every bank in the world, every law firm in the world, anybody who had a, a sort of an apprenticeship model. I don't doubt that that's a real issue. But that's different from the constraints that, that all of this work ends up resulting in that you still have not been able to automate. We had a customer conversation two weeks ago and, and, and this is just gonna sit with me forever. I'm gonna always have this example. They've automated or they're working on automating patient referrals when, you know, when you wanna go and see the radiologist or the, the high-end doctor for whatever issue you have. They're automating that, which is awesome, so now you don't have to be on the phone for, you know, a week or whatever. Well, guess what? You can automate anything, but if it still is 18 months out before an appointment is available, what, what-- your ultimate constraint is still the healthcare institution and the amount of doctors we have, and actually the amount of, of, of real labor we have across those organizations. So yes, maybe, maybe you, you don't want to, you know, stake your career on being a frontline, you know, customer ser-service rep in healthcare right now, but, but first of all, that same person will have a lot of other types of jobs that they'll have access to. Um, but you still will end up having all of these other constraints that, that, that eventually we will need to produce more and more jobs to go and resolve. So automation is gonna actually just force us to see the next set of bottlenecks that are in all of these industries that we didn't perceive that we had before because everything was so slow and manual.
- 9:27 – 12:57
What Role Does Not Exist Today But Will Be So Common in 5 Years
- HSHarry Stebbings
What job title does not exist today that will be incredibly prominent in five years' time?
- ALAaron Levie
Yeah. So I'm workshopping, and, and a bunch of people are, are doing this, so this is like not my invention. But I'm, I'm workshopping-
- HSHarry Stebbings
Aaron, Aaron, Aaron, you've, you've gotta take attribution. As a venture investor-
- ALAaron Levie
Yeah. Sorry
- HSHarry Stebbings
... it's all about coining a term, okay? This was your original thought, okay?
- ALAaron Levie
This is my original idea. Yeah.
- HSHarry Stebbings
Have it in the shower.
- ALAaron Levie
Yeah.
- HSHarry Stebbings
Aaron Levie's.
- ALAaron Levie
This is-
- HSHarry Stebbings
Share it with me.
- ALAaron Levie
I've been influenced-
- HSHarry Stebbings
Yeah
- ALAaron Levie
... by nothing I've seen online. Uh, this is all from me. Um, so there's some kind of... And, and who knows if this sustains as, as a full-time role or where it gets diffused into. I'm not, I'm not, uh, I'm not 100% clear on that. But there is 100% a role right now that there's gonna be 500,000, a million jobs created for, and, uh, and, and it's basically some kind of agent operator. And, and this person is, um, is actually gonna be needing to be, uh, somewhat technical. They're gonna have to like be deep in the AI world. They're gonna have to understand MCPs and CLIs, and they're gonna have to know how to write skills. They're gonna have to understand agents.md files. The, the-- it's gonna be this group of people that will know how to go into your marketing team or your legal team or your operations team or your life sciences research team. And, and this is the person that is basically going to enable that function to get leverage from agents. And, um, and the problem that the real world has that startups and, and frankly many of your guests don't understand is that, is that when, when you start a company from scratch, you've got like, you know, the world is your oyster, right? You can design your workflows however you want. There's really no risk if you-- if something goes wrong 'cause you don't have much scale to begin with. There's no real regulator that, that is sort of calling on you to say, "Hey, you know, are you doing things the right way?" It's, it's, it's effectively infinite upside and white space. When you go into a, a, you know, a Fortune 1000 pharma company or bank or, or, you know, consultancy, um, that doesn't-- that's just not the case, right? These guys have-- They're regulated. They have data fragmented across their organization. They have employees that are, that are sort of wired to do workflows a particular way. So there needs to be somebody that can basically say, "Hey, if we actually wanna get real leverage from automation, we need to start to redesign the workflow that we're doing." And the workflow needs to be redesigned for agents.Not for people. And so, so what do you do when you, when you reimagine a business process where the agent is now doing much more of the work than, than what, you know, than what the human used to do in that process? And that just means it's a, it's a very different sort of implementation cycle. It's, there's real change management. You've gotta get data organized in the right way. You've gotta connect up systems in the right way. Guess what? The second a new model drops, your workflow probably breaks, um, because the way you prompt that agent now is different. There's a different way that, that it wants its index to be, to be handled. So that, it just requires care and feeding and, and, and a real level of kind of technical and business process acumen. Uh, so I think we're gonna create a, you know, an untold amount of jobs that look like that. Some of those people will come from IT. Some of those people will come from operations. Some of them will come from engineering, uh, if you're in a maybe more technically inclined company where it's like the next generation of if... You know, there's a, a limit to, again, the number of software you need to build that looks like an app on your phone. There's an unlimited amount of software you need to build that looks like a background system process that's connecting different data sources, automating workflows. That's where the, the work is gonna
- 12:57 – 18:57
Is Your SaaS Tool Actually a Valueless Database in an Agentic World?
- ALAaron Levie
go.
- HSHarry Stebbings
But this was really gonna be one of my main questions, which is, you know, Jensen very clearly said, "AI won't kill software. It will explode the amount of software needed."
- ALAaron Levie
Yeah.
