Skip to content
a16za16z

Aaron Levie on AI's Enterprise Adoption

a16z General Partner Martin Casado sits down with Box cofounder and CEO Aaron Levie to talk about how AI is changing not just software, but the structure and speed of work itself. They unpack how enterprise adoption of AI is different from the consumer wave, why incumbents may be better positioned than people think, and how the role of the individual contributor is already shifting from executor to orchestrator. From vibe coding and agent UX to why startups should still go vertical, this is a candid, strategic conversation about what it actually looks like to build and operate in an AI-native enterprise. Aaron also shares how Box is using AI internally today, and what might happen when agents outnumber employees. Timecodes: 00:00 Introduction to AI in the Enterprise 00:31 Aaron Levy, CEO of Box 01:32 AI in the Enterprise: Challenges and Opportunities 03:07 The Evolution of AI Adoption 04:54 AI's Role in Workflow Automation 05:55 Faster Buy-in Than Cloud: CIO Attitudes Have Changed 08:08 SaaS vs. AI-Native: Who Wins? 10:00 Is AI Just a Consumption Layer? 12:00 Business Models and the COGS of AI 15:00 New AI-First Categories Are Emerging 19:25 Box's Journey and AI Integration 21:39 The Future of Software and AI 27:41 AI in Decision-Making Processes 29:53 The Impact of Memo-Oriented Meetings 31:03 AI in Research and Strategy 32:18 AI's Role in Enterprise Budgets 43:03 The Future of Entry-Level Engineers 48:28 AI's Influence on Small Businesses 55:36 Predictions for the Next 5-10 Years Resources: Find Aaron on X: https://x.com/levie Find Martin on X: https://x.com/martin_casado Stay Updated: Let us know what you think: https://ratethispodcast.com/a16z Find a16z on X: https://twitter.com/a16z Find a16z on LinkedIn: https://www.linkedin.com/company/a16z Subscribe on your favorite podcast app: https://a16z.simplecast.com/ Follow our host: https://x.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures.

Aaron LevieguestMartin Casadohost
Jul 14, 202559mWatch on YouTube ↗

EVERY SPOKEN WORD

  1. 0:00 – 0:31

    Introduction to AI in the Enterprise

    1. AL

      AI is going to take over the enterprise. We know this is going to happen, and we- it needs to happen to us faster than it happens to our competitors, which is a totally different dynamic than we saw with cloud. What is the journey over the next decade? It's about the speed at which humans can change their workflows. How fast can somebody use a computer to do something? To type an email, to write code, to generate a marketing asset. When that's no longer a limiter, how do these jobs begin to change?

    2. MC

      It's so strange to me how many disruptions are happening all at the same time. [upbeat music]

  2. 0:31 – 1:32

    Aaron Levy, CEO of Box

    1. MC

      Aaron, thank you very much for joining us.

    2. AL

      Thank you.

    3. MC

      Ev- everybody here already knows you. However, I still think you should intro yourself-

    4. AL

      Sure

    5. MC

      ... just for, just for completeness.

    6. AL

      Okay. Uh, Aaron Levie, uh, CEO, co-founder of Box. And, uh, at Box, we help enterprises basically take, uh, all of their unstructured data or enterprise content and turn it into valuable information. Um, and AI is, is absolutely this, uh, incredible accelerant for, for that problem.

    7. MC

      I just learned that we're investors in you. [laughs]

    8. AL

      You... Well, many years ago.

    9. MC

      Yeah.

    10. AL

      Many years ago. So, uh, uh, no, no claims post-IPO. Actually, um, uh, Ben Horowitz had this, um, early, uh, kind of blog post on basically, I think it was the, titled The Fat Startup.

    11. MC

      Yeah.

    12. AL

      Um, and-

    13. MC

      Yeah, yeah, yeah. In response to [laughs] -

    14. AL

      The lean startup movement

    15. MC

      ... enterprises, the lean startup theory.

    16. AL

      Yeah.

    17. MC

      That's right.

    18. AL

      And we, uh, let's just say we very much took that to heart.

    19. MC

      Yeah.

    20. AL

      Um, and we, uh, uh, we, we basically, like, deployed every single lesson, which was, like, the name of the game is you get big fast, you scale a- aggressively. And,

  3. 1:32 – 3:07

    AI in the Enterprise: Challenges and Opportunities

    1. AL

      um, and that was, that was a very important kind of period in, in our company's journey.

    2. MC

      So the, the, the notional topic of this is AI in the enterprise. Um, but I think it's good to be kind of nuanced about this 'cause it's less obvious than people think, and you've been talking a lot about AI on X, but also, you know, you're thinking about it in the terms of your business. So let me just kind of set up the first question-

    3. AL

      Sure

    4. MC

      ... as follows, which is what's... You know, AI has historically been this very B2B enterprise thing.

    5. AL

      Yeah.

    6. MC

      Like chatbots or whatever, you know, um, personalization systems. But what's unique about GenAI is a lot of the use cases are actually, like, consumer or prosumer, right? Think, like, creativity or developers, and it actually hasn't made intros as much into the enterprise yet. It's just starting now.

    7. AL

      Yep.

    8. MC

      So maybe just a couple of questions. First off, like, A, does that match with your experience? And then, B, how are you thinking about, you know, this transition to the enterprise?

