Episode 7: CEO Drew Houston on solving the biggest problem with modern work

For our seventh episode of Working Smarter we’re talking to Drew Houston, the co-founder and CEO of Dropbox. 

If you’ve been online long enough, it’s likely Dropbox was your introduction to the cloud. The goal is still more or less the same—give you one organized place for all your stuff—but it’s no longer just about storing and syncing files. A hundred files on your desktop is now a hundred tabs in your browser, and Houston believes AI is what will finally bring calm to the chaos that’s been created by the tools of modern work.

For Houston, AI’s potential is so great that its arrival feels like a civilization shift. It’s also not just a professional preoccupation; AI is a personal interest too. A few years ago he decided to teach himself machine learning in his spare time—and some of the AI tools Houston now uses to run Dropbox are ones he built himself.

Hear Houston discuss why it’s gotten so hard to find the information you need to do your job, the types of tasks we'll increasingly offload to our silicon brains, and what Dropbox is doing to help make modern work more meaningful and fulfilling.

Full episode transcript

This might sound quaint, depending on your age, but there was a time when I could hold my entire digital life in the palm of my hand. Like, literally—on a floppy disk or a CD or a thumb drive. You know, when people still had files to store.

Then the cloud came along, and everything changed. I didn’t have to think about where I put my stuff anymore, or how much of it I had. Wherever I worked, whatever device I used, my documents were just there. And once our files moved to the cloud, so did our apps. We took this antiquated concept of files and folders—and blew it all up.

But when we started to sprinkle bits of our jobs and our lives across all these apps and services, we also turned knowledge into something more…amorphous. More abstract. Something that was harder to hold.

I sometimes think about how simple things used to be. How I could use my hand-me-down laptop for days at a time without ever going online, because everything I needed—all three of my programs and my floppy full of files—were right there on my desk. Versus today, where doing my job can sometimes feel like playing one of those arcade games—you know, with the joystick and the claw? 

It doesn’t matter if you know what you’re looking for. You’re probably not going to get it on the very first try. Or the second. Or the fifth. I know that’s by design. But maybe it’s time we built a better claw. 

I’m your host Matthew Braga, and on today’s episode of Working Smarter, we have a very special guest. Someone who knows what it’s like to be able to fit your whole life onto a thumb drive—but then forgets that thumb drive at home, and gets so frustrated by the experience that he builds a product and a company so that it never has to happen to him again.

Drew Houston is the CEO and co-founder of Dropbox. If you’re anything like me, Dropbox was your first introduction to the cloud. But these days, the company does a lot more than just store and sync your files. Today, Dropbox is building AI-powered tools that help make people’s digital lives more manageable and accessible—that make you feel like everything you need is right at your fingertips when you sit down to do your job. 

Like a claw that actually works. 

Drew and I talk about the transformative potential of AI in the workplace, the types of tasks we'll increasingly offload to our silicon brains, and how an AI-powered Dropbox can help you and your team collaborate, find focus, and spend more time on the work that matters most. 

That’s coming up next on this episode of Working Smarter.
 

~ ~ ~

Drew, thank you so much for being here today.

Thanks for having me.
 

This is great. So I want to start off with maybe a little bit of a bigger picture question here. You've worked in the tech industry for a long time now. What about AI—especially in this kind of present moment—would you say is different from some of the other kinds of hype cycles that you've experienced?

I mean, first, AI is clearly the next cycle, the same way that mobile and cloud were when Dropbox started. Like, those enabled Dropbox to exist in the first place. 

Some things are the same. This hype cycle, we had this sort of explosive interest after ChatGPT. That was kind of the starting gun for the world and AI when it really came into people's consciousness in a much bigger way. And then we're starting to see, on the one hand, exuberance and crazy 2021-type valuations and things in startups—and at the same time, large scale implosions of some of these AI startups. 

So I think the kind of mayhem is a bit similar. But what's different is the magnitude. And when I think about the cycles that I've lived through in my lifetime—like, I was born into the PC era, and then saw the rise of the internet when I was in middle school, high school, and then the rest after—it feels like AI would be more up there with computing or electricity or the printing press or fire.
 