- HSHarry Stebbings
And when I thought about that, you know, the thesis there is obviously you kind of, you have this kind of core AI that crawls over 15 SaaS tools, and they really become databases that agents crawl on top of. Is that what it looks like? And are they not just valueless SaaS tools then?
- ALAaron Levie
Yeah, I mean, I, I think that, that I'm, I'm sympathetic to that argument in some, in some categories. I think there's some software where because the person was the user of the, of the software and they were clicking all the buttons, that your sort of ratio of buttons to underlying APIs was, like, more in favor of buttons, and I'm, I'm oversimplifying. But there, there are some tools where you open it up and there's, like, 93 features, um, that you're kinda clicking around on, and the, and the user has been so accustomed to exactly how to do that that the, that the software's value proposition was correlated to, to roughly that, that sort of mass. In a world of APIs a-and a world of agents being able to do more of the, of the work that you used to do on clicking those buttons, then, then again, the, the value goes more to the API layer. So then the question is, how many APIs do you have? Not, not in, like, a you just need 1,000 APIs, but, like, like how robust and useful and, and proprietary and how much business logic is embedded in those APIs versus it's just calling a database and pulling a record. Like, does the API surround a, a, a set of business logic of, like, no, it actually secures the data, or it knows exactly what person each piece of attribute should have access to inside the organization. That's, you know... A-at the end of the day, all software has a database behind it, so you could oversimplify it and be quite reductive to that. But there's a lot of business logic in the layer above the database that, that software players have. Like, if you're an ERP system, y-you know, you're way more than a database at this point because you've written a tremendous amount of business logic of how your supply chain should be automated and work and how you should do accounting. None of that goes away. So then the question is, what changes is the user interface that the, either the user or the workflow is interacting with?
- HSHarry Stebbings
Mm.
- ALAaron Levie
The user interface might be now you're just chatting with an agent. I think increasingly the right way to do this is there's some kind of agent in the background that's connecting multiple systems, so you're, you're not even, like, the user's maybe not even seeing half the value that's happening, but the agent is sort of working across an ERP system, a CRM system, an HR system, a, you know, a document repository, and then doing work across those systems, which means that the value proposition has to be how good are your APIs? How well designed are they? Are they ready for agents? And then can you monetize that in some way that makes sense? And, and we, we are treating software too, too much like one gigantic sort of, of monolithic industry and, you know, it'd probably be better to have some kind of two by two, which is like how much business logic is there? How much sort of human to agent, you know, collaboration does there need to be? And, and, like, so the reason I bring that up is the moment you have human and agent collaboration, you need some kind of, you know, usually you need something that the user can pop into to, to experience the work that the agent did and, and that probably doesn't go away so much. Um, uh, and then, and then, you know, when more agents are working on the software, which parts of software do the agents need those APIs even more than humans ever did? And I think there's a lot of categories of software where actually agents using the tools is a massive boon for the technology as opposed to a dilemma.
- HSHarry Stebbings
Where will agents use the tools more than humans do, and those API calls become much more frequent?
- ALAaron Levie
Yeah. I mean, an easy one is just unstructured data. You know, agents are going to be this incredible consumer and creator of your unstructured data. They're gonna read through every one of your contracts and generate all of your contracts. They're going to generate marketing assets. They're gonna write reports for you. And so when it becomes trivially easy for you to generate all this new information or have agents review it all, well, guess what? You still need a backbone that kind of manages and coordinates and creates the guardrails of those workflows and, and all the agents doing that work. So we're about to see an explosion of unstructured data as an example. Um-
- HSHarry Stebbings
With the greatest of respect, Aaron, can I just interject?
- ALAaron Levie
Please. 100%.
- HSHarry Stebbings
Does that increase the value of your business?
- ALAaron Levie
Yes.
- HSHarry Stebbings
When I think about that, I, I asked Aaron from Monday, if you become, you know, a data repository which agents crawl on top of, how do you retain value in that?
- ALAaron Levie
Yeah.
- HSHarry Stebbings
I would ask the same to you with respect.
- ALAaron Levie
Yeah, 100%. It, it's, it's the question on the mind of every investor on the planet right now, so we're, we're used to it and it's not a [chuckles] it's not, it's not a scary question. Um, o-one thing that, that helps us is we've always had a, a, a, an API sort of maybe not first, but equal strategy. Um, uh, if, if I told you the number of API calls we did last year, you'd, and, or you guessed first, you'd probably be off by an order of magnitude. Um, uh, so, so the volume of, of API usage on our system is already enormous and already, you know, is outsized relative to any of the end user interactions in the system. Um, and that's just a virtue of you use content in a variety of applications and workflows that, that-You know, far exceed what, what people, you know, kind of, you know, open up their finder and, and upload a document into. Like, like an ERP system generates files. A, a wealth management portal, you have clients uploading documents into the portal and they never see Box. Um, you have workflows of invoice processing that's happening behind the scenes. So the headless version of Box has been alive and well for, you know, almost since the day we started the company. And so agents to me, just again, represent a, a force multiplier on that. So it's not a-- it's, it's, it's actually an exciting proposition for us. We already know how to monetize it. The question is like, will the exact dollar and cents be the same between an agent user and a previous application user? We don't know. But we do know that if the number goes up by 100X or 1,000X, that that's actually more opportunity for us in the future. Now, that's not all the same for all software providers, but for where we sit in the workflow, where you just, you generate a document, it needs to go somewhere, you have to secure that, you have to protect it, um, you have to govern it over the long run, that's just more data going into our platform and, uh, and that's, you know, so that's why it's just all upside
- 18:57 – 22:41
The Cybersecurity Tsunami: Why Agents are Your Biggest Threat
- ALAaron Levie
for us.