    9. AL

      Yeah. I think if you were to pro- probably, like, do the, the idiosyncrasies of AI and then reverse engineer why that was the journey, um, basically up until, let's say, pre-ChatGPT moment, AI was extremely hard to use.

    10. MC

      Yeah.

    11. AL

      It required, in many cases, having custom models for basically every problem you tried to solve.

    12. MC

      Yeah.

    13. AL

      And so there was almost no way that a consumer ecosystem could flourish-

    14. MC

      Yeah

    15. AL

      ... you know, based on that. It was not, it was just, it was not generalizable enough. There was really few products other than, like, maybe Siri, Alexa, et cetera-

    16. MC

      Yeah

    17. AL

      ... that you'd interact with that would even have some sense of AI. Um, and so enterprises were the, the, you know, early adopters of AI systems to bring automation and, and workflows, uh,

  4. 3:07 – 4:54

    The Evolution of AI Adoption

    1. AL

      to, to, uh, work- workflow automation to their, um, companies. Then boom, ChatGPT happens, and all of a sudden it's the exact right form factor-

    2. MC

      Yeah

    3. AL

      ... for mass adoption.

    4. MC

      Yeah.

    5. AL

      There's no startup costs. It costs, uh, you know, two minutes to, uh, two seconds to learn the product. You just... It's a chat interface.

    6. MC

      Yeah.

    7. AL

      So it was, like, perfectly ripe for just taking off in the consumer space.

    8. MC

      Yeah.

    9. AL

      Um, and, and then, you know, you have, uh, also these incredible conditions, uh, set up for mass adoption. You have billions of people on the internet.

    10. MC

      Yeah.

    11. AL

      It was set up as a free product.

    12. MC

      Yep.

    13. AL

      Um, uh, again, it, it kind of solved this sort of latent kind of question mark that everybody had, which was like, when are we gonna see AI, you know, touch-

    14. MC

      Work [laughs]

    15. AL

      ... touch work and touch our lives?

    16. MC

      Yeah, yeah, yeah.

    17. AL

      And so all, everything was kind of, like, the perfect conditions to, to get mass consumer adoption. On the enterprise side, um, you, you, you have, uh, unfortunately kind of the opposite, right? You have, you have-

    18. MC

      Lots of workflows

    19. AL

      ... lots of workflows that have been kind of ingrained for decades and decades. You have lots of legacy IT systems that have data-

    20. MC

      Yeah

    21. AL

      ... kind of not set up well to be accessed by AI.

    22. MC

      Yeah.

    23. AL

      Um, you have a sort of shadow IT problem-

    24. MC

      Yeah

    25. AL

      ... which is, which is most corporations don't want, you know, end users just injecting, you know, text into prompts that might contain information that the AI models could learn off of. So it's sort of a, a, a difficult environment for that same level of virality.

    26. MC

      Yeah.

    27. AL

      With the exception of a few of these prosumer categories, um, I'm... you know, I have talked to, uh, large corporation, um, you know, CIOs that are seeing people just show up with Windsurf and Cursor and Replit.

    28. MC

      Yeah.

    29. AL

      And so you're getting actually this sort of shadow IT version that we saw 15 years ago.

    30. MC

      But dev, dev tools has always been thought of that way.

  5. 4:54 – 5:55

    AI's Role in Workflow Automation

    1. MC

      Yeah, yeah.

    2. AL

      Even, even separate from the people that pay for it.

    3. MC

      Totally.

    4. AL

      Um, so now the question though is, is, like, what is the journey over the next decade-

    5. MC

      Yeah

    6. AL

      ... for the real change management of deployment of AI systems-

    7. MC

      Yeah

    8. AL

      ... that drive the more, like, GDP-changing-

    9. MC

      Yeah

    10. AL

      ... productivity gains?

    11. MC

      Yeah.

    12. AL

      And that, that's something where I do think we have to be prepared for. This is many years. It's about the speed at which humans can change their workflows as opposed to how kind of quickly the technology can just, you know, sort of evolve and advance.

    13. MC

      Yeah.

    14. AL

      And so we in Silicon Valley, and certainly anybody tuning into this, sort of imagines like, well, why doesn't the breakthrough that we just saw get released, why isn't that, you know, permeate every corporation-

    15. MC

      Yeah, yeah, yeah

    16. AL

      ... within six months?

    17. MC

      Yeah.

    18. AL

      And it's because, like, people just, like, have meetings and they have budget, you know, pr- processes, and they have to go through a governance council, and they have to get compliance on board.

    19. MC

      Yeah. The operational models.

    20. AL

      And they have to figure out, like, who has the liability when the-

    21. MC

      Right

    22. AL

      ... when the thing recommends this stock, and then they, you know, uh, the financial services provider shares that with a client. Like-

    23. MC

      Yeah

    24. AL

      ... that takes years.

    25. MC

      Yeah.

    26. AL

      And there's gonna be case law that needs to happen.

    27. MC

      Yeah, yeah, yeah.

