Dropbox CEO drops in & Softbank's $40B OpenAI bet
49:24 Share

ACAST powers the world's best podcasts. Here's a show that we recommend. They said over and over and over again, and we just ignored it. What if a warning siren sounds? Look for cover.

From Lawfare and Goat Rodeo, this is Escalation, a podcast about the United States and Ukraine. We were all living in a bureaucratic process, and Putin's living in the world of hard power. And the word escalation, everyone just says it. This latest escalation... ...paralyzed because of nuclear escalation. And you know what is interesting about the war? When you don't want to understand the war from the distance, it comes closer and closer.

Escalation is out now, wherever you get your podcasts. ACAST helps creators launch, grow and monetize their podcasts everywhere. ACAST.com This episode of the Times Tech Podcast is sponsored by Vanta. Let's talk about something that might be keeping you up at night. Cybersecurity. According to Vanta's latest State of Trust report, it's the number one concern for UK businesses. And that's where Vanta comes in.

Whether you're a startup, growing fast or already established, Fanta can help you get ISO 27001 certified and more without the headaches.

And Vanta allows your company to centralize security workflows, complete questionnaires up to five times faster and proactively manage vendor risk to help your team not only get compliant, but stay compliant. So stop stressing over cybersecurity and start focusing on growing your business in 2025. Check out Vanta and let them handle the tough stuff. Head to vanta.com forward slash the times tech to learn more.

Because when it comes to your business, it's not just about keeping the lights on. It's about keeping everything secure. So you know how last week, one of the things we're talking about was vibe coding? Well, I was successful because I've actually vibe coded a video game, guys. Yeah, you managed to do it. Yep. Do you want to know what it's called? Go on. Meatball Mania. Okay. And how do you play? You maniac. So hold on. I'll bring up the exact prompt.

Meatball mania. Can you share screen? No, that's too complicated here. Says the techie, the newfound techie. I can't share my screen. Anyway, I put in my prompt. I want to make a video game retro style kind of suit like Super Mario Brothers where the Swedish chef is fighting the meatball monster. Yes. And then you press enter and then all of a sudden the window opens on the right and all this code starts coming down the screen like the Matrix. And then what did you do with that?

Nothing. Then it goes, it's showing its work while it's writing this code. And then it's like, ready, ding. Would you like to play my game? Because I made a landing page. Yeah, well, you know. You're getting all techie. Can you find the chat box? Studio chat. I'm struggling.

Hello and welcome back to the Times Tech Podcast with me, Danny Fortson. Danny in the Valley and me, Katie Prescott. Katie in the City. And Danny, the Valley is where we're going to start and finish today because it's where all the money is. All the money. Yeah.

You always seem to get all the money and certainly you've got $40 billion more as of Monday in a deal that's been on the cards for a while now going straight into the coffers of OpenAI and it's the largest private tech deal on record. And I say it's been on the cards for a while. I know when we had Sam Altman on the podcast, there were rumours bubbling away at the time. And the money has, of course, come from the deep pockets of SoftBank.

What's interesting about that is we've seen this movie before. Masayoshi Son, who is the founder CEO of SoftBank. He is this kind of legendary figure in ways both good and bad. He was perhaps most famous for being the money man for WeWork and Adam Neumann. And that whole kind of disaster that unfolded where he was just

And then 15 minutes later, he said, "Yeah, here's billions of dollars." And obviously we all know how that went. It went terribly. Let's go back to what SoftBank is, because it's one of those, it's a bit of a misnomer. It's neither particularly soft or is it a bank. But Masa Son, it's often called Japanese, but Masa Son actually is a Korean guy born in Japan, Korean heritage.

who spent a lot of time in Silicon Valley. I think he was at university there at the end of the 70s when he really rode the wave of that period of Silicon Valley just as Microsoft was founded and Oracle and then brought it back to Japan. And that's how he made his money. It was bringing software like the soft into Japan. And then he made it. I mean, the WeWork story is brilliant and there are many more of those. But he also, he made his money, didn't he, in Yahoo, right?

was investing a massive chunk of money in Yahoo, which exploded just as search engines were taking off. Well, it was also his... By far, by far, the thing that he means he will never have to do anything else in his life ever again is he put $20 million into Alibaba in 2000. Just after the Yahoo investment. And that turned into...

At its height, like $75 billion. And they were quite reluctant at the time, weren't they? When Jack Ma was offered the money, he was like, oh, I'm not sure about this. And they went back and forth on it. But yes, it was that $20 million for 30% of the company or something like that. Yeah, $20 million to $75 billion, not bad. But yeah, as you say, the WeWork story is a brilliant one. And I remember meeting someone inside SoftBank who said,

You know those business books, like the sort of business disaster books? We could probably have our own shelf in a library of them. Because if you look at the WeWork investment, Greensill. Oh, yeah. Uber. It might not be a disaster, but it's a good story and it's got its ups and downs. Wirecard. Do you remember Wag? I don't. Wag was a dog walking app.

And this is, again, when SoftBank decided, you know, we're going to raise this $100 billion vision fund, the biggest fund ever for startups in conjunction with...