- HSHarry Stebbings
You said secure and protect it. We mentioned our mutual love for Rory O'Driscoll. I do a show with Rory and Jason every week. Jason's bluntly said that this will be the golden age for cybersecurity because the security threats are going through the roof.
- ALAaron Levie
Yep.
- HSHarry Stebbings
Are you concerned with the system vulnerabilities and the security threats that are coming with AI? And what do we not know about security that we should know?
- ALAaron Levie
I am concerned, but not, uh, in any kind of like new concern sense. Uh, this, this to me was kind of priced in the moment that we were generating code with AI. So if you can generate code, you have two problems. One, you're gonna generate way more code than anybody's ability to review that code. So, you know, starting with GitHub Copilot six years ago or whatever the date was, five years ago, like that was just priced in, which is, which is as soon as, as AI writes most of the code or-- and then like 90% of the code, and then 95% of the code, then by volume, we're just gonna produce this unbelievable amount of code. And, and any, you know, any change in a system, uh, you know, everybody kind of thinks about security as like, um, you know, is there a zero day where there was an unpatched, you know, component of your technology or somebody found a clever new packet, uh, package that, that, that you could kind of slip into. Uh, every time you, you ship a new feature, you have a chance of a security vulnerability because the, the AI could have written in, "Oh, you know, we wanna actually open up that port in, in the system because we need to do something," and maybe that was the wrong decision for the agent to go and do. So, so we're gonna be living in this new world of, of cyber risk, uh, in the form of, of using agents more. And then on the other side, obviously if you have the offensive side able to use AI, probably more in the form of open models and, and whatnot, then they can find more vulnerabilities because they can scan across the internet far faster than before. So you actually have two new forms of risk in the development process, and you only have one benefit, which is agents can also review the code and, and try and keep it secure. So, so it's, it's, it's gonna be a, a very dynamic, um, uh, you know, period. I t- I think, you know, for better or worse, agents are the solution to the problem that agents have caused and, um, and, and that's why there's gonna be a lot of money made in agentic security, uh, as well.
- HSHarry Stebbings
You said agents are the solution, um, to the problem that agents have caused. It almost reminded me of when Jensen went on TV and was like, "Oh, every engineer should be spending..." I can't remember the amount. I think it was either 250 or 500,000.
- ALAaron Levie
Yeah, or like half the salary essentially.
- HSHarry Stebbings
Yeah, yeah, yeah. And it's kind of like, you know, drug dealer, you should buy drugs. Well, [laughs] okay, no shit. Uh...
- ALAaron Levie
I, I, again, I'm, I'm, you know, obviously Jen- I mean, you know, listen, we love Jensen for that level of, of, uh, uh, of, you know, grandiosity and, and charisma. So I, I actually, I, I, but, you know, whether he's off by half or not, I mean, directionally the idea is actually pretty salient, which is, which is you're gonna be spending more on compute per person in the future than, than you ever thought, and that you certainly are today.
- HSHarry Stebbings
What percentage of salary are you going to spend on compute in Box in five years?
- ALAaron Levie
Uh, great, great question. Uh, I don't think we've modeled that out in, in five years, and obviously the joy of being public is-
- HSHarry Stebbings
Well, this is your chance. This is-
- ALAaron Levie
No, no, totally. I, you know, I was told not to model long-term financial projections on podcasts, so, um, so let me, let me... Yeah, it's a weird SEC financial thing.
- HSHarry Stebbings
That's so boring. God.
- ALAaron Levie
It's, it's a true story. Don't ever go public if you don't wanna model-
- HSHarry Stebbings
[laughs]
- ALAaron Levie
... on, on, uh, on podcasts, so.
- HSHarry Stebbings
Well, I'm, uh, this is why the Collisons don't. Everything else is great. They just didn't podca- Yeah, cheeky Pine Hill level.
- ALAaron Levie
I, I, you know, I don't know if you, I don't know if you'd be able to pin Patrick or John on the same question for their five-year view, but, but, you know, it'll be a larger number for sure than it is today, so.
- HSHarry Stebbings
I'll smash them with four tequilas and then ask them.
- ALAaron Levie
[laughs]
- HSHarry Stebbings
Um,
- 22:41 – 33:53
Token Maxing: What Every Company Needs to Know About Budgeting Tokens
- HSHarry Stebbings
you, you said one of your observations in your very viral tweet, uh, was about token maxing-
- ALAaron Levie
Yes
- HSHarry Stebbings
... and kind of token allocations within enterprises. I'm really intrigued. How do you think about advising CIOs on token allocation-
- ALAaron Levie
Yeah
- HSHarry Stebbings
... token maxing, what we should know that we don't know?
- ALAaron Levie
Yeah.
- HSHarry Stebbings
How do you think about that?