    28. AL

      And we still have lawsuits that are going on about-

    29. MC

      Yeah

    30. AL

      ... who owns the IP of this stuff. So, so that, that

  6. 5:55 – 8:08

    Faster Buy-in Than Cloud: CIO Attitudes Have Changed

    1. AL

      part is gonna take years. What, what's interesting, and I think you'll, you'll especially appreciate this on the cloud side, is, um, I remember when we-First we're scaling up in the enterprise, let's say 2007, 2008, 2009. You know, let's say that three to five-year period.

    2. MC

      Mm.

    3. AL

      Post-AWS, post kind of cloud starting the, its journey, basically to a T, every conversation you'd have with a CIO or a group of CIOs was basically like, "Yeah, that's nice. Maybe some little corner of our organization could use this."

    4. MC

      Yeah.

    5. AL

      "We are never gonna go fully to the cloud."

    6. MC

      Yeah.

    7. AL

      You know, they, they had their arms, you know-

    8. MC

      I remember

    9. AL

      ... wrapped around their servers.

    10. MC

      I remember, yeah.

    11. AL

      Yeah, and, and, and, and basically they did not wanna give up the infrastructure. They... There was too many questions, too many compliance issues. There was, you know, l- just job, you know, existential job questions-

    12. MC

      Yeah

    13. AL

      ... of like, "Well, what happens when-

    14. MC

      Sure

    15. AL

      ... this, you know, gets delivered as a service?" Here's what's super interesting. Let's say we're now two years into, two and a half years into the ChatGPT moment.

    16. MC

      Yeah.

    17. AL

      That same s- group of CIO conversations, none of that. It is basically assumed-

    18. MC

      Wow

    19. AL

      ... it is, it is basically fully assumed that AI is going to take over the enterprise.

    20. MC

      Whoa.

    21. AL

      Um, like the CEO, the CIO, the CDO, every jo- every, you know, org leader is basically like, "We know this is going to happen."

    22. MC

      Yeah.

    23. AL

      This is not, uh, this is not like a, a we're trying to kinda push it off. It is purely a sequence of events. Who do I deploy? How do I deploy it?

    24. MC

      Yeah.

    25. AL

      How do I drive the change management? Is the model ready? So what's really interesting is I think the level of buy-in you have now in the enterprise is, like, five times be- you know, greater than we had in the early days of cloud. And you can even see it. Like, the, to me, the classic witness, uh, test was, um, if you remember like 15 years ago, I think Jamie Dimon was probably most famous for saying like, "We're never gonna go to the cloud."

    26. MC

      Yes. [laughs]

    27. AL

      So, like, they basically said JPMorgan will never go to the cloud.

    28. MC

      Yeah.

    29. AL

      You know, today, that equivalent commentary, whether... I don't have a perfect Jamie, uh, a Jamie Dimon quote, but David Solomon at Goldman Sachs has given this anecdote of they can write now an SEC filing or an S1 for an IPO-

    30. MC

      Yeah, yeah

  7. 8:08 – 10:00

    SaaS vs. AI-Native: Who Wins?

    1. AL

      cloud.

    2. MC

      So do you think this has implications for companies today that are building products that are, you know, pre-AP- or pre-AI products? So for example, with the cloud wave, you know, you basically had a bunch of cloud native companies that ended up, you know, taking over, right?

    3. AL

      Yeah.

    4. MC

      Like, so for example, Snowflake is a great example of this, which is like the ones that, you know, decided not to go all in and were hybrid. Like, hybrid i- it kind of became known as like means it won't work. [laughs]

    5. AL

      Right.

    6. MC

      Like, you know, anything called hybrid, like, hasn't worked.

    7. AL

      [laughs] Yeah.

    8. MC

      And they get to do the-

    9. AL

      Yeah.

    10. MC

      So, so do you think because the buyer and the enterprise is more ready that, like, companies that are pre-AI have more of an opportunity, or do you think that, you know, you're gonna see the same thing with a lot of, like, AI native companies do well?

    11. AL

      I, I'm gonna basically give you the non-answer of, I think both.

    12. MC

      Yeah.

    13. AL

      Um, and one benefit that, that the cloud cohort has or the SaaS, you know, kind of post like us all understanding and agreeing on what SaaS would look like, what, what we all have is we were, we... Whether, whether we li- you know, adhered to this perfectly or not, you know, is a question, but we basically all tried to build API-first platforms.

    14. MC

      Yeah.

    15. AL

      And so, um, or at least like API kind of like equal platforms. So we have the UI, and we have the API.

    16. MC

      Mm-hmm.

    17. AL

      And if you think about it, like age- AI and AI agents are like the perfect, you know, consumers-

    18. MC

      Yeah

    19. AL

      ... of an API.

    20. MC

      Yeah.

    21. AL

      Right?

    22. MC

      Yeah.

    23. AL

      And, and so they basically become these super users within your system-

    24. MC

      Yeah

    25. AL

      ... on your APIs.

    26. MC

      Yeah.

    27. AL

      So, so if I had to just say, "Okay, I want to au- I wanna deploy agents to go and automate my ServiceNow workflows-"

    28. MC

      Yeah

    29. AL

      ... I think I'm, I'm better off just deploying the ServiceNow agent to go do that-

    30. MC

      Yeah

  8. 10:00 – 12:00

    Is AI Just a Consumption Layer?

    1. AL

      pre-cloud to post-cloud was a, an entire rewriting of your software.

    2. MC

      Yeah.