Saudi Arabia's sovereign wealth fund, and that really kind of completely changed the dynamics of funding throughout Silicon Valley. He just started spraying money. They started spraying money into all these crazy ideas, including WeWork. But one of them was this company I called WAG, where they put $300 million into a dog walking company, just an app to connect you with a dog walker. I remember thinking at the time, like, okay, we have jumped the shark. This is where...

I mean, there's a lot of crazy stuff happening, but that felt like the one, like this just feels like insane. And this is back in 2018. Obviously WAG is no longer with us. All dogs go to heaven. But SoftBank really is. And it bought Arm in 2016. There's another one. It's a mega success story. So the British microchip designer, the Cambridge microchip designer, Arm,

Much to everyone's upset over here because we don't create many big tech companies. To see it disappear was a source of upset. And it's recently floated that on the NASDAQ and it's just doing extraordinarily well. Yeah, so when he goes big, I mean, he doesn't do things. It's like go big or go home. He just always goes big. And sometimes it is WAG and WeWork and sometimes it's Alibaba and Arm. But whatever he's doing,

is at a level and valuations that just everybody else is kind of shaking their heads. But this past year, he's been like, okay, AI is the next thing. I'm doing it again. I'm going to go all in on AI. I'm going to start making very big bets. And of course...

A couple months ago, he showed up with Sam Altman at the White House to announce Project Stargate, this supposed $500 billion data center megaproject throughout the United States. And so he's back now.

Star Wars, exactly. He's back. There were some misses, though, weren't there? I mean, on the AI front, he invested in NVIDIA. He owned 5% of NVIDIA at one point, which he sold in 2019. Oh, bummer. It was, and he said at the time, he said, the fish that got away was big. That 5% of NVIDIA would now be worth $133.5 billion. Wow.

With a B. Oh yeah, billion, billion, billion. I will say about Massa, he has had some epic failures, but I think the Alibaba thing might be the best.

the greatest single investment ever. Yeah, investing in Chinese e-commerce before e-commerce was a thing in China. Yeah, yeah. Bada bing, bada boom. What do you think about this OpenAI investment? I rather wonder if he regrets not going in sooner. I mean, it's a lot of money, but it feels quite late. I mean, the thing is, OpenAI is growing so fast. They now have 500 million users, which is 100 million more than they had a month ago. So this is a massive company.

But again, I think I've said it before. I think the question I have is like, you know, it does feel like all these models are getting so good. What OpenAI is very good at is turning them into products people like. But ultimately, it feels like their biggest asset is the brand, ChatGPT brand. Because all of these models, it kind of seemed to be heading toward the same place, the same power, same capabilities, right?

with differences on the margin, it does feel like, you know, open source is very powerful, especially with things like deep seek and otherwise. And you're kind of wondering like, well, what's the, how, what's the quote unquote moat for open AI beyond its brand, but maybe that's all it needs. I don't know. I just think it's, it's quite a big swing because it could go,

not great for open AI. Yeah, especially investing at what feels like a peak like that. I think the other interesting thing is speaking to someone close to SoftBank the other day, is it shows how open AI is changing its corporate structure. And that was one of the caveats of them investing, as I understand it, was, you know, this kind of total laser focus on the commercial, as you say. Yeah. And I think the valuation is 300 billion, which is

Because when it was at 86 billion not long ago, people were like, whoa, that's crazy. But, you know, if you've gone from a few million users to 500 million users, I mean, that is not normal. Whether it's free, whether it's paid, anything. That is extraordinary. But one other thing I think to your point around the structure, this kind of strange structure opening I had where it had this kind of

It's set up as a nonprofit and they had this nonprofit board that was theoretically in control. And then there was, you know, if you go back two years when Sam Altman was fired very dramatically, you know, there's been all these questions.

around like what actually happened there because he was fired then rehired five days later it was this massive drama at the most important tech company in the world should it be allowed to make money Microsoft had invested a huge chunk of cash which direction was it going in exactly but there's this new book from a Wall Street Journal reporter Keech Hagee I think it's how it's pronounced yeah the book is called The Optimist Sam Altman Open AI and the Race to Invent the Future and

And of all the stuff I have read around what happened with that Sam Altman firing, rehiring, the reporting in this is by far the most detailed and seems to give the fullest picture. Because of course, if you go back when they fired Sam Altman, it was like he was not quote unquote consistently candid with the board. But there was nothing beyond that. It was like this parlor game of like what actually happened? Why was he fired? What is actually going on inside that company?

And in this book, which was excerpted in the Wall Street Journal recently, there's a lot of detail in there about what was going on. And just to give you a few kind of bits from that, one of the issues was that Sam Altman had personally set up the OpenAI Startup Fund, which is a fund to invest in AI startups using OpenAI.

And turns out it was in his name personally, even though publicly it was supposedly in the, it was controlled by the company. It was actually his personal investment vehicle. So investing in the companies that were building on top of the company's technology, uh,

And apparently this didn't come out until the board really started asking questions. The revelation came. It was either at a dinner party or an overheard conversation that one of the board members heard. And then it's kind of, they started asking questions. They're like, oh, he actually owns this thing personally. And apparently he was like, it's temporary or it was for tax reasons, but it went on for two years before it was found out. So that was one thing.