Like a fundamental kind of shift. 

Like, civilizational shifts. And the reason why is, you think about things like cloud or mobile, they certainly unlocked all kinds of new behaviors for us. It's hard to point to aspects of life they didn't change in some ways. But when you think about the industrial revolution, for the first time we could break free of the limitations. Like, energy was no longer limited by your muscles or your animal's muscles, and suddenly you could conjure up energy on demand. You were able to offload a lot of your physical heavy lifting, quite literally, to machines, but then also create this renewable—it became this thing that you could kind of bottle up, energy became this thing you could bottle up. And then demand wildly increased. After you have things like electricity, suddenly our consumption of energy went up a lot. Then that translated to a much better standard of living and all the abundance that's happened since then. 

I think what's really consequential about AI is it's sort of analogous to what you unlocked with physical energy before, now you're able to unlock with cognitive energy. And the promise of AI is like, now, for the first time, we can bottle up intelligence and convert electricity plus silicon into cognition and reasoning. That's a super powerful thing. And then bringing it more down to the concrete, there's all kinds of new ways to automate a lot of our busy work and offload new kinds of work to machines. So that same kind of unlock we saw in the physical realm, now we're going to start to see in knowledge work.
 

When we talk about some of that cognitive unlocking that AI will allow us to do, what exactly is it we’re trying to solve? Like, what is the problem with the way we work, with the way that we do our jobs, that you think AI is uniquely primed to help us fix?

Well the timing is really good for AI because one of the biggest challenges with modern work, I would argue, is cognitive overload. The information coming at us, or the number of things we need to juggle, is a lot more than our brains were adapted to handle. In the last four or five years, we shifted to remote and distributed work after the pandemic, and now we're working much more out of screens then out of physical offices. We're putting a lot more stress on our digital environments, and they've become very fragmented and overwhelming places. 

I think we're all familiar with a lot of these challenges. It’s really like a thousand paper cuts, right?It's a lot more challenging these days to get the information that you need. We know from brain science that you're most productive and most engaged when you can focus, or you're in some kind of flow state. And yet, go look at what's on your laptop screen. If you wanted to design an environment that made it impossible to focus, or impossible to get into some kind of flow state, or something that was always interrupting and distracting you, what we have is just a disaster cognitively.
 

Totally. Well, and maybe you can tell me, just for a second—there's a device that you have in front of you that seems uniquely designed to solve some of that stuff.

So I have a reMarkable tablet, which is sort of like an infinity paper notebook—although it's electronic, it's an E Ink tablet, basically. Over the years, it started off with paper notebooks or legal pads. I'd fill a bunch of them up a year, walk around with them anywhere, and they just kind of helped me focus and stay on track. This is sort of an augmented version of that, where it doesn't run out of space so I can have basically everything I've ever written with me. It's synced to the cloud so I don't have to worry about losing it, and a number of other benefits. It’s something that more aligns to focus. It can't send you notifications, it's really just the screen. It helps me be more present in meetings and use the time better.
 

You kind of alluded to this a minute or two ago but where would you say that we tend to lose the most time or the most focus when we're working today? What kinds of things are sapping our attention and making it harder to do our jobs?

People have done time studies that sort of look frame-by-frame, or second-by-second, at what people are doing at work. Even several years ago, McKinsey did one of these studies. They found that knowledge workers spend probably about 60 percent of their time on basic tasks like finding information, communicating with people, things like that—and only 40 percent of the time, the minority of the time, on the jobs you were hired to do.

And it doesn't mean that the 60 percent is totally a waste. But it does feel like—we've called this “work about work”—that is becoming a large and growing percentage of time where we're doing all this coordination and these information tasks. And that was several years ago. So I imagine those numbers or those trends have only continued and gotten more acute after COVID. But yeah, it's a thousand paper cuts. It's like, “Oh, I can't find the thing I'm looking for, and now I've got 10 different places where my content lives, and I've got 10 different inboxes,” and just the complexity has exploded. Which means that the idea of having one organized place for your stuff is further and further out of reach for people every day. 