    3. AL

      You had to go from single tenant to multi-tenant.

    4. MC

      Yep.

    5. AL

      The scaling of the systems were totally different. Even the functionality and, and application logic was different because, like, well, it should be real time. It should be collaborative. It shouldn't be as, as sort of async and batches as-

    6. MC

      Yeah

    7. AL

      ... the on-prem systems were. And so in a cloud world, it's, it is a reinvention of the user experience-

    8. MC

      Yeah

    9. AL

      ... and what, what you're doing in the system, and we should, you know, definitely get to that.

    10. MC

      Well, well, I-

    11. AL

      Yeah

    12. MC

      ... I just, I wanna make sure I tease this out 'cause it's actually, uh, it's a very interesting point, which is so your claim is, is to go from pre-cloud to post-cloud, like that ripped through the entire stack all the way down to like the infrastructure, for example, like tenancy.

    13. AL

      Yes.

    14. MC

      Like you have to rewrite everything, and then what you're saying about AI is more of a consumption layer thing, which is like you just kind of like treat the existing systems as they are-

    15. AL

      Yeah

    16. MC

      ... and then, like, you know, the AI becomes the consumption layer.

    17. AL

      Yep.

    18. MC

      Do you think this is like a 1.5 step, and like the 2.0 step-

    19. AL

      Yeah

    20. MC

      ... kind of rips through the entire stack, or-

    21. AL

      Well, okay, so, so let's, let's bookmark that one-

    22. MC

      Yeah

    23. AL

      ... for one second.

    24. MC

      Yeah.

    25. AL

      So, so but like if you, if you do pure Clay Christensen sort of, you know-

    26. MC

      Yeah

    27. AL

      ... approach, you, you know, sustaining innovation, disruptive innovation.

    28. MC

      Yep.

    29. AL

      Disruptive innovation is this thing that looks like so much harder, so different, so less profitable.

    30. MC

      Yeah.

  9. 12:00 – 15:00

    Business Models and the COGS of AI

    1. MC

      Yeah.

    2. AL

      So I think you have a lot of TAM expansion. Now, the good news for startups-

    3. MC

      With one, with one caveat-

    4. AL

      Yeah. Right

    5. MC

      ... which may- maybe we've bookmarked, and we're gonna get to, but let me just say the one caveat.

    6. AL

      Yeah.

    7. MC

      The one caveat is you now have a component that has a very different COGS model if you're a software provider.

    8. AL

      Yes.

    9. MC

      And so like now it's almost like, it's almost like when we went from like on-prem to cloud-

    10. AL

      Yeah

    11. MC

      ... we went from perpetual to recurring.

    12. AL

      Yeah.

    13. MC

      And it feels like with AI, you kind of have to go from recurring to usage-based just because

    14. AL

      Yeah. I, I, um... Okay. So, so business model will, will shift for some of the use cases.

    15. MC

      Yeah.

    16. AL

      Because even if you look at the Cursors, Replitz, you know, Windsurfs of the world, there does seem to be this baseline seat price, and then, and then your consumption usage thing is sort of this add-on.

    17. MC

      This overage. Yeah.

    18. AL

      And so, and so, you know, SaaS providers are kind of well, well-structured to be able to have that kind of dynamic.

    19. MC

      Yeah.

    20. AL

      Um, if it was 100% usage and the, and the user seat goes away, I do agree then you have this... Then, then you have a, then you have a little bit of a business model crisis, which-

    21. MC

      Oh, so you think but right now it's not clear that that's gonna go all the way over?

    22. AL

      Well, until you, until the human literally is not a seat on the system, I d- I think you don't remove the end user license as a component.

    23. MC

      Okay.

    24. AL

      And, but again-

    25. MC

      Right

    26. AL

      ... that, that could be, like, the much bigger disruption.

    27. MC

      Yeah.

    28. AL

      Now, now, the, the, um... Just, just to, you know, f- fully lay out though the, the, the market dynamics, I think SaaS incumbents are es- especially, you, you have a couple other idiosyncrasies right now versus the on-prem days. Another idiosyncrasy is I would say, like, on the margin, you tend to have founders still leading the, the SaaS companies.

    29. MC

      100%. Yeah.

    30. AL

      And so-

  10. 15:00 – 19:25

    New AI-First Categories Are Emerging

    1. MC

      like, professional coding.

    2. AL

      Yeah.

    3. MC

      But everything above that is one of these. So on the consumer that's very clear.

    4. AL

      Yep.

    5. MC

      Is that, is that clear on the enterprise side?

    6. AL

      I, I, I absolutely think so. I think if you looked at... Just, just take, um... If we did a snapshot 10 years ago-

    7. MC

      Yeah

    8. AL

      ... of the size of the, uh, contract management market or the legal document market-

    9. MC

      Okay

    10. AL

      ... it's, like, sub two billion.

    11. MC

      Yeah.

    12. AL

      I'm making up the numbers.

    13. MC

      Yeah, yeah. Sure, sure.

    14. AL

      It could be plus or minus a billion.

    15. MC

      Yep, yep, yep.

    16. AL

      Would you agree that in five years from now, the AI agent related spend on legal services should be in the many, many b- billions to double digit billions?

    17. MC

      Absolutely.