And then Mira Murady, the CTO and Ilya Soskovor, the chief scientist. Now no longer with OpenAI. Yeah. They had started kind of building a case about his quote unquote toxic management style, taking screenshots of things he was doing, trying to manipulate people, turn, you know, pit people against each other at the highest levels, et cetera. They were getting really worried. And so they had actually started building a case.

And one of the inflection points was Altman apparently told the board that the safety board, which is supposed to approve kind of public rollouts, had improved enhancements, several enhancements to Chad GBT before it had its big launch in November. Apparently they hadn't. Oh, and it's so interesting because they left so soon after that, didn't they, Ilya Sitskova and Mir Marati? Well, no. So this was the original launch. Okay. Yeah.

They were like, the board, you know, the board is supposed to, the safety board is supposed to sign off on things. And they, he just put it out in the wild and said, oh, actually the safety board had signed off on this and they hadn't. And they hadn't done it. Yeah. Or not on all of these enhancements. I think two out of the three they had not.

How fascinating. And then there was all this kind of Game of Thrones stuff around the board and Altman apparently saying, I want to get this woman, Helen Toner, out, who's very critical of him, and trying to kind of not be totally honest with people to try to kind of create a movement to get her out. All of this stuff. So when you step back, you're like, oh, all of that together, like, I get it. Because at a certain point, Maradi and Satsgavur

put together a document, sent it to the board and said, this is what your CEO is doing. Here are the receipts. And then days later he's fired. And then of course, what they didn't realize is that 95% of the company would be like, if you fire Sam Altman, we're also leaving this huge mutiny happens, uh,

And then Suskover and Marotti joined the mutiny and be like, bring him back. Yeah, I was going to say that's what happened, wasn't it? I remember them on Twitter saying, we support Sam. Do you remember there was a little phrase people were using to kind of...

To support Sam. And then, of course, Satya Nadella offered him a job and said he'd sweep up anybody who wanted to follow him. Come into the arms of Microsoft, he said. That sounds good. The report has done well, right? Because there's a huge omata around that. I remember trying to get hold of some of those people and it was very difficult. It's like the walls came right down. Yeah. And if you think about if the board had just been straight, it would be like, okay, here's the list of alleged malfeasance.

it would be much more clear what had happened. Because when you hear all of that and you put it together, you're like, okay, that actually makes sense. And the question, of course, which a very good tech writer out here, Casey Newton, asked, is like, would it have mattered? Would it have mattered if they put out like a one-pager and be like, these are the reasons. Would the mutiny have been so forceful? Would he still be running OpenAI right now and running the largest, the most valuable private tech company on the planet?

It's all super interesting, but I just thought it was a really interesting kind of blow by blow of the inner workings of what was happening at that time. But that, my friend, is the joy of being a private company. And that's part of the reason that whole episode was so strange because it's so heavily intertwined with Microsoft. There were massive questions about what it meant for their investment, what it meant for their AI strategy. And there was very little information trickling out that way. Yeah. And so we have asked OpenAI,

their take on this book. We have not heard back from them. Well, I love a good corporate kiss and tell book. We don't get enough of them. I'm halfway through the Facebook book at the moment. Have you started it? Oh, yes. I have not. The bit that struck me the most at the moment and that's taking over my LinkedIn feed are all the allegations about Sheryl Sandberg and whether she really believed in Lean In or not. Hmm.

Lean In, which was a book she wrote about how to succeed in the workplace as a woman and how you could really do everything and you just needed to try your best and assert yourself. You can have it all. You really can have it all and look at me and my perfect life and family and everything.

The way she is alleged to have behaved at Facebook is completely the opposite to that. So I think there are a lot of disappointed women like, oh, right, you really can't. Well, let me ask you, Katie. You're a woman.

Because I feel like for a while, Sheryl Sandberg was like this almost legendary figure. Absolutely. But that feels like a long time ago. Like the book came out, Lean In, and it had like, there were groups that started around the world, Lean In groups, and all these women supporting each other and blah, blah, blah. But it feels like that has faded a lot, especially since she left Facebook and also as more and more about like what Facebook is now.

how it is run i mean the book the title of that book you're referring to is careless people the one you're reading right now and from what i've i've read some excerpts and talked to people who have read it and who are around the company and it's like you know the the general vibe of just growth at all costs we know this is maybe doing harm in this way or that way but we kind of don't care

It does feel like the kind of the shine that the Sheryl Sandberg shine has kind of dulled and has been dull for a while, but I don't know if that's your reading. I think there was, as you say, this huge moment where...

She really seemed to capture the zeitgeist about how a lot of women were feeling about the workplace. And I don't necessarily think that was a Facebook thing. It was more just having someone who was a very, very successful female business leader who was willing to talk about the subject and engage with it. But yes, the reality, as is depicted in the book...

by the author who has a baby as if one that you've got to totally hide your family and your children have to be absolutely secondary to work. And it is, you say, work at all costs. And so I think to your point, yeah, the shine has... And maybe to the point you made last week, which I thought was really interesting, about actually people have thought about Facebook in this way for a while. But I hadn't thought about Sheryl Sandberg that much for a while. Right. The movement went, and I didn't think...