It's a bit ironic because 20 years ago, or even before that, when I would go visit my dad at work, it was different. Like, a lot of things are the same. I'd go into his office, he had a PC, he had a desk, he had a phone. But a lot of things were different in that, you know, he got like five emails a day and not 500.
 

How nice.

Right? And he had an office with a door that could close. And you could see how every time he added a new tool—like, a phone call could save a meeting, and then an email saves a phone call. Or a google search saves having to research something. Or all the time that having a spreadsheet or Excel would save. And it was very clear back in those days, as you stacked on these tools, how much they were all kind of new superpowers. But between then and now, as we went from ten tools to a hundred tools to a thousand tools, something kind of shifted where the tools went from helping us do the work to becoming the work in a lot of cases. 

Think about how much time people spend in their email, and now their Slack. There's this kind of dysfunction in the system where we're giving people, individuals, more and more powerful tools to spend a hundred other people's attention. So any one interruption or any one little friction point is not a big deal, but they stack up to become—you know, the majority of the time we spend, you could argue, is friction with our tools.
 

It's comforting in some ways to hear that you struggle with the same problems of overload as pretty much everybody else.

Absolutely. Yeah.
 

It’s a universal problem. We've talked a little bit about some of the ways that you try to deal with this. You have this single purpose device in front of you that you use for note taking. How else are you finding focus in your day? What are other things that you do? Or is it a matter of using tools to help deal with the tools? What are you doing?

Well, it's a constant struggle. I'd say I deal with it with mixed success. And honestly, this is kind of my playbook as an entrepreneur. I started Dropbox because I kept forgetting my thumb drive, and now we're tackling a lot of these problems because I'm just like a frustrated end user. I’m like, the whole point of knowledge work is you hire these people for their minds, and then it feels like none of us have the time or space to think. 

A lot of what I do is try to adhere to productivity 101—kind of motherhood and apple pie things. Like, get clear about what your priorities are. Be proactive about managing your calendar. Make sure that your time aligns to your priorities. Some of my favorite books—like “High Output Management” by Andy Grove or “The Effective Executive” by Peter Drucker—have really good thinking, some of the best thinking I've ever seen, about: how do you be effective and have high output? There's a lot of concepts like not just managing your time, but what are high leverage activities? What gives you a lot of output per minute? Things like that. But I also found that trying to implement those things is always an exercise left to the reader, and is often a lot of rote mechanical tasks. Like, okay, I’ve got to manually audit my calendar. I’ve got to manually do this, manually do that, reconcile this, live out of my inbox. 

So what I've done over the last several years is apply automation in new ways. I grew up as a kid programming and knew I wanted to start a company eventually. But when I actually went from startup founder to really being a CEO and managing a large team—like, man, there's a lot of repetitive tasks. I finished my undergrad, but didn't do a PhD or a master's in computer science. I would have liked to. But I never really got to study machine learning. So several years ago I started learning the things you would learn in a grad level AI or machine learning class, in applied statistics, things like that.
 

Where were you going for that? How were you teaching yourself those concepts?

For me, it was a lot of just tinkering and playing with things. There's a lot of good open source libraries for machine learning, and just building little toy apps to be like, “Alright, well, instead of me manually tagging my calendar, or having my team do it, I wonder if I can train a classifier to do that.” And then machine learning is exciting for an engineer because when you learn programming basically the way it works is the human figures out an algorithm and then applies it to the data. Machine learning kind of inverts that where, instead of the human writing a program, the human gives data and then basically the machine figures out the algorithm for you. So that was just a really intellectually exciting concept.

But yeah, a lot of toy apps. I use machine learning to classify my calendar. I use machine learning to triage my emails—like, which ones of these are important, which one of these needs a response. And things like tracking my screen time. So a lot of little toy apps to try to automate some of the more repetitive aspects of it. And this was all pre-large language models. It was 2016, 2017, 2018. But there's a lot of new colors you can paint with now.
 