    18. AL

      Okay.

    19. MC

      No question.

    20. AL

      So all of a sudden there's, like, not these natural incumbents-

    21. MC

      Yeah, yeah

    22. AL

      ... that were like, "Oh, we, we captured all that market."

    23. MC

      I see. Yeah.

    24. AL

      AI agents all of a sudden expands the size-

    25. MC

      Yeah

    26. AL

      ... of the software related spend in that space. So I can underwrite that for healthcare, legal, consulting services. I think there's e- entire areas of financial services. Like, we always think, "Oh, finance has been wired up for so many years." No. Banking, you know, like, con- you know, consumer banking has been wired up. Trading has been wired up. Investment banking never went digital.

    27. MC

      Yeah.

    28. AL

      Wealth management never went digital.

    29. MC

      Yeah.

    30. AL

      Like, these were not categories where, where you ever had, like, major software platforms to help these entire categories of the economy. And, and the reason was because the work was unstructured. It's very ad hoc, very dynamic, lots of unstructured data as opposed to stuff that goes into databases. All of that is now ripe for AI, and that will then largely be ripe for many startups because there won't be a natural incumbent in those spaces.

  11. 19:25 – 21:39

    Box's Journey and AI Integration

    1. AL

      those type of things first.

    2. MC

      Just, I mean, I can't imagine a, a listener n-not knowing what Box does, but just for completeness, maybe can you just talk to us very quickly about what Doc- Box does-

    3. AL

      Yeah

    4. MC

      ... and how you're thinking about how that dovetails with AI?

    5. AL

      Yeah, so, um, we started the company on a really s- with a really simple premise, make it easy to access and share your files from anywhere. And we, we pivoted about two years in to, to the journey to focus on the enterprise market, and the whole idea was enterprises are awash with all this in, you know, unstructured data, so corporate documents, research files, marketing assets, uh, M&A documents, contracts, invoices, all of this. And as companies move to the cloud and as they move to mobile, they need a way to access that information. They need a way to collaborate securely on it.

    6. MC

      Yeah.

    7. AL

      Um, they wanna be able to integrate that data across different systems, so we built a platform to help, to help companies do that. We have about 120,000 customers, about 65 or so percent of the Fortune 500. And so what's incredible right now is we've had this ongoing problem since the creation of the company, which is with structured data, the stuff that goes into your database-

    8. MC

      Yep, yep

    9. AL

      ... you can query it, you can synthesize it, you can calculate it-

    10. MC

      Yep

    11. AL

      ... you can analyze it.

    12. MC

      Yep.

    13. AL

      Your unstructured data, the stuff that we manage-

    14. MC

      Yep

    15. AL

      ... you create it, you share it, you look at it, and then you basically kinda gets forgotten about. Like, it goes into some folder, and you almost never see it again. And, and maybe you kinda find it once every five years for some task you're doing, but that's about it. And so most companies are sitting on most of their data being unstructured-

    16. MC

      Mm-hmm

    17. AL

      ... and getting the least amount of value from it relative to their other, you know, structured data.

    18. MC

      Yep.

    19. AL

      AI is basically the unlock. So AI lets you finally say, "Okay, we can ask this data questions."

    20. MC

      Oh, cool. That's cool.

    21. AL

      "We can structure it, so we can take, we can look at a contract, pull out the 10 most important fields. Once we have all that data, we can analyze that information, we can get in-insights from it." And then you can start to do things like workflow automation that was never possible with your unstructured data. So if I wanna move a contract through an automatic process, I can't do it if I don't know what's in the contract, and the computer previously was not able to know-

    22. MC

      Yeah

    23. AL

      ... what's in the contract. So for us, this is a huge unlock of now what you can finally do with your information and your content, so we're building an AI platform to handle all of the kinda plumbing, user experience, uh, to make then your content AI-ready effectively.

  12. 21:39 – 27:41

    The Future of Software and AI

    1. MC

      I don't wanna be, like, too bullshitty and-

    2. AL

      Sure

    3. MC

      ... provocative, but I have to, I have to ask this, which is-

    4. AL

      Please

    5. MC

      ... um, I've been in enterprise software for a very long time.

    6. AL

      Yes.

    7. MC

      A lot of the business model is predicated on the fact that, like, building software is hard and takes a long time.

    8. AL

      Yep.

    9. MC

      Like, to what extent do you worry about that not being the truth going forward? Like, do you think we enter, like, this time of bespoke software being upon us?

    10. AL

      Um, I'm, I'm, I'm bearish on the, on the, uh, extreme version of that, um, uh, of the essence of that. So the extreme version of that, if you just im- you know, if you imagine the poles of this, like, the extreme... Like, like, you know, on one pole, basically all software is prepackaged. It, you know, it's the Ford Model T. It's gonna work v- the only in one way.

    11. MC

      Yep, yep.

    12. AL

      Everybody uses the same thing.

    13. MC

      Yeah.

    14. AL

      Okay, like, that's not gonna happen. We, we get that.

    15. MC

      No. Yeah.

    16. AL

      The other extreme is, like, everything is just, like, home brew.

    17. MC

      You wake up in the morning, you utter something-

    18. AL

      Yeah

    19. MC

      ... you get your software for the day. [laughs]

    20. AL

      You get your software for that thing, and then, like, the next day you do it again and you change it.