But the lean in had gone away. The book kind of popped that bubble for me. It should be said that the company is very angry about this book. I think they put the author under some type of gag order. Yeah. So she is she was doing press and then she stopped. And they're basically painted her as somebody who was fired.

And that she's bitter and that a lot of this stuff has been out there before. And it's like a kind of a recycling of kind of old grievances and whatnot. But they are pushing back really strongly against this book and reacting in a pretty heavily emotional way. They called it, quote, a mix of out-of-date and previously reported claims about the company and false accusations about its executives. That's their take.

And that leads us on nicely to our guest for today, who's actually a Meta board member. Oh, look at that. What a transition.

and transition beautiful so i did ask him about the book and he also pushed back on it we have drew houston yeah he's a he's a drop he's uh he's the dropbox guy he's the dropbox guy so he's actually he's a he's a he's the founder and remains the ceo and you know he's one of those dudes who from when it was like him and a him and a dog and a

PowerPoint presentation and now he's running this big public company. And a lost data stick, which is how he started the business. But I'm going to let him tell that story. It's now almost 20 years old, which I couldn't believe. That's wild. It is, isn't it? It does make you feel your age. But it was really the pioneer in this space of sharing, which all of the big tech companies do now. You know, think about Apple's iCloud, for example. But back in the day,

it was a really, really new thing. And it's an interesting business because Dropbox has managed to keep that brand awareness, even though it's surrounded by so many other people who are involved in file sharing. Because, yeah, it was like this kind of, it was like, oh my gosh, I don't have to carry around a little thumb drive or anything. I can just like put it in Dropbox and access it anywhere. And it was a real kind of,

And now it's like you've got Google Drive and everything else. Like it's like it's the way things happen. And that's why I'm always interested in companies like Dropbox who like come up with a big idea, build a company around it, build a brand around it. And then everybody does what you're doing.

And then it's like the dot, dot, dot. Okay. Like, all right, well now what do we do? Well, should we get to the interview? I'm fascinated by, by old man Dropbox. He's an old man. He's just in his early forties. No, I'm talking about the company because 20 years, 20 years in tech land is a long time. Well, let's hear from him. He was coming, he was coming through London, visiting some of his teams in Europe. And he, he dropped by into Times Towers on his way through.

I don't want to make you feel old, but I can't believe it's been 18 years since you started Dropbox. I mean, I'm about the same age as you, and I can remember it so, so well when it launched here in the UK.

How does it feel to you looking back on that 20 years? I mean, you know, has it flown by? It's a wild ride. Yeah, it feels both simultaneously very short and very long. But it's been, it feels like it's been multiple companies along the way. And obviously a lot has changed in technology since we started, but a lot has stayed the same. I mean, we really think about how do we, I mean, I started the company because I kept forgetting my thumb drive. So I was the initial customer that I had in mind.

And that was the initial problem I was trying to solve. But it was really this bigger umbrella of like, you know, why is it so difficult to find the information we need at work? Why is it so hard to organize it? Why is it so hard to share our information? Why is it so hard to keep it safe and protect it? And in the beginning, that was our stuff at work was largely files, and they were scattered across these different devices.

And today, we still have a lot of files on our computers, but we also have 100 tabs open in our browser windows. And so much of our work has shifted to these cloud tools, which have solved a number of problems. So we use all these new things for good reason, but it's created a bunch of new problems around, again, how do you organize and share information across all these platforms? And how do you deal with just sort of the fragmentation and information overload of that?

of that experience? And then how do you bring intelligence to that whole experience after Gen AI? So, you know, when I look at the last 18 years, it's never a dull moment. We're really proud of the scale that we've achieved, but then there's all kinds of opportunities for us to solve some of these foundational problems. Like how do I find and organize and share my stuff?

in new contexts. And so moving Dropbox through products like Dropbox Dash to helping people wrap their arms around all the clutter and basically build this self-organizing experience where you can search across all of your information. You can ask a lot of, Dash will answer a lot of questions that ChatGPT can't because it's not connected to your stuff. So being able to search in natural language across every platform or ask questions in natural language is

We think it's going to be a powerful upgrade that helps. And hopefully a lot of the same ways that Dropbox did just

taking something really complicated like AI and making it really simple and pervasive. I'm glad you got into that because that has been one of the things I've been wondering about with Dropbox going forward is how much you can innovate in the storage space. Looking back to 2007, you can see exactly what problem you were tackling. But I wonder now when customers come to you, what sort of things they want to see next? Well, first, I think we're all just overwhelmed by the sheer volume of information we're dealing with, right? So whether that's all the different cloud tools we're using or

or other communication tools we're using. Somehow we've ended up in this world where instead of one search box, we have 10. Instead of one inbox, we have 10, and we just get new ones every year.