- ALAaron Levie
This one's tough. It's, it's, uh, it's, you know, the, uh, the general advice will end up sounding kind of like, um, uh, you know, kinda generic by definition. Um, it, it, you know, usual- I mean, it's gonna have, it's gonna have something to do with, with your tokens will have to correlate to where there is the most amount of, you know, value generated for your company. Like, like most bland statement of all time, but just obviously has to be true. So in the software industry, we're into token maxing because guess what? Like generally the value proposition of your company will correlate to how much software can you produce. And so, so, you know, if you're trying to drive a lot of change and you wanna make sure everybody's shipping lots of soft- software and you wanna be able to teach the best practices faster, then token maxing and leaderboards are an interesting way to do that. I, you know, it's not obvious that you're gonna see that across every industry. Um, uh, we've, we've seen a couple of interesting examples. One company had this sort of like Shark Tank pitch-a-thon type thing, which is, you know, teams have to show up and they have to go pitch for, for compute, you know, token budget, and then you kinda allocate it in some central fashion like a VC would, and then you sort of... You know, I don't know their exact interval, but I would imagine you review that three months, six months in being like, "Okay, did you get the upside that you thought on, on that token usage?" So that, that's an interesting one. I think you have aAnother company had a kind of a view of like, you know, it's, it's some kind of like natural stratification of, you know, 5% of your users are doing the most valuable things, 20% are doing the next tier of most valuable things, and then everybody else is sort of doing general productivity. I'm, I'm making up their numbers, but, but the idea would then be like, well, for that 5 or 10%, give them the g- the best models with unlimited capacity. For the next 20%, have some, have some limits, maybe it's a, a little bit more efficient of a model. And for everybody else, it's sort of like we're gonna just use the cheapest thing on the market. It's, it's not gonna be like the game changer of the employee productivity. Um, and so I think everybody's kind of working their way through this. The, the part that, that back to Silicon Valley's again, kind of, you know, sometimes more, um, uh, you know, y- y- let's just say like, like positively naive view is, is like c- real world they have like budgets, and they have like annual budget planning cycles 'cause they have EPS numbers they, they commit to Wall Street. And so you don't get to just be like, "Oh, we're gonna token max across the enterprise where everybody gets unlimited token budgets," because obviously then that company would just miss their, their earnings throughout the year. So you have to like wait for the earnings... You have to wait for the budget cycle, you have to figure out what teams, you know, make-- are, are, are most interested and, and have the best use cases. That's a, that's a natural journey. One final bookmark... One final, one final bookmark, um, that, that I think is, is well understood now at this point is the budget of tokens will have to move out of IT spend and into regular kind of OpEx spend. This can't be treated like a, "Oh, you know, I'm gonna trade off between Salesforce licenses or, or, or compute tokens." Like, like it's gonna more be, "I'm gonna trade off, you know, this next marketing campaign, uh, and, and instead I'm, I'm gonna go and drive more automation in our, in our marketing engine." Like it's, it's gonna be that kind of, of set of trade-offs.
- HSHarry Stebbings
What happens to that token budget when it transitions to that different spend category?
- ALAaron Levie
Um, well, first of all, it goes up because-
- HSHarry Stebbings
[laughs]
- ALAaron Levie
... because, no, because IT spend as a percentage of revenue of, of large enterprises is, is-
- HSHarry Stebbings
But is this the same as the kind of classic VC blog post which every fucking firm has written, which is like, "You know, AI, it's moving from software budgets to labor budgets." And every partner goes and likes-
- ALAaron Levie
Yeah
- HSHarry Stebbings
... the tweet, and there's like no fucking shit. [laughs] Like really?
- ALAaron Levie
Well, yeah. Yeah. Uh, I mean, if you do it in that voice, it sounds, it sounds kind of like, um, you know, maybe, uh, you know, simple. But like, yeah. That, that, but like that's just like a very big deal in technology. We've never had-- There's never/rarely been a technology that you could sell into an enterprise where you weren't capped by that company's corporate IT budget. And so now for the first time ever, you have a technology where you can go into the line of business and you can say, "I can now offer you a, a, a new tool in the form of an agent that will augment a workflow that will make you 50% or 100% more productive, and so maybe I should be able to get 5% of your OpEx budget this year to go and do that." Like that, that is a new budget to tap into, and I don't think it like, you know, 10X's the size of, of IT spend or, or technology spend globally, but it certainly doubles it.
- HSHarry Stebbings
I mean, current, uh, enterprise technology spend is estimated between 10 and 12%. To see that going to like 20% as you said there is like, I think relatively feasible. You said about kind of companies being like based on earnings per share and actually having budgets that they have to adhere to. Very strange not to have venture-funded companies. Um-
- ALAaron Levie
[laughs] Yeah. They don't have unlimited VC to, uh, to go and solve this.
- HSHarry Stebbings
Can't we just go to our venture investor and ask for more money? Um, the one thing that I worry about is we see this insane demand side pull. Every company in the world needs an AI story. Everyone wants to kick the tires with something, and I think we project the same demand side pull and extrapolate it continuously. Do you worry that we are in a momentary 18-month period on the demand side pull, and that may not always be lasting?