    21. MC

      Yeah, yeah, yeah.

    22. AL

      Okay.

    23. MC

      Yeah.

    24. AL

      The, the downsides of that model, of why basically I think it doesn't work, is I think if you ask, like, 90%... Uh, if you ask, uh, the world population, uh, you know, you'd probably find that 90-plus percent just don't care enough. They just don't, like, they don't care about-

    25. MC

      [laughs] That's true

    26. AL

      ... about the tabs on their software-

    27. MC

      Right, right, right, right, right

    28. AL

      ... and the modules on their dashboard.

    29. MC

      Right, right, right.

    30. AL

      They d- Like, it's like they want someone else to just be like, "This is what you should look at in the morning."

  13. 27:41 – 29:53

    AI in Decision-Making Processes

    1. AL

      that basically, like, these things are gonna live together.

    2. MC

      Cool. Let's move from software to decision process. So, uh, I won't say the name of the company, but I, I just spoke with a very, very legit company, household name-

    3. AL

      [laughs]

    4. MC

      ... where, uh, it's a private company though, it's not a public company, where at the board level, for every decision they ask the AI [laughs] for like, like, a basically a more information for the decision.

    5. AL

      Okay.

    6. MC

      Right? And they were like, "This has actually been great-

    7. AL

      Yeah

    8. MC

      ... from like-

    9. AL

      That's funny

    10. MC

      ... from like discussion fodder-

    11. AL

      Yeah

    12. MC

      ... to be provocative."

    13. AL

      Yeah.

    14. MC

      And it also shows how, like, fundamentally unoriginal [laughs] the board members are.

    15. AL

      Yeah.

    16. MC

      Like, this founder was telling me, he's like, "It's literally better than half of my board members," right? And so [laughs] like, how much have you thought about bringing AIs in to like help with decision process?

    17. AL

      Yeah.

    18. MC

      And by the way, I think the board is like low-hanging fruit, because boards tend to not have a lot of context to the business, and so kind of like the instance are probably less anyways. But is this something that you've thought about, or...?

    19. AL

      Um, well, no, uh, the, the board one is an interesting one, so maybe we can, we can, uh, unravel that one. But the, um... Well, like, I already use it for, let's say, our earnings calls, where we'll do a draft of the initial, uh, earnings script.

    20. MC

      Okay.

    21. AL

      And then, I mean, again, because Box AI deals with unstructured data-

    22. MC

      Yeah

    23. AL

      ... I just load up the earnings script, and I'll use a, a better model and say, like, "Give me 10 talk... Give me 10 points that analysts are gonna ask about this, and like, how would I improve the script?" And it just spits out a bunch of things, and it's-

    24. MC

      And how, how good is it at predicting what analysts are gonna ask?

    25. AL

      Oh, extremely good. Oh, 100%.

    26. MC

      We're so-

    27. AL

      Because, because, but the, but the thing is like that's not surprising. Like-

    28. MC

      Yeah

    29. AL

      ... like, it has access-

    30. MC

      No, of course

  14. 29:53 – 31:03

    The Impact of Memo-Oriented Meetings

    1. AL

      whatever.

    2. MC

      Yeah, yeah, yeah.

    3. AL

      So it's a, it's a quick way to just do, do some analysis on something. The, um, but yeah, I mean, you know, it's, it's funny, we, we, uh, uh, so Bezos, you know, famously had this me- memo-oriented-

    4. MC

      Yeah

    5. AL

      ... essay-oriented-

    6. MC

      Yeah

    7. AL

      ... kind of meeting structure, and, um, w- we never did that, but I was always fascinated by the companies that, that could do it. And actually we're, we're entering a world where probably you could just pull that off, right? So if, imagine if, whether it's a board meeting or product meeting, you just do a quick deep research essay on the topic. Like, obviously every meeting, every strategy meeting in, uh, in history would be better off if you probably had that as a starting asset-

    8. MC

      Yeah

    9. AL

      ... to get everybody informed.

    10. MC

      No, the, the, um, the, I think the argument against that would be the reason Bezos said it is because it forced people to think clearly about what they're doing-

    11. AL

      Yeah

    12. MC

      ... and writing it down. So the exercise meant the people walking in the meeting had more context.

    13. AL

      Yeah.

    14. MC

      This would almost argue that they would have less context because something else did the thinking.

    15. AL

      Well, well, two things. The, the, the, it was, uh, it was to make sure that the person doing the thing had the clarity to write it.

    16. MC

      Yeah. Yeah. Yeah.

    17. AL

      For sure.

    18. MC

      Exactly.

    19. AL

      But it was also still to inform everybody else that could, that didn't do that work.

    20. MC

      Right. True.

    21. AL

      And so it certainly would've helped everybody else in the room.

    22. MC

      Yeah.

    23. AL

      And I'm not 100%. I mean, we should do a full longitudinal analysis of like the people that wrote the essay, did they actually have the better products?

    24. MC

      They were-

    25. AL

      Or like, like, I mean, there's

  15. 31:03 – 32:18

    AI in Research and Strategy

    1. AL

      some Amazon products I don't like, and so d- they obviously wrote an essay also for those.

    2. MC

      Yeah.