I was going to say, are you still using permutations of these apps that you built, these toys that you built today?

Oh yeah, yeah. All the things I described—I mean, I even have a little focus button. So in my menu bar, next to my Dropbox icon, I can be like, “focus on.” And then it will actually kill all my distraction-generating apps. It closes my email, closes Slack, puts on do not disturb. It's kind of a taskmaster for me, so if I start meandering off into—well, it also shuts down internet rabbit holes. So it's like, you can't go on Hacker News or Reddit or Twitter or things like that. It blocks all that to create a little bit better conditions for focus.

And then, after large language models and after ChatGPT and AI, now I'm doing a lot with AI-assisted writing, coding to some extent. There's a lot of stuff where I kind of hit a wall before large language models. For example, trying to prioritize or triage your email. Classical machine learning is really good at things that involve numbers or structured data. But as soon as you give it a blob of human generated, malformed text—English—before large language models computers didn't really know what to do with that. And the whole natural language processing field was pretty primitive compared to what we have now. But then, the large language model completely blew through that limitation. So now things like, “Take this long email that you've received and then identify: what exactly is the request? When does it need to be done? Do I impose some structure on this unstructured data?”—there's things like that. That's a pattern and an unlock that's super powerful. 

In some of my little pet projects I figured out, if the toy apps I was making were level one or two, now you can go to, like, level 10. And now we have actually a little team in Office of the CEO that's like a little lab or a little incubator for a lot of these concepts. Maybe it starts with the seed, the toy version that I came up with, and then handing it off to a team to build it out fully. 
 

And that's a really good segue because, I think up until now, we've been talking about ways that you’re trying to find focus in your day—things that you've been experimenting with and building. But what is Dropbox as a company doing? You kind of alluded before to there being this shift in how Dropbox is trying to tackle some of these problems.

We started by syncing your files. Along the way—as I found myself running this big organization, in meetings and inboxes all the time—I was like, “Wait, knowledge work seems to have lost the plot a little bit.” And we shifted the company's mission to “design a more enlightened way of working.” In the long run we want to go to the other end of the spectrum with our company's mission. The company spends most of its resources in the middle. It's like, okay, there's a pretty big gap between file syncing and solving knowledge work. But there's a lot that we can do in the new world now that AI is coming online.

So we're focused primarily on what I would describe as solving the 2024 version of the problem that we started with. When I started Dropbox, the problem I was trying to solve was I just kept forgetting my thumb drive, or I kept having to email myself files—you know, all the little paper cuts back then that we experienced from not having our stuff in the cloud. And the solution back in 2007 was like, “All right, well, my stuff is in files on these different devices. Let's sync it up.” And that was how you solved this problem of like, “Ahh, my stuff is everywhere, I can’t find it.” Okay, give people one organized place for everything, and access it from everywhere. Fast forward to today, and that’s kind of the problem we have again. It’s like, alright, well, our stuff is everywhere. Can't find it. 

A lot is different. We have all these new tools. A hundred files on your desktop is now a hundred tabs in your browser, right? And a hundred files on your desktop. There have been these regressions or these steps backwards. And so on the one hand, we're adopting all these new tools because they're great. They give us new superpowers. But we are trading one set of problems for a new set of problems. Everything's very fragmented and fractured, and this manifests in all kinds of ways. 

I mean, search is probably the most obvious example where we're living in this bizarre scenario where it's easier to search all of human knowledge with a Google search at home. Then you go to work and you have 10 search boxes that each search 10 percent of your stuff. No one would really design something like that—and actually, that was much better 20 years ago when, if you wanted to find something, you just searched your hard drive and it's there or it's not. So this is sort of a new problem. 