- ALAaron Levie
You know, it's very possible I should be more sensitive to that. Um, uh, but, uh, I, I, I would take the opposite side of, of, of that particular wager at the moment because, um, partly because we, I, I already saw one diffusion cycle with cloud and actually how long that ended up taking and the, and, and the, the, the kind of spiky early nature, you would've just been like, "Oh my God, this is, this is on fire. It's, it's... and how could this last?" And 20 years later it lasted and got way bigger than we ever realized. If it works, the market's always larger than you ever think. And, um, and then the only, the only part why, why 18 months is like not even a relevant window to me is I think diffusion is gonna take longer than Silicon Valley thinks, and it's back to the very first kind of that new role idea. When you go to most companies, they can't yet just deploy an agent to do, you know, full, uh, you know, financial proposals for all of their, their clients without a human reviewing the thing and because the SEC will just show up and be like, "Hey," like, "you, you just, you just gave this person bad financial advice, and you're gonna lose your license." Like that, that will just start to happen kind of across the board. And so, um, and, and so that, that's why, you know, people take time. That's why we, we, we-- there's a lot of regulatory controls and compliance teams, security teams have to figure this out. That just takes time in the economy.
- HSHarry Stebbings
I, I had, I think Matt Fitz- Fitzpatrick from Invisible, which is like a Turing or a McCool competitor.
- ALAaron Levie
Okay.
- HSHarry Stebbings
And he said, "You cannot sell into enterprise without an FD model. It is impossible."
- ALAaron Levie
I mean, it rounds to being true. So-
- HSHarry Stebbings
Okay
- ALAaron Levie
... yeah.
- HSHarry Stebbings
Super interesting to hear that because we're seeing like the rise of, "Oh, we go PLG, and then we like seep up into enterprise." So isn't that-
- ALAaron Levie
Well, P... Well, I, I, I don't-- Oh, sorry. I wouldn't... I, I, I don't, uh, think of those as, as mutually exclusive for what it's worth. Um-
- HSHarry Stebbings
I guess what I'm saying is when you think about adoption within the largest enterprises-
- ALAaron Levie
Yeah
- 33:53 – 39:23
Is Silicon Valley Secretly Being Powered by Open-Source CCP Models?
- HSHarry Stebbings
You said before when I've tried the latest model, it's got like 85% of the way there.
- ALAaron Levie
Yeah.
- HSHarry Stebbings
I speak to many of the best early stage and more mature, you know, We- West Coast based companies and they say, "Hey, we use frontier models to set where we can be, and then we use open source Chinese models to get as close as we can to that frontier benchmark." Is Silicon Valley being funded by a generation of open CCP funded open models?
- ALAaron Levie
I mean, that, that must be kind of empirically true. Um, I, uh, I, I, I don't have the same kind of like, uh, oh, that's so scary, you know, kinda element. Now, obviously, again, holding out some, some element of risk of, of some, some backdoor waits that, that can get triggered at some moment or some parameters, but like, like, like I'm not ... I, I just ... Like, that's not how I'm perceiving it. But, um, uh, but yeah. But also that's ... I, I would say that's kinda orthogonal to my point about, like, the best frontier model still will go and do the wrong thing. Uh, and so thus I have to be in the, I have to be in the workflow loop to make sure that I review its, its work.
- HSHarry Stebbings
You know, as a venture investor, I specialize at making bold statements with little substantive evidence. Um, if you-
- ALAaron Levie
It, it, it, it worked for the greats, so, uh, yeah.
- HSHarry Stebbings
Do, do you know what? I'm just following their lead.
- ALAaron Levie
Okay.
- HSHarry Stebbings
Jason Lemkin, my dear friend, says, "Why has no public company created any good agent product? Everyone creates 60% shit agents" But he's like, "The one person who's done it is Palantir, and no other public company has created a sufficiently good agent product." Why is that?
- ALAaron Levie
Um, you know, I, I don't know that I can fully endorse the point, but I'll, I'll, I'll ... I can give you the ... Because I, I, I would argue our agent is sort of the best agent for working with content. You know, this is a very fast moving, uh, space, and you have to be kinda wired in at a level that, that I don't think you've ever had to be wired in in tech. Um, uh, like I am, uh, you know ... And, and the information sources aren't the classic ones. It's not the, it's not the roll-up review two weeks later from your traditional news publication that is gonna give you any kinda alpha. It's, it's, it's the practitioner who's the, you know, literally the engineer at the, you know, agent sandbox company and their, their long form article on how they are handling, you know, memory and the harness, and, you know ... Uh, like, like, like if you're not wired into that ecosystem, it's very hard to then have your team, you know, be at the kind of forefront of all of what is happening. And so it just is a, it's a different pattern than what we've ever had to do. Like, like, you know, COVID was, was pretty crazy. Like, we all had to kinda like hunker down and be paying attention to daily news cycles on, on COVID-y stuff. Um, but it wasn't like a tech problem. Like, it wasn't hard technologically. Um, so there's, like ... There's, but, but there's not been a moment before where the speed of change and responsiveness you have to have is quite literally on a multi, uh, you know, multiple times a week cycle.
- HSHarry Stebbings
Is your job harder than ever?
- ALAaron Levie
Yes
- HSHarry Stebbings
Because of that speed of transience of superiority of technology?
- ALAaron Levie
Yes. Uh, you have, you basically have this, this, you know, component of, one, there's a tsunami of change that you can just feel. And so you're, you're like, "Okay, we gotta like run faster than ever before." And then there's a... And then there's just like the pure technical underpinnings, which some of it has business and strategy implications, some of it has product implications, some of it has partner ecosystem implications, uh, because of that tsunami that you have to very quickly kinda wire up what you are doing a- about that shift, uh, and, and where the market is going. Uh, at the exact same time, you have to also be like, you know, find a way to be a bridge for your customers that, that, you know, also don't wanna get, you know, crashed into by the tsunami, and they wanna be able to, to be able to have a bridge into the future. And so there's just, you're juggling a lot right now.