    3. AL

      So I, you know, I don't know the hit rate ultimately on the essay specifically-

    4. MC

      Yeah

    5. AL

      ... as much as the idea of like write down a strategy, think it through, and so why not have an agent do like 90% of the heavy lifting?

    6. MC

      Yeah.

    7. AL

      So a lot of my, my workflows are, like, if I have a topic where, like, like, maybe the direct change of, of my workflow on this front is, is the kind of thing that three years ago I might sort of lob over to the chief of staff and say, "Hey, can you, like, go rese- research like the pricing strategy-

    8. MC

      Yeah

    9. AL

      ... of this ecosystem or something?"

    10. MC

      Yeah.

    11. AL

      That's just a deep research query now.

    12. MC

      Oh, yeah.

    13. AL

      And then I'll wake up and-

    14. MC

      Sure. Yeah

    15. AL

      ... and it'll look at this thing.

    16. MC

      For sure.

    17. AL

      And so the, but what that does is because now I'm not having to calculate that person's time, their tasks, their trade-offs.

    18. MC

      Yeah.

    19. AL

      I just do it for the most random things.

    20. MC

      Yeah.

    21. AL

      Which means, like, I'm expanding and exploring way more spaces mentally than I would've before.

    22. MC

      Yeah. Yeah.

    23. AL

      And these are the kind of, you know, parts where like, like... And, and, and again, this is equally why I'm like actually more optimistic on the jobs front-

    24. MC

      Yeah

    25. AL

      ... because what we, we, we do too many times within AI is we, like, look at today's way of working, and we're just like, "AI will come in and take 30% of that." And it's like, no, no, no. A-

    26. MC

      Yeah

    27. AL

      ... we'll just do totally different things with AI.

    28. MC

      Yeah.

    29. AL

      I wouldn't have researched that thing before when it was people required to research it-

    30. MC

      Yeah

  16. 32:18 – 43:03

    AI's Role in Enterprise Budgets

    1. AL

      an inane, you know, task to send to somebody.

    2. MC

      Yeah, yeah, yeah. So o- one thing, so when we run the numbersAnd by run the numbers, I mean look through how AI companies are doing, where does the value accrue. There's basically one takeaway.

    3. AL

      [laughs]

    4. MC

      And that is, like these markets are very large-

    5. AL

      Yeah

    6. MC

      ... and growing very fast. And value is kind of accruing at every layer.

    7. AL

      Yeah.

    8. MC

      Everything from like literally chips up to apps, right?

    9. AL

      Yeah.

    10. MC

      Everybody, you know. And so, like the only real sin is zero-sum thinking.

    11. AL

      Yeah.

    12. MC

      To be like, "Oh, like the models are not gonna be defensible," or, or whatever the, your zero-sum thinking is, that just hasn't proven out.

    13. AL

      Yep.

    14. MC

      Now, this has still largely been a consumer phenomenon, so like what, what I've been thinking about, and I don't, I don't have an answer. I'd love to hear your thought is, is when it comes to enterprise budgets, like you can't just create budget out of thin air.

    15. AL

      Mm.

    16. MC

      So, like you actually do have a limited resource. And so as budgets get reallocated, to what extent do you think this is like zero sum, like the old budget gets robbed versus like budget accretive? Or like how do you think-

    17. AL

      Yeah

    18. MC

      ... about that? Because again, like where we've come from, that has not been an issue.

    19. AL

      Yeah.

    20. MC

      I think in the enterprise it probably will be.

    21. AL

      So it does have to come from somewhere. It's fu- fully, fully logical. Um, couple things.

    22. MC

      Yeah.

    23. AL

      A large number for startups can also be a very small number for a large corporation.

    24. MC

      Yeah, of course.

    25. AL

      Uh, so, so you have that dynamic playing out.

    26. MC

      Sure.

    27. AL

      Um, uh, like I, I'll, I'll make up random stats, but you could probably take a meaningful engineering team and s- and probably for the price of five of those engineers or 10 of those engineers, you could probably pay for Cursor licenses for the entire engineering team.

    28. MC

      Right.

    29. AL

      And so-

    30. MC

      But this would argue that it's actually coming out of headcount.

  17. 43:03 – 48:28

    The Future of Entry-Level Engineers

    1. MC

      get interns and go into management, and so maybe we're just skipping that step.

    2. AL

      Right.

    3. MC

      So the obvious question is, is what happens to entry-level engineers? Like, does this change how people get introduced to computer science, for example?

    4. AL

      The cool thing is probably more people will even now get introduced to computer science.

    5. MC

      Yeah.

    6. AL

      So, um, because you'll be able to, uh, you-

    7. MC

      Anybody can learn

    8. AL

      ... anybody can learn it.

    9. MC

      Yeah, yeah. Sure.

    10. AL

      And, and like, uh, you know, it's been 25 years for me, but like in the early days of like programming basic applications or putting together websites, it was just extremely frustrating that you'd spend-

    11. MC

      [laughs] Oh, my

    12. AL

      ... you'd spend a- you know, days and days being like-

    13. MC

      Oh

    14. AL

      ... "Why does that thing not work?"

    15. MC

      Yeah, yeah, yeah.