And from my vantage point, it seems like the tech industry was sort of adding more, but no one was really better organizing what we have. So it saw an opportunity for Dropbox similar to where we started. And so customers often express these challenges like, oh, there's just not enough time in the day or my attention is so fractured across all these different apps and I'm just so overloaded and overwhelmed. And then on the positive side,

Our customers recognize that AI is going to be transformative in our experience, but often don't know how to connect the dots between what they see in sort of their personal lives with things like chat GPT, but then how do we bring some of that intelligence to work? And so I think there's a big opportunity for Dropbox to help take AI the last mile to people at work and take a lot of what we kind of, or what we experience and take for granted in the personal realm around these intelligent experiences and

You can search all of human knowledge with a Google search. You can ask any question to ChatGPT and it'll give you a good answer. But then when you go to work, that one search box, that Google or that ChatGPT isn't really there. And actually, if you try to take work information and paste it into consumer AI tools like ChatGPT, you create all these security problems. So Dropbox Dash is...

gives you that value, that one search box that searches everything. It also answers the questions that ChatGPT can't, like, oh, when does my lease expire? Where's that slide from last year's product launch? And does it in a way that's secure and designed for business from the ground up. So is the idea it's like a large language model powered by your own content? That's right. It's a bit of a shift. So when we started, we're storing all of your files and the value that Dropbox provided was storage.

It's a little different now. We're shifting from storing your files to really organizing all of your cloud content. And so what that entails is basically you connect Dropbox Dash. It's a new product. You actually don't even need files on Dropbox. But you connect Dash to all your different apps. And then Dash will go index everything, bring it all into one place.

And then you're able to both do conventional search or kind of the way you would do with a Google search, or you can ask a more direct question like, you know, what time does my flight take off from Dublin? And it will understand, Dash is able to understand what you're asking and not only find the document, but find the exact paragraph where it answers your question. I think a lot of our customers and prospects and probably most people listening hear a lot of different things about AI. On the one hand, all the things that are exciting. On the other hand, some things that are pretty scary in terms of

you know, what happens to my data if I give it to one of these models? Like, you know, is my data just going to be used to

target me for better ads or is my day or even worse is my like life's work and then be used to train some you know some company's next foundation model and with dropbox like don't have to worry about that because we keep your data private we actually self-host our own ai and this is where i see us able to contribute really able to contribute like how do we take something that's sort of exciting but amorphous like ai and really deliver concrete benefits and help you with very commonplace things like

helping you find the thing you're looking for or be able to get that question answered without having to go through like a million pages of court transcripts. These are the kinds of things that Dash can do. This episode of the Times Tech Podcast is sponsored by Vanta. Let's talk about something that might be keeping you up at night, cybersecurity. According to Vanta's latest State of Trust report, it's the number one concern for UK businesses and that's where Vanta comes in.

Whether you're a startup, growing fast or already established, Vanta can help you get ISO 27001 certified and more without the headaches.

And Vanta allows your company to centralize security workflows, complete questionnaires up to five times faster and proactively manage vendor risk to help your team not only get compliant, but stay compliant. So stop stressing over cybersecurity and start focusing on growing your business in 2025. Check out Vanta and let them handle the tough stuff. Head to vanta.com forward slash the times tech to learn more.

Because when it comes to your business, it's not just about keeping the lights on. It's about keeping everything secure. Okay, well, before we got to the world of generative AI, obviously there was a use case that you saw the niche in the market for back in 2007.

Can you tell us that story of how you started the business? You saw the idea for it, and then you got picked up by Y Combinator. Is that right? So it's the big Silicon Valley startup incubator. I just graduated from college. I was 24. And I'd started a different company before Dropbox, doing online test prep, which is a whole other story. But one challenge I had is everything in that company I kept on this little thumb drive.

that I kept losing. And so I actually took this bus ride from Boston to New York. I realized I'd forgotten my thumb drive. And thumb drive, you mean like USB stick? Yeah, USB stick. So it was stuck in my computer at home and not with me on the bus. So I suddenly had four hours of...

self-loathing, you know, because I'm like, why can't I be more organized? And I just, I'm like, I never want to have this problem again. And so I started writing some Python code, having no idea that it would turn into a product or a company, but it was pretty clear that this is something that all of us were going to have increasing challenges. Like, how do we get to our information across all these different devices? And then this, when we started Dropbox, it was in 2007. This was also the same year that the iPhone launched and

and a year after, Amazon Web Services and sort of the cloud launched. And then suddenly this went from a fairly niche use case when people mostly just had one laptop or something to suddenly when we have multiple computers in our lives.

this became a universal problem that we could address. And so I think the same kind of thing is happening with AI now. But yeah, in the beginning, it was really just sort of, you know, one of many little side projects I was working on that took on a life of its own pretty quickly. Teamed up with a classmate of mine from MIT. We moved out to California, got our first funding from Y Combinator. Steve Jobs had his eye on you. Yeah, we did.

Yeah, all these little things along the way. So the phone call from Paul Graham saying we were accepted into Y Combinator. He was running, it was his baby. He was running Y Combinator. Yep. We go back to the car. Our car got broken into. Our laptops were stolen. But our stuff was on Dropbox, so we didn't actually lose anything. So, you know, lots of stories like that in the early days. And what happened with Steve Jobs? Along the way, we got Apple's attention. And, you know, over time, got kind of escalated or introduced to more and more senior people there until...