- HSHarry Stebbings
You said your agent product is the best product. Again, Jason Amero said this, and Rory might kill me for this 'cause he gets a little bit more sensitive about when I quote him or misquote him more appropriately. But like, he basically says it, and this is Jason again, "If you can't, like, charge way more for your agent product, that Wall Street doesn't give a shit, that you have to re-accelerate revenue with agent products." Can you charge significantly more for an agent product?
- ALAaron Levie
The, the answer is yes, but, but there's a little bit of nuance, which is our, our business model is we have a new plan tier that we just introduced last year that basically houses our, you know, best workflow capabilities, our business automation, you know, uh, our app- application development capabilities, and then the agent is sort of central to that because it's gonna help you automate the work that, that you're actually doing with your content. So it'll read a document and extract metadata from it. It'll process information inside of a, of a, of a workflow. So that is actually causing a real re-acceleration of our revenue growth. Last year, we, we, we saw an inflection in our revenue growth, and so that, that... It's already happening in our business. Um, and, and so we are, we are doing the thing that I think Rory is, is sort of probably saying is the new benchmark. Now, to be fair to what's, what's happening though is I think Wall Street still is sort of saying we kinda need to just step back and see where everybody lands in this because of how much change there is. So, um, so this is, um, very much a year where if you're in software or infrastructure or building agents, you just g- it's an ex- a year of complete unrelenting execution.
- 39:23 – 46:30
The Brutal Truth: Is This Generation of CEOs Too Low-IQ for AI?
- HSHarry Stebbings
Do you look at the ticker?
- ALAaron Levie
Yeah.
- HSHarry Stebbings
Every day?
- ALAaron Levie
Yeah.
- HSHarry Stebbings
[laughs]
- ALAaron Levie
But I was like a day trader.
- HSHarry Stebbings
I've never met, I've never met a public company CEO who hasn't. But, um, uh, the Nevan CEO was on the other day, and he's like, "Multiple times a day. Multiple, multiple."
- ALAaron Levie
Yeah, no, 100%. I'm-- But, like, but, but, like, partly I just, I have, like, you know, ADHD or something, and so I just ne- like, I'm like, "Ah, there's, there's-"
- HSHarry Stebbings
Is the stock... It's-- Will we look back on this period and be like, "What the fuck? Companies trading at three times cash flow." Like, way over exaggerated or not?
- ALAaron Levie
Well, thr- three times cash flow is, is very much over exaggerated. I would say that we're in a period right now where basically the market is being treated roughly as i- in a, in s- a, as a kinda indiscriminately, you know, kind of bucketed sec- sector. And the next year, two years or whatnot, you'll start to see some separation and parsing between the companies because, as I noted in the beginning, agents will be really good for some parts of software, and agents will put pressure on other s- other, other parts of software. So... And it'll mean some companies have to fully pivot, and some companies can just sort of ride the wave and, and if they respond, you know, effectively. Like, clearly 3X free cash flow is, is like, you know, that, that seems like aggressively low territory. But I also think that at times in software, things have been aggressively overvalued, um, beyond the, the realm of, of likely what the terminal value is of, of a particular, you know, category or, or, or company as well. So, so I think we're, I think there's just a pendulum that, that needs to kind of find its equilibrium right now, um, and, uh, and that, that'll play out over the next year.
- HSHarry Stebbings
Okay, I'm going for spicy. Uh, I, I interview most of the public company CEOs that you know and I know, and Jason Lamkin said on the show, "If Aaron Levie is the best," which y- you're a phenomenal... No, but you're a phenomenal AI-first mind and leader. Uh, so just roll with me on this.
- ALAaron Levie
Yeah.
- HSHarry Stebbings
And 'cause it gets worse, sorry. And even he is struggling with-
- ALAaron Levie
Wait, wait, how am I struggling?
- HSHarry Stebbings
Well, I mean, I-
- ALAaron Levie
I just said I'm tired. I... Well, Wall Street, who... I mean, Wall Street does its things. Like, we're, like, we're not... Like, we're cranking.
- HSHarry Stebbings
Dude, I, I... It's Jason. Blame him. I hate him.
- ALAaron Levie
I even, like, I, I do think, I think that, that there's a little bit of-
- HSHarry Stebbings
Do you think this generation of CEOs that you have around you though is equipped for the AI transformation that is ahead? 'Cause I don't. Like, I think... I, I'm not blowing smoke up your ass. You are. You're so versed in this. You're so fluid. But a lot are like... Now, one said to me the other day, "No, we don't have the AI chops in-house. We might need to bring it in."
- ALAaron Levie
This one's hard. I think, um, I think you still have a lot of kind of founder-led or, or tech, uh, f- you know, forward, whether they were an engineer or just they're, they're just very technical, you know, c- category of folks that are, are pretty dialed in. And like, like I, I have Slack channels and, and WhatsApp groups where people on the weekend are, are just, like, working with Cloud Code or Codex building stuff, and they're public company CEOs. So to, like, like, th- they are clearly wired in, tapped in. They, they can feel the technology, and they are not gonna let their company lose. Um, you know, assuming that, that as a, as a category, they're in a spot where, where there's a lot of upside. Um, so yeah, but, like, every, every technology wave, there's winners and losers. I don't know that the... This won't be any different. Um, uh, I and I, I just think that you just have to, you just have to be super dialed in and work through it.