    16. AL

      And it was like, and like I had n- I had very few resources of like-

    17. MC

      Yeah

    18. AL

      ... figuring out why the thing didn't work.

    19. MC

      Yeah, yeah.

    20. AL

      It would've been 100 times easier if I could've had an agent write the thing. I would've learned, I would've learned 10 times faster.

    21. MC

      Yeah, yeah.

    22. AL

      Be- and, and, and-

    23. MC

      I mean, honestly, what you did is you were like, well, not, not 25 years ago, but 10 years ago, you'd go to Stack Overflow.

    24. AL

      Yeah.

    25. MC

      And so it was like the slow version of-

    26. AL

      Yeah, but so think about how many people missed the window pre-Stack Overflow-

    27. MC

      Yeah

    28. AL

      ... that got sort of pushed out of the ecosystem-

    29. MC

      Yeah

    30. AL

      ... because they're just like, "This is too frustrating."

  18. 48:28 – 55:36

    AI's Influence on Small Businesses

    1. AL

      s- things that you, you can actually relieve your team to go and work on is incredible.

    2. MC

      Yeah. Yeah.

    3. AL

      And the other big like boon for, for, uh, the economy, and this is again where the economists just totally miss this stuff, is think about every small business on the planet, of which there's-

    4. MC

      Yeah

    5. AL

      ... millions, tens of millions, whatever.

    6. MC

      Yeah.

    7. AL

      That for the first time ever in history, they have access to resources that are somewhat approximate to-

    8. MC

      Yeah

    9. AL

      ... the resources of a large company.

    10. MC

      Yeah. The biggest company.

    11. AL

      Like-

    12. MC

      Ever. 100%

    13. AL

      ... they, they can do any marketing campaign. Uh-

    14. MC

      I know

    15. AL

      ... did you see the NBA Finals video for, from Kalshi?

    16. MC

      No.

    17. AL

      Um, the VO3 video?

    18. MC

      Oh, yeah, yeah, yeah.

    19. AL

      Yeah.

    20. MC

      Yeah, yeah.

    21. AL

      Like-

    22. MC

      Yeah, yeah

    23. AL

      ... like you can now put together-

    24. MC

      Yeah. Right

    25. AL

      ... a, an otherwise million-dollar marketing video-

    26. MC

      Yeah

    27. AL

      ... for a couple hundred bucks of tokens.

    28. MC

      I know.

    29. AL

      And, and that being applied to every domain-

    30. MC

      Right

  19. 55:36 – 58:55

    Predictions for the Next 5-10 Years

    1. MC

      conversation about like the current impacts-

    2. AL

      Yeah

    3. MC

      ... and the near term impacts. If you do a longer view, can you dare to guess what things look like in five to 10 years or-

    4. AL

      I think, um, uh, so Sam, Sam Altman and Jack Altman had a podcast recently.

    5. MC

      Yeah, yeah, yeah.

    6. AL

      And-

    7. MC

      It was very good

    8. AL

      ... I'm gonna, I'm gonna, you know, paraphrase probably, you know, in, in some wrong way, but like they were going back and forth about how like we, we just got, you know, what we would've predicted as AGI five years ago, and it's just like we use it.

    9. MC

      Yeah.

    10. AL

      And it's like-

    11. MC

      Like, yeah [laughs]

    12. AL

      ... and like it's now just built in.

    13. MC

      The most anticlimactic-

    14. AL

      Yeah

    15. MC

      ... or the most climactic anti-climactic note.

    16. AL

      And I think that's my instinct for a lot of this-

    17. MC

      Yeah

    18. AL

      ... is five years, 10 years, whatever your number is, um, we should... And this is why I'm so optimistic-

    19. MC

      Yeah

    20. AL

      ... is like o- on just society and jobs and all this stuff, is like I don't think it's the Terminator-

    21. MC

      Yeah

    22. AL

      ... kind of crazy outcome, you know, scenario of we automate away everything. I think, I think the human capacity for wanting to solve new problems, for creating new products, for serving customers in new ways, um, uh, for delivering better healthcare, to try and do scientific discovery, like all of this stuff just will, w- is just like built in us.

    23. MC

      Yeah.

    24. AL

      And it will continue.

    25. MC

      Yeah.

    26. AL

      And AI is this kind of upleveling of the tools that we use to do all those things. And so I think the, the way we work will be totally different in five years or 10 years.

    27. MC

      Totally.

    28. AL

      But y- you're already seeing enough of probably what it will look like, that I think it's an extrapolation of that. It's when you want the marketing campaign done, you have a set of agents that go and create the assets and ch- choose the markets and figure out the ad plan, and then you have a, a few people review it and debate it, and say, "Okay," like, "let's go in this direction instead." And then you deploy it and you're onto the next thing.

    29. MC

      Yeah.

    30. AL

      And so each company, their, their units of output grow. As a result of that growth, you know, we're all still in competitive spaces so, so some of it gets competed out, and others will keep growing faster than they would've before, so they'll hire more people, and those will be... You'll have new types of jobs, like we'll have jobs for people just to manage agents, and like you'll have operations teams. You know, Adam D'Angelo had this cool role that just kind of got announced of-

Episode duration: 59:07

Install uListen for AI-powered chat & search across the full episode — Get Full Transcript

Transcript of episode EjLkFc1xgR4

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