One day, Steve Jobs wanted to meet us, and so we drove down to Cupertino. And we were going through our rituals. I was like, oh, man, maybe he hasn't seen the product. Let's do our little demo ritual. And I'm with my co-founder in a zip car pulling up to Cupertino.

one infinite loop there. But anyway, so we go in the boardroom and Steve comes in and basically they were interested in buying the company. So Steve launched into the pitch around why we should join Apple, but we also recognized that we were having a good time building an independent company and also for what Dropbox does, it kind of connects to every platform. So

We tried to steer the conversation away from acquisition, which he didn't love. He started saying that they're going to have to build their own competing product, which later became iCloud. But actually, it was a great meeting. The formal part of the, will you join Apple? That was only like 15 minutes, but he stuck around for another half hour just talking about it.

starting companies in Northern California and his return to Apple and things like that. So it's a pretty special experience. Extraordinary story saying no to Steve Jobs as well. It's amazing to think about it actually. 20 years ago, 2007, you say the iPhone launches and you've got the rise of the cloud. And now almost 20 years later, generative AI. Do you remember where you were when ChatGPT launched? Did

Do you remember using it for the first time and how that made you feel? I'd use some of the precursors. I'd use GPT-2 and 3, and those had some notoriety in the engineering community.

but hadn't become mainstream. Okay, so you'd had a chance to mess around with it. Yeah, and I'd been passionate about engineering and machine learning for a long time before, generative AI. But it was clear that there was a... As you started getting to like GPT-3, 3.5, and chat GPT, suddenly these models were getting really good at a lot of things, going from pretty incompetent to pretty capable across many dimensions simultaneously because they were able to take so much data and compute data

and create these larger models. And so probably to my wife's regret, I became quite enamored with ChatGPT and then using the GPT-3, GPT-4, or I got a preview of GPT-4 and

on our honeymoon. I don't think you're supposed to become enamored of other things on your honeymoon. I spent all my time on that. But it was definitely a distraction where I'm like, oh my God, these things can... How did we go from the GPT-2 where it would sort of... These models would generate sometimes coherent prose, but it would be kind of nonsensical to... Even back then, this is late 2022, suddenly not only...

to some of these models can read, they can write, they can write code. Clearly, it was kind of one of those shot heard around the world kind of moments for tech. And we've talked a bit about what it means for the Dropbox product. What does it mean for you inside the company when you talk about the quality of the code, for example, that large language models can write? How are you applying it? Yeah, I think about AI generally as being this

In some ways, there's a lot that's new about it, and that's what kind of captures our imaginations. But in other ways, it's a progression of something that's been happening for a long time. And I think we live in a world where we have our human brains, but increasingly we have this silicon brain that's paired with our brain. It's not one instead of the other. They're complementary. They have complementary strengths. And so as you have tools like ChatGPT or generative AI or more obvious expressions of that silicon brain because they can...

express intelligence in ways that seem very familiar and human-like. But I think at the same time, you could also say things like a Google search or a tool like Excel is kind of a similar amplifier or a silicon brain where a Google search can function as kind of infinite memory. Or a lot of these verbs that you would do as a human, like calculating things or even the word computer, used to refer to a person back in the 1600s. And so...

On the one hand, there's a lot of new capability that AI unlocks, but at the same time, there's been this general progression of

or perhaps a story of human progress is really we invent these machines that we offload a lot of our busy work to the machines and then we free ourselves up to invent better ones. And so I think that cycle has been going for many, you know, since the dawn of our species. So, you know, maybe I start there, but then within Dropbox, there's all kinds of ways that in each function, you have a little bit more of a silicon brain. So for engineers,

a lot of the toil of writing tests or doing these big migrations, or even the need to write all the code yourself, type it by hand, that's clearly shifting dramatically in the last six to 12 months as these models have become more robust at that.

The data point is so interesting, isn't it? I mean, it's such a hot topic, what happens to copyright. As you say, people are absolutely terrified, and we've seen numerous cases here, and I know you have in the US as well, about people worrying what do they do if their life's work is being used by companies to power their language models. How do you think they should be properly compensated? Well, I think, again, with some perspective, I think this is a, you know,

a complicated, you know, human and societal and legal process, right? Where you've seen it with journalism, you've seen it with music, you've seen it in pretty much every form of digital content. There's sort of this tug of war and, you know, lots of, you know, anxiety and hand-wringing over it, but eventually these things get settled by society and by the courts, and I think AI will be the same thing. Obviously, I think...

And to some extent, there's already some... There are some foundational elements of this around fair use and just copyright law in general, and certainly in the US. But there's something that will be pretty different. And yeah, I think we're sort of at the early innings of that whole political process. And you sit on the board at Meta as well. How do you consider it when you're wearing that hat? Well, I think...