And then there's just a lot of stuff that, you know, anytime you have this change—as we've evolved from physical papers on a desk, to files on a hard drive, to now tabs in a browser—you gain things and you lose some things. You often forget or you sort of relearn why those things were helpful in the first place. So, in addition to search being broken, organization is broken, persistence is broken. When you reboot your computer your files are still there. When you go to sleep your papers are still on your desk. But in the browser realm, if your computer updates itself the wrong way—or more commonly, you just declare tab bankruptcy—you're clearing out your whole workspace when you close your browser. And then you literally have to start over from scratch.
 

Because that’s the best choice you have. 

Yeah. And we could keep going. Sharing has a lot of issues, which is strange, because there's no concept of a collection. Files have folders, songs have playlists, but when you start to think about stuff on the web—URLs, web content—there's not really a consistent concept of like, “Oh, here's a set of links.” Most of us have to do other workarounds, like, “Oh, I'll have a cloud doc that links to other cloud docs” or “I'll put all these things in an email.” But to me those are very reminiscent of “I'll just carry the thumb drive” or “I'll email myself this file” because then I know—blah, blah, blah.

But I think we'll find it very strange that there was no sharing primitive, right? That we had no concept of persistence or organization. When you're working on something with other people, if you're re-modeling your house, you're getting ready for a board meeting—if you have a Google Doc, and an Airtable, and a 10 Gb 4K video, there's no common container that holds all those things. And so that contributes to the kinds of paper cuts. 

You also take search and then, one way or another, you're ending up in a world where you can't find the information you need. It’s very hard to do knowledge work without the knowledge. You do dysfunctional things like, “Oh, I can't find it myself, so I'm going to go interrupt you and Slack you and use you as my search engine”—and then be a search engine for someone else. Get a Slack, get interrupted. So Dropbox is really trying to fix, tackle, a lot of the information problems—most notably with our new product, Dropbox Dash, which does AI-powered universal search.
 

Can you elaborate just a little bit on what that means in practice?

It takes you back to one search box for everything. So Dash connects to not just your files but your Google Docs, and your Slack, and your email, and your Salesforce, and whatever you have. It indexes everything, identifies all the stuff you're working on, and then makes it easy to search. When you do a search in Dash, it will search all of your apps and content. It's like a private Google. 

It also takes advantage of a lot of what we're seeing with the kinds of benefits you see from things like ChatGPT—and Dash can answer a lot of the questions that ChatGPT can't. For example, if you ask ChatGPT, “When does my lease expire?” or “Where's that slide from last year's product launch where we talked about that thing?” ChatGPT can't answer those questions because it's not personalized to you. It's not connected to your stuff. But Dropbox is and Dash is.

So by connecting to everything, we can use large language models to do not just universal search, but get universal answers—to be able to ask personalized questions, or basically have this silicon brain that has read everything your company's ever written, and be able to converse with it in natural language like ChatGPT.
 

I'm curious, for people who still associate Dropbox with cloud storage and files and not for some of the AI powered tools and features that you're describing right now, why is Dropbox the company that can pull this off?

It's a pretty natural evolution for us. We've already helped millions of people and millions of companies or organizations with organizing their important files at work. So it's not that much of a logical leap to go from syncing your files to organizing all your cloud content and solving a lot of the problems we were talking about around search, around being able to access your information, being able to organize it. Our biggest limiting factor to growth today is the fact that most people have some kind of file syncing solution. There's not some new continent of people who haven't heard of Dropbox. And so we have a bit of a saturation problem in the business we're coming from.

But if you sort of think about the market for organizing all of your stuff, and things like universal search—How do you organize all your cloud content? How do you like, have a better work environment?—literally no one has these problems solved, right? There is totally a green field up for grabs. A billion knowledge workers and growing have these kinds of problems, and so it's a huge opportunity and a natural evolution for Dropbox. 
 

What about this moment in time, as well? Like, how did this moment in time from the pandemic to Virtual First inform your thinking about what Dropbox should be building and who it's building for?