- HSHarry Stebbings
A hard one before we do a quick fire.
- ALAaron Levie
[laughs] Okay.
- HSHarry Stebbings
Who has the world turned their back on who you think should be much more appreciated?
- ALAaron Levie
I, I don't... I, like, uh, you know, I'll give maybe a shout-out to, like, um, Atlassian, uh, as an example. Um, uh, I, I think, um, I think that, that feels like-Like oversold territory. You know-
- HSHarry Stebbings
You think 78 cents a bit harsh?
- ALAaron Levie
I, I think, I think, [laughs] I think, I think possibly. And, uh, and, and, you know, it's in the category of I just... But like, like what they've, they've been fighting this, this narrative, and I'm not gonna speak too much for them, but like I think what I perceive is, is, "Oh, no, engineering gets commoditized." And so like where in the stack was, was their engineering, you know, revenue generation? And again, with my headset, I'm like, "No, there's gonna be more engineers." And so now does that mean that Atlassian's product set will have to-- will look exactly like it does today? No. Like, obviously it's got to evolve and whatnot, but, but I think if you're like a company selling infrastructure for engineering to be more automated, uh, like that seems like a good spot to be in. And, and, you know, I-- you look at what Linear is doing, and it's, it's fantastic and it's awesome to watch. Um, but I don't-- I think there'll be, you know, multiple plays in that space just given how big the market is. Um, I think right now this is a moment where you need to be deep in the workflow, and you need to have data. Uh, you, you, you have to, you have to have data, uh, in your platform, and you have to be the best place for that data to go, and you have to be the best place where agents want to work with that data. That's like the mandate right now is if you are not the best place that for-- that an agent would, would sort of intentionally choose for working with data of that particular category or automating the workflow in that particular, you know, area, that's a tough spot to be in, and that, that's the, the job for all of us, you know, if you're building software.
- HSHarry Stebbings
And the best place where agents want to work is defined by great API?
- ALAaron Levie
Great APIs, great pricing models, um, uh, y-y-you like the, the surrounding features to the API. So if you were to say, "Hey, I wanna be able to wire up a workflow where I'm..." This is the, you know, the Box sales pitch. "I wanna be able to wire up a workflow where an agent is interacting with FINRA-compliant documents where you-- when..." You know, FINRA-compliant document means the things that gets generated or seen or shared with a customer, and it, it can't ever be deleted and, and removed af-- you know, for a certain amount of time. Then, then on one hand, the APIs have to be super clean for the agent. On the other hand, you have to have a bunch of surrounding capabilities to ensure that that company can go to, uh, you know, uh, go to their regulator or auditor and say, "Yeah, we, we are complying with FINRA." So that combination is what makes it so you would build that kind of agent on something like Box, and that, that persists across a variety of industries.
- HSHarry Stebbings
I'm gonna do a quick fire round with you.
- ALAaron Levie
Okay.
- 46:30 – 54:43
Frontier Labs: Why Aaron is Still Betting Everything on the Labs
- HSHarry Stebbings
What acquisition did you not make that you wish you had made over the Box journey? Jensen said in the show, "Oh, I wish we'd invest in frontier models. That was my big mistake." What acquisition would you not make that you wish you had done?
- ALAaron Levie
I honestly don't-- I, I don't think I have any, uh, M&A regrets. I actually-- The, the, the, the... It's the deals that I wanted to do that, that we ended up not doing that I don't regret, um, is probably more the, the situation, so-
- HSHarry Stebbings
Wait. Which one is-
- ALAaron Levie
Where, where-
- HSHarry Stebbings
Which one is that?
- ALAaron Levie
I'm not gonna tell you those, but, but-
- HSHarry Stebbings
[laughs]
- ALAaron Levie
There, there are some where left to my own devices I would've done, and I look back and I'm like, "Oh, thank God that there was, uh, there was more rational logic in the process."
- HSHarry Stebbings
Who is going to win the enterprise race, OpenAI or Anthropic?
- ALAaron Levie
Oh, God, that, that's impossible. Back to the cloud piece, I think, um, you know, I think it's, it's totally fair to think about it as a race, and certainly if you're, if you're in either of those companies, you have to treat it like a race because, because, you know, like you obviously want 80% market share, not 55% market share. So like you have to treat this as a, you know, we gotta dominate. Uh, that's exactly how-- They, they should be executing that way. Everything is going according to plan. If you compare it to other areas of compute, um, a-and I, I ran this analysis recently. In 2010, 2010, not like-- You know, maybe you were 12, but like the rest of us, we were just like in companies doing things. Uh, in 2010, AWS made $500 million in revenue. Azure had just launched, and GP-- a-and GCP was called Google App Engine, and it had a little like a turbine logo, uh, with like wings or something. So that was the state of cloud. Fast-forward to this year, and it's a couple hundred billion dollar a year revenue ecosystem, right? So, so in 15 years, right? So and we were in that moment being like, "Who's gonna win, AWS or Azure or GCP? What's-- How is this all gonna play out?" And it just turns out the market was so large, like obviously it was due to their execution that they kept it going and kept it large and the competition kept up, but it, it just didn't really matter. Like everybody, everybody kinda won. And so I, I sort of think of AI in a similar fashion, which is I, I can't predict if it's gonna be OpenAI 60% and Anthropic 40%, or it gets flipped, or I'm off by another 10% here or there. Uh, but, but no matter what, these markets are just fantastically large. Companies are, are going to adopt multiple of these systems. They don't wanna be single vendor. Uh, they don't wanna be-- They don't wanna have single vendor in this stack. One service goes down, or one changes its APIs, or one has a new commercial model. You're gonna-- It-it's gonna be a multi-vendor, multi-AI world. And so, uh, and so tha-that's why it's, like, very hard to kinda, like, call it at this stage.