The place Meta occupies in the landscape is they're the pioneer in open source. So I think that's been phenomenally successful. I mean, with my technologist hat on, and I grew up as a little kid coding, so the whole AI era is super exciting as an engineer, and I love it. And so it's been great to see

meta take really courageous steps to take something that could have been we could have had an alternate universe where this was all where intelligence was locked up among a few companies and you know we had to pay all of them rent to access it instead the world looks pretty different we've seen

of AI inference and the quality of these models and basically the price performance of intelligence, machine intelligence, just keep increasing 10 or 100 times every year. And I think open source is a huge lever in that and has been a really successful strategy with Meta and their Lama models being downloaded over a billion times. So it's a, you know, it's

It's been a lot of fun to have a front row seat to that. And I think Meta's driving a ton of innovation there. I'm glad that you brought up the cloud earlier because I really want to ask you about something that's boring but quite important. Just data center infrastructure. Because what you do is storage. And you've got, is it 90% of your customers' data stored on your own data centers? Yeah, probably even higher than that. Right, okay.

And so how do you manage that data center estate? Carefully. Huge cost to the company. Yeah, well, it's a huge strength for us, too. I mean, we've made, I mean, just a little bit about the history. We actually, Dropbox started on the public cloud. Yes, AWS, was that right? AWS, yeah. And they were and have been a great partner. That said, we recognize that there are a lot of, that first there are a lot of unusual things about Dropbox, given our scale and the nature of our business.

And the fact that we had all these great engineers, we realized that we could innovate not just at the application or user experience level, but if we went all the way down to the data center and the rack and the server and even the design of the server, like down to the metal, we could get huge advantages both in terms of cost because we could bring all of this in-house, but also performance. We could really specifically...

tune our infrastructure to the specific storage workloads or the things that our customers really needed. Whereas with the public cloud, you're sort of taking building blocks, but there's only so much customization you can do. Whereas we wanted to really get, you know, open up the building block and tune it even further. So that's been phenomenally successful for us. And are they mainly in the U.S.? And they're mainly in the U.S., although we offer...

various options to have local data residency. Sometimes, in some cases, we also have like points of presence globally to be able to make sure the service is fast and reliable globally. But then there's also options for local data residency if you want to have your data domiciled in one country. And sometimes we'll do that ourselves. And sometimes we'll do that through partners actually like AWS. I'd love to ask you about politics because you've just had a major regime change in the US. Yeah.

How do you feel about that as a tech boss? What does Donald Trump's presidency mean for you? Well, I think, you know, I think there's like a new episode every day. So I think we're keeping up like everybody else. You know, I think on the positive side, I think there's benefit to shaking some things up. And, you know, I think there's every large organization or every government can be more efficient. That said, obviously, like...

It's pretty chaotic and unpredictable, so I can't really... I don't know if I have any special insight as to where this is all going to go. Well, no, as you say, watching it unfold day by day, it just feels very... It feels like your world, the business world in Silicon Valley, has become intensely political as a result, particularly because you've got certain individuals, Elon Musk, who've been brought into government in a way that wasn't happening before. Totally unprecedented, yeah. It's pretty wild. And again, yeah, there's not really a...

Not really following a conventional playbook, and I think a lot of us can get behind more efficiency in government, but I think there's... I think what you'd also want is for these things to be kind of... Or you want these transitions to be orderly or compassionate, and so maybe...

And I think we're in the early innings here. We'll see how it all unfolds. Just want to know what's going on. Yeah. I mean, another uncertain thing is, of course, geopolitics for tech, specifically because you've got what's happening in the US and then you've got increased regulation in Europe and then the UK, I think, sort of straddling the two. And then even more extreme is the relationship between the US and China. How do you see that evolving over the next few years? And what do you think it does for Dropbox? Yeah. I mean, you know, we've had...

Or as I was saying, Dropbox has been a global company almost since inception. I mean, it took off in places outside the US as fast as it took off in the US. And so... You must have watched those tectonic plates move around. Totally. You know, ranging from things that sort of started in the US, like a lot of questions around civil liberties and the Snowden disclosures and how should we strike this balance between privacy and national security? Right?

And so that's played out in the U.S., and certainly cloud services have been part of that conversation. So that's probably the first of these kinds of issues. And it wasn't just the U.S. Obviously, every country cares about privacy and civil liberties and national security. And we've had some experience in navigating these kinds of things. And fortunately, the regulatory environment in Europe has improved.

It's been generally good for Dropbox. We haven't really had a lot of friction. Meta has. I mean, it's constantly crunchy for the big tech companies. Yeah, Dropbox has had a pretty straightforward relationship and a lot of things are common sense around striking a reasonable balance. We talk a lot about Meta and AI now, but obviously they're also innovating in the hardware space, right, with glasses and headsets. Mm-hmm.

For you, do you think that's where the future might be? Do you think the idea that we might be wearing a computer on our faces going in 10 years' time is what we might be doing? Well, I think we're clearly seeing that there's a world beyond the paint of glass in your hand, right? And I think it'll take many shapes. And I think there's also a fascinating intersection with AI or AI unlocks all these use cases that might not have been on the horizon before. So...

you know, the idea that you can put on sunglasses and then suddenly have a dialogue with machine intelligence and be able to have computing in all these new contexts I think is really exciting and meta was super early and...

with a lot of courage and conviction there. And I think we're starting to see a lot of the fruits of that. Do you make your own personal investments? Are there companies that you... I do. Yeah, little startup investments. So often these are like friends or former employees at Dropbox or just people from the community. What catches your eye? What sort of things do you put money into? I don't spend a ton of time... I mean, Dropbox definitely keeps my hands pretty full. So I'm not really like a serious angel investor. But it's been fun to be involved and...

you know to bet on founders pretty early on so companies like cruise or scale or figma the seed investor and all those oh wow okay um so yeah it's a great way to sort of keep to meet a lot of the next generation of founders or try to be helpful as they scale their companies and and also it's you know i learn a lot from observing what you know each new generation of entrepreneurs comes up with so what do you make of dash this this is their thing it's like okay

we're putting a chatbot essentially on on top of your i'm skeptical i'll be honest i i like i said i i what we're kind of to what we were talking about before like in a world of google and drive and all all of these kind of like ecosystems that they're trying to carve out their beachhead i just think it's tricky you know and this is a company that

They're worth like $8 billion. I was just checking their shares. And that's roughly what they've been worth effectively since they went public. Since the float, yeah. Yeah, since 2018. If I had to bet today, I would bet that Dropbox just kind of fades over the next five years or just kind of stays where it is. It's just like this becomes more and more of a kind of a niche player that some people... You know, it's like my...

My father-in-law still uses AOL for his email. You know, some people will be just wedded to the system and will just stick with it. But I just, I'm not, I'm skeptical it's going to kind of have a second life. It feels like

other companies can do it right. I mean, if you've got Google and Microsoft, which are also developing their own AI, that it's this natural direction of travel for them. But just looking at their results as well, I mean, as you say, their share prices has stayed about flat to what it was when they floated, but they're slowly, slowly growing their user base and their average revenue per user sort of.

ticking up. I think they've got sort of 18 million paying subscribers. But it does feel like they're going to have to pull a rabbit out of the hat at some point. Otherwise, they're just going to kind of bumble along. One of the really impressive things about meeting him was just how much of a geeky engineer he's still stayed. You heard him talk on the tape a bit about playing around with AI. So I

I bet he's going to go and do something quite exciting if he does move on from Dropbox. And, you know, he is obviously super well connected. He's on the board of Meta. He's like...

giving money to all these like next generation founders and stuff you know it's a it's a very you know silicon valley in that sense is like a village all these guys know each other they're mostly guys and invest in each other startups and all of that stuff and now danny the master coder you're in you're in the box you're in the mixer well you know you'll be next you said it you said it not me but i'm not gonna argue with you i'm a coder teaming up with drew soon

Yeah, exactly. Doing some AI gaming startup. I've got to set up my angel investing operation, though. But also, I need millions of dollars. So if you have millions of dollars or pounds, I'll take pounds. I don't, but you're in the right place for them. Or ask Massasson. SoftBank's got a few, Bob. I'll ask him. I'll ask him. See if the Saudi and Arabian sovereign wealth fund. And I hear he likes giving it away.

Meatball mania. Hello. Start game. And there's literally a little, oh my God, they're just brown blobs. What am I doing? What is going on? There's brown blobs dropping from the sky. Yeah, you have to shoot them. How do I shoot? Oh, I see. Left and right with the keys. Yeah, how's your instructions? Oh, left rise arrow to move space to jump.

Z to throw label. Cool. But Danny, it only shoots sideways. Yeah, well, I mean, come on now. I mean, like I said, this is the product of about two minutes. This is amazing. It's just a little red rectangle.

That's the Swedish show. Zoom. Oh, yeah. Zooming back and forth on the button. This is brilliant. And it looks a bit, as you say, it's like old school graphics, like Lemmings-esque. Yeah, exactly. Very blocky. Okay, I'm coming out now. Otherwise, I won't do anything else. Brilliant. Well done. Thank you. Vibe coder of the week. I'm the vibe coder of the week. Maybe don't show it to anyone else. Yeah, exactly. Exactly. It's all right. It's not your real job. Okay.

Can I ask one other question, which is probably pointless. When you're shooting your laser, do you have to say what direction it's going in or does it just auto do that? First of all, it's not a laser. It's a ladle. Ladle, yeah. If you press the arrow key in one direction or another, then that's the direction it shoots. Got it. Okay. Cool. Well, I look forward to a few more games. Yeah, enjoy. Enjoy. See you next week.

This episode of the Times Tech Podcast is sponsored by Vanta. Let's talk about something that might be keeping you up at night, cybersecurity. According to Vanta's latest State of Trust report, it's the number one concern for UK businesses, and that's where Vanta comes in. Whether you're a startup, growing fast, or already established, Vanta can help you get ISO 27001 certified and more without the headaches.

And Vanta allows your company to centralize security workflows, complete questionnaires up to five times faster and proactively manage vendor risk to help your team not only get compliant, but stay compliant. So stop stressing over cybersecurity and start focusing on growing your business in 2025. Check out Vanta and let them handle the tough stuff. Head to vanta.com forward slash the times tech to learn more.

Because when it comes to your business, it's not just about keeping the lights on. It's about keeping everything secure.