I think the pandemic was quite a turning point in a lot of ways. For Dropbox and for me, I mean, first it was just the disorientation of like, “Well, okay, I guess we're locked down now and we're all working from home." And then also just dealing with the incredible trauma of all the uncertainty and things like having family members that were either sick or at risk of being sick, or getting furloughed or laid off. So first it was just the mayhem of that, but then pretty quickly recognizing that this is a huge opportunity, or there's a big silver lining here—where, for the first time our work could be decoupled from our physical location and environment. I don't think anybody wants to think about if the pandemic had happened 10 years before. Before Zoom or all these cloud tools, it might've been a very different and worse story. 

But I think we had already been on this transition to where the percentage of our time at work we spent on a screen was already increasing. Then COVID kind of finished that migration. Now we primarily work out of screens more than offices. I love all the technology, but some of my heroes are also the management thinkers and folks who have thought about, like, what is the fundamental nature of work? How does it change? What's the history of it? I mentioned people like Peter Drucker—like, people that rethought what work is and what it could be. Everybody gets that the flexibility is great, not having to commute, all that. Those are huge benefits. Importantly, once you give people that kind of flexibility, they don't want to give it back. 

But there are some problems, right? It can be this very literally disembodied experience. It can be fatiguing to be on Zoom calls all day. One of the biggest things you lose by not being together in person is you lose all this context and peripheral awareness and information. So when you think about “Why did we pick Dash?” or “Why did we pick this problem?” part of it connects to that issue. If you're in a distributed world, and you can't just lean over to the person next to you and get a quick answer, suddenly you have to do all of that through a screen. And that screen was just not designed to handle, you know, even an order of magnitude or a fraction of the stress we're putting on it.

So that very basic challenge of “How do I just get the basic information I need to do my job?” is actually harder in this distributed world. We made a very deliberate decision to be like: no, we're dealing with these problems too. They mostly look like problems, but they could actually be a huge opportunity, because literally all of our customers are dealing with the same things—and let's have Dropbox become this kind of lab for distributed work, because no one's really doing that.
 

Earlier I heard you use the phrase silicon brain when you were referring to some of these tools—the superpower, for lack of a better word, or ability that it gives us. Why does that feel like the right frame to you for thinking about these kinds of tools?

We've kind of had a silicon brain for a long time. We've been offloading cognitive heavy lifting to machines for a while—starting with writing. The fact that you can write something down in a persistent way, or have language, is actually a technology that we had to invent that gave us an infinite memory and allowed you to pass things down through generations and things like that. There are many of these kinds of things, but computing was another example. Excel or your microprocessor can do a lot more calculations or computation than any human can. 

But basically, the human brain and silicon brain have complementary strengths. Historically, humans have been really good at the conceptual, relational. That’s one set of things. And then computers have a mechanical intelligence. They will follow instructions tirelessly and do them really fast. The computer will not complain or get tired if you give it like a million documents to parse or something. But historically it has been very brittle. The computer, to a fault, will do only exactly what you tell it to. You end up having to be very prescriptive about each step of that process. 

And so AI and augmentation actually fixes a lot of those limitations where computers, for the first time, can have that kind of world knowledge or conceptual understanding, or can do a lot of tasks or get a lot of skills that were previously reserved to humans. Computers can now see and hear and listen and talk and read and write natural language. Soon robots will be able to do that more in the environment. So now we have all these new ways to offload a lot of our busy work to our silicon brains and actually use our silicon brains to help our human brains focus.

A lot of that distraction or friction that I talked about, computers have no problem with that. And they will happily—if we're able to re-segment and redesign work, and build the right tools, build the right environment, we can actually have a much better division of labor.
 

Well, and not only that, but I think I've heard you say that it also frees us up to do more of the work that matters. 

Exactly.
 

And I'm wondering, like, what is the work that matters to you? Like, how do you define that or think of that?

There's a lot of enthusiasm about AI right now, for sure. Justifiably so. But there's a lot of things that only humans can do for the foreseeable future. Computers don't really have soul or taste, or don't have the same kind of creativity, or the ability to synthesize information the same way. I find it better to talk about it more as like an alien intelligence where it's not really better or worse. It's just sort of different. And instead of fixating on, like, “Oh, here's all the ways it's going to substitute humans or ways it's not, as it isn't or never will be as good as a human”—like, that's true, but I think the useful question is: what can it do? How do we just offload as much as possible? 

And I think this augmentation concept is really powerful because, I mean, think about Google Maps, right? That's a great, positive example of what I think this is going to look like. Look at self-driving cars. That idea first came around ten years ago, and everybody thought, you know, “In the next 10 minutes no one's going to ever drive again” and things like that. That turned out to be a bit off in terms of the timing. Things like Google Maps have had a much more transformational role in the driving experience than self-driving. And like, why is that? Well, you have this symbiotic, positive partnership with a machine where Google Maps will route you to your destination and recalculate all these things. 

New generations of kids will live and die without the experience of, like “Hey, did I miss my exit?” Or my dad would have an atlas in the back seat of our car, or you'd have to, like, print out things on MapQuest—these things that just seem totally archaic. Now we don't have to really worry about navigation. A 16-year-old who just got their driver's license can kind of out-navigate a career cabby or whatever, you know? Humans are still driving, humans are still picking the destination, but it's just made easier by the silicon brain in Google Maps.
 

I like the Google Maps example because it’s easy to forget there was a time when we had to learn to trust these things. We had never seen something like this before. We needed to understand it and get comfortable using it. And now of course, everybody uses this stuff. How do you get more people to trust and embrace AI, and AI powered tools, when this is something that they haven't seen before? That there's not really a good past precedent for?

I think it'll be similar to what we saw on the internet, with the rise of the internet or computing or the iPhone. There's a lot of enthusiasm, right? So I think the first thing that drives adoption is utility—like, it's just super useful. The internet was like that. I remember the first time I could send basically send a postcard around the world in a hundred milliseconds. Mind blowing. ChatGPT was that for a lot of people. Suddenly here's a thing that has a PhD in every subject—that not only has all this knowledge, but then also has all these useful functions. It can write for you. There's a lot of busy work it can help either assist or automate. And unexpected things, like being able to generate visual memes, or being able to write poetry—like, write a sonnet about whatever you want. I think it's done the capturing people's imagination part.

But then, on the flip side, there's a lot of fear, justifiably, as there was with the internet. “I'm not going to put my credit card online to buy something, that's insane” or, like, “Oh, I'm going to get hacked and all my stuff's going to get…” you know. We had to navigate a lot of this. Even with Dropbox, we had to navigate a lot of these kinds of fears. So I think, for good reason, people approach this with some apprehension. Because there's very real trust, safety, security, and privacy concerns—concentration of power, bias, all these kinds of things. People are, rightly, out in front of these things more than maybe some of the last generations of tech where people just focused on the wild-eyed futuristic piece. And then when we realized, like, "Oops, yeah, we let something out of the barn that we probably could have steered in a better direction”—I think, fortunately, we're learning from that. 

But we're going to go through that kind of process. It'll be super useful paired with big new problems. Importantly, you can't really stop these tidal waves, but you can kind of shape them and steer them. And that's a lot of what I spend time thinking about, certainly for what we build at Dropbox. Like, how do we steer it to be an aggregated force for good?
 

One of the advantages of AI powered tools is freeing up time to do some of the work that matters. But of course it's not always about work, right? I mean, it could also allow you more time to do things that matter to you in your personal life, in your personal time. I'm wondering, what does that look like for you? What are you doing for fun in your personal time with the space that some of these tools are creating.

Well, first, I'm a new dad.
 

Okay, congratulations.

My wife Erin and I had our son Charlie. He's five months old. He's made a couple appearances, little cameos on our all-hands meetings and things like that. So that's a lot of fun. But honestly, I'm sort of back to being that little kid who was programming and just in love with technology. I think AI is—I'm now coding like an 18-year-old again, and because I track my time, I was like, yeah, okay, last year I spent 350 hours coding or something, which is the equivalent of eight 40 hour weeks if you think about it. Some of it's building the little toy apps that I was talking about. A lot of it's just playing with large language models. I've got some GPUs in the cloud, and I'm always trying out the latest models and sort of twisting them and bending them and seeing what they can do and not do and kind of trying to become a bionic CEO. 

For me, this is fun. I think a lot of people might be like, it's not very balanced. But I really love this stuff. I'm very interested in this stuff personally. Dropbox is also very interested in AI as a company. Fortunately, these things line up. And so actually, it's been a lot of fun to have that be integrated. Both have a lot of the intellectual side of it—off hours, to be writing lots of code, but then also on the Dropbox side, really working on: how do we build the best products that really take advantage of all the new colors we're painting with?
 

I want to go back to something that you alluded to earlier on. You mentioned the Dropbox mission, which is designing a more enlightened way of working. I'm wondering if you could tell me a little bit about what success there would look like for you?

I think there's a lot to do to return work to a more engaging and productive state, and just sort of realign the experience of work with what's good for us cognitively and spiritually. Some of this is why I say more enlightened way of working, and not more efficient way of working, more productive way of working. It's not just about making the treadmill go faster. We're both most productive and most engaged, most fulfilled, when we're calm, when we're able to focus, when we're doing meaningful work. We all deserve that in a lot of ways. We're sort of tripping ourselves up with the last generation of tools. I think that'll look very different in the next generation of tools, both in helping us focus, but also freeing us up—eliminating a lot of the busy work and having every minute be time well spent. 

It'd be awesome if we got a future that's much less busy than today and a lot more meaningful. I think there's a lot that Dropbox can do to start to put a dent in that problem and reverse some of the trends of overload and disengagement.
 

Well, and to bring it back full circle, I know that we started this conversation talking about where AI fits into some of the hype cycles that you've seen over the past 20 years in tech. What are you most excited for, or optimistic about, in terms of where AI is going next? 

Well, I think it's just so broad in its application, right? Things like computing, the phone, they're so pervasive in our lives, and I think AI is going to be that to an even greater extent. I think this’ll probably be the biggest wave of our lifetimes. So I think that's super exciting. 

These are moments, when you have these new eras of computing, when the concrete unfreezes. Suddenly, anything's possible again. And then things harden up after a few years. So we're in the middle of that right now. All the things we're describing around having these smart AI co-pilots and self-organizing workspaces—and, you know, we haven't even really talked that much about the more exotic, like having AI coaches and mentors and confidants. AI is going to play a really weird role in our lives too. But we're prognosticating these things. They will exist. It's similar to like ‘99 or ‘98 when people were like, “Oh yeah, in the future you'll order groceries online, you'll listen to music online” and people are like, “No, that's ridiculous.” You either thought it was ridiculous and impossible or you thought it would happen tomorrow.

The same with self-driving. It takes a while. The time constants are longer than you think. But the fact is, it's like a jump ball. It's anyone's game. All these things are really just limited by our imagination and our resources—some on purpose, but some by just luck. I'm hopeful that once again, our board is sort of positioned in the right place on this wave, and it's going to be a lot of fun.

Drew, I think that's a good place to leave it. Thank you so much for joining us.

Thanks Matt.
 

~ ~ ~
 

If you want to learn more about Dropbox Dash—and try Dash for free—visit dash.ai.

Working Smarter is brought to you by Dropbox. We make AI-powered tools that help knowledge workers get things done, no matter how or where they work. 

You can listen to more episodes on Apple Podcasts, YouTube Music, Spotify, or wherever you get your podcasts. And you can also find more interviews on our website, workingsmarter.ai

This show would not be possible without the talented team at Cosmic Standard: Our producers Samiah Adams and Aja Simpson, our technical director Jacob Winik, and our executive producer Eliza Smith. 

At Dropbox, special thanks to Benjy Baptiste for production assistance and our illustrators Fanny Luor and Justin Tran. 

Our theme song was created by Doug Stuart. 

And I’m your host, Matthew Braga. Thanks for listening.

~ ~ ~
 

This transcript has been lightly edited for clarity.