- HSHarry Stebbings
What does everyone think they know about enterprise adoption with AI that they get totally wrong?
- ALAaron Levie
Uh, what they think is that, is that, uh, that the, the outcomes that you're seeing in AI coding will quickly come for other areas of knowledge work and, um, and that is a, a, a slight misread on the other areas of knowledge work. And, uh, and, and some of it is the idiosyncrasies of, of coding, and some of it is the broad, you know, kind of just elements of the rest of work and how it happens.
- HSHarry Stebbings
If you were a venture investor today, which category would you be most excited to invest in? Obviously, I'm just hyp-
- ALAaron Levie
Yeah
- HSHarry Stebbings
... hypothetically speaking.
- ALAaron Levie
Yeah, yeah.
- HSHarry Stebbings
If that was-
- ALAaron Levie
God, I would-- I, I think I would still be probably loading up on all of, uh, all of the, the hyp- uh, the, the frontier rounds. Uh, it's... Yeah, like these numbers could, could continue to get much larger. And, um, and then I think that the-
- HSHarry Stebbings
Could they get, I mean, much larger? I mean, like this is where am I... At 850 billion, you've got like a, a 3X to like a 2.1 trillion style-
- ALAaron Levie
Yeah
- HSHarry Stebbings
... with dilutional multiplier.
- ALAaron Levie
I, you know, I, I always think it's hard because, um, I, I, I kind of have said that on the way up of many companies. [laughs] Like, like, like, you know, just like-
- HSHarry Stebbings
I did it with crypto.
- ALAaron Levie
Yeah.
- HSHarry Stebbings
Like how much further can it go? It's like... [laughs]
- ALAaron Levie
Yeah. Well, well, that one I'm gonna put in a different category 'cause that's, that can just sort of be memed to life. Um, I- I'm just talking-
- HSHarry Stebbings
Well, no, but I actually, no, but I actually think, but I, I think to the point on Atlassian and, Bluntley, you and your whole category is the casinoization of the stock markets, which is like if you're a momentum trader today, you still buy Palantir because the market's a casino right now.
- ALAaron Levie
Yeah. Well, to, to be a little bit more fair, um, I think you have some one-off companies that have done a, a, um, you know, amazing job capturing the zeitgeist on that. I do think you have-- I, I think, I think the broad story right now is the sector rotation story, which is, which is sort of, hey, this AI thing's happening. Right now I can get a higher return if I, if I get closer to the semi stack and the, and, and the kinda, you know, where the workloads are going and where the data center build-out is happening. And I get less of a return if, if I'm in software, you know, with, with kind of pure licensing. And so I think that, that is probably more of the color of, of what we're seeing. Now, some of the, some of the data center and infra, you know, uh, names maybe have been memeified also, and so that's kinda helping the case. Um, but it's just a we- it's just a really weird time, you know, overall, uh, that, that's, you know, hard to, hard to think through.
- HSHarry Stebbings
So you would not buy Oobirds as an AI company?
- ALAaron Levie
I mean, maybe you would because of that exact point, so I think you'll have-- that will be in the kinda one-off category, so New Bird AI or whatever it, it's called. Um, but, but I think, um... No, I, I am still... I, I hear-- This is a generic statement. There will still be a lot of money to be made in the companies that can take the innovation that we're seeing in Silicon Valley and in the labs and apply it to the real-world, you know, work that happens inside of enterprises. And whether that looks like vertical AI, whether that looks like, you know, the new kinds of tooling that, that companies will need, um, there's a, a, you know, a company and a new category emerging on, on, uh, like agent observability and evaluations. I'll give a shout-out to BrainTrust as an example, not an investor, um, where I can just kinda sit back and be like, "Shit, like we thought that, that agent builders were gonna need evals," so that's like a Silicon Valley TAM. And then I'm like, "Oh, actually everybody on the entire planet, if you're putting agents into an enterprise workflow, needs evals because you need to know if all of a sudden your agent just stopped producing, you know, like, like, uh, you know, loan origination documents the right way." And so I, you know, that's a category where, you know, it's probably not gonna be owned by one of the labs. You kinda want it to work across all the labs. It's, it's, it's, you know, it's a very relevant kinda new form of, of, of infrastructure for an agentic enterprise. I think you're gonna see a dozen, two dozen, five dozen of, of things like that that, that start to emerge.
Episode duration: 54:54
Install uListen for AI-powered chat & search across the full episode — Get Full Transcript
Transcript of episode qrxQikecJL0
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome