This CEO is betting on AI to solve the world’s biggest problems
47:17 Share

This is Most Innovative Companies from Fast Company, where we speak to visionary founders to understand how they think, how they innovate, and what lessons they may have for you and the businesses that you run in every shape and size. I'm James Vincent, a founding partner at Foundr. So this is a story about the App Store and how to think about reimagining the world and how you go about

Making that huge leap from where something is today to where something might be. I think one of Steve's great quotes is, if you look at the world around you, you realize it's only invented by people no smarter than yourself. That's the moment when you realize you can change things. So the way that things are isn't necessarily the way that things are going to be.

but in going about making that mega change i think sometimes those things are easy to say on stage but actually really hard to implement and the example i want to give today is just you know the iphone came out it didn't have any buttons there was this wowness it actually didn't come out with the app store at the beginning and the app store came along a year or so later but i really think that was the crux of its success

when you were able to show that this supercomputer in your pocket, this Swiss army knife could solve all manner of challenges, click a button and a car comes, click a button and the travel gets booked, click a button and something gets fixed, click a button and you're connected with people, click a button and then more and more and more examples of that. There are the big leaps and then there are just the one foot in front of another. And I think it took three or four or five years. I know because we did a campaign around, there's an app for that, there's an app

for that. There's an app for just about anything on the iPhone. I think there were like 278 spots with three or four examples in each one. And we just kept showing people that it could do this and it could do that. And in the end, everyone was just like, OK, I get it. And so I think these big leaps require small steps. And my guest today, Alex at Scale, believes in the big leaps that AI can make. But he's not scared to make the one foot in front of another steps.

He rolls his sleeves up, he listens to customers, and he builds AI so it's useful. He makes it useful one example after another. And I think when you hear him speak today, you'll hear examples of things that you're just like, oh, that's how that works. And actually, scale is behind the scenes making things useful that are helping humanity progress. So I'm very excited for you to listen to my conversation today with Alex Wang from Scale.ai.

With me today is Alexander Wang, who is the founder and CEO of Scale AI. Alexander was at MIT at 19 years old and decided to create a company that has now become a $7 billion artificial intelligence company.

We worked together three or four years ago, and I have to credit you with really teaching us about AI back then. And you've done just an incredible job to shave your blushes. But you're on newsstands because you're the youngest self-made billionaire. And I'm sure you don't go running around telling people that.

But apparently someone's figured that out. So congratulations, Alex. It's great to see you. How are you? Yeah, great to see you. Thanks for having me. I'm excited to chat and reconnect. And, you know, I'll have to say that three or four years ago, you taught me about storytelling and I taught you about the boring world of AI. And you taught me how to connect that to what people cared about, which I deeply appreciated.

Well, thank you, Alex. Let me just go backwards to go forwards a little bit for the audience. I remember 10 years ago, I was at Harvard and the big topic of the week was big data. There was sort of over-promise and under-deliver back then, right? It was like all of this future, it's going to be awesome and

When I met you, Alex, I really felt like you were talking at a much more useful level. You felt that your job was to roll your sleeves up and build a company that turned raw data into usefulness. Is that a fair description? Yeah, that's exactly right. You know, the origin story was that I was, as you had mentioned, I was at MIT and I was enamored with AI and machine learning. So the fall of 2015 was when DeepMind released AlphaGo, which was

beat Lee Siddle in Go, which was sort of this incredible accomplishment. And then the spring rolled around and it was sort of, you know, it was all I could think about, all I could study. It was incredibly exciting. And I remember wanting to try to use it in my day-to-day world. So I was trying to build a camera inside my refrigerator that would tell me when my roommates were stealing my food.

And it took me, you know, I tried working on this for weeks and weeks and weeks and realized like it was going to be almost impossible because there was sort of no way to get the data to actually be able to build this thing to be successful. And I kind of like realized, you know, those maybe the first nugget of my general observation with

the world of artificial intelligence, which is there were so many people talking about this fantastical future and sort of all the risks and all the benefits and, you know, this amazing what happens if we get these like all powerful AI systems and what does that mean for the world? And then I tried to use the technology and I realized, you know, it's actually just really hard to get going. And data was this like fundamental bottleneck.

And that was maybe the first moment where I realized most of the community is going to be focused on these far out fantastical outcomes. And there need to be people who are focused on what are the foot by foot step interactions that are required to actually get artificial intelligence to actually have an impact on the world. Because you didn't have to squint to see what's

how artificial intelligence could be used for the world's most important problems. You know, if you think about climate change as one example, you know, it didn't take this sort of like really far out there view to see, oh, you could use artificial intelligence to at minimum just help us use

or energy more effectively. A lot of energy today goes to waste because of inefficiencies in the utility networks or the global energy industry. Or you look at the medical industry, you don't have to think so far in the future to think about all the potential benefits of being able to scale medical care beyond the very limited number of doctors that we have globally. And so there are these incredible problems that can be solved with a relatively direct application of the technology today, but

there were these huge barriers in what were the actual implementation steps and what were the actual like near-term problems in getting AI to solve some of these, frankly, world-changing and very impactful real-world problems. So that's what I got excited about. That's what we've dedicated Scale to is how do we enable, you know, the most ambitious organizations in the world to utilize artificial intelligence and their data to solve the most important problems and the sort of most transformational problems today.

You talk about every action is a piece of data. Everything in your business or in your personal is a piece of data and it's useless to you unless you can sort out the data so we can make sense of it. So I'd love a few case studies.

Tell me a few things that scale does. You're probably one of the biggest companies that people don't know about. So give us a few examples. Yeah, totally. And I think this is one of the big problems with AI is I think that when we hear about it, we think about the Terminator or Ex Machina or these sort of sci-fi scenarios rather than, you know, frankly, very real world problems that can be supercharged with AI today. So to walk through a few of them, I'll walk through sort of three examples of

The first is work that we did in Ukraine, actually. The second is work that we've done with some of the largest e-commerce companies. And then the third is work that we've done in the medical field. So to start out with in Ukraine, one of the most valuable assets in the Ukraine conflict has actually been satellite imagery and satellite data and using that to sort of understand what's actually going on in the region and using that to divert humanitarian resources as

as well as figure out how to best manage the conflict. So one thing that we did, and we actually open sourced a lot of this data, is we basically used satellite data and algorithms that ran artificial intelligence algorithms that ran on top of that satellite data to map out, A, map out all the major cities. So Kiev, Kharkiv, Dnipro, Mariupol,

And then we mapped out on a day-by-day basis, what is the level of damage to every single structure within these cities? So literally building by building, what is the level of damage that is caused by the war with Russia? And how can we use that information to real-time

divert humanitarian assistance resources? How can we divert medical attention? How can we divert infrastructure projects? How do we divert our resources to basically immediately address wherever there's meaningful damage on a day by day basis? We open source all this data. We worked with both the US government as well as the Ukrainian government and we're told, "Hey, this hasn't existed before.

There was no way in the past to get this level of granularity in the analysis. And through the use of artificial intelligence, we're able to work through this very, very rich data of sort of like satellite imagery,

of all of Ukraine to actually get to insight and actually get to something that's useful for coordinating response. The second example that I want to walk through is actually in e-commerce. So a lot of us buy things online. Instacart is one of our customers, for example. If you shop on Instacart, behind the scenes, scale is helping to power things.

I think all of us have had maybe the uncanny experience where we go to an e-commerce company, maybe it's Amazon, maybe it's another one that you may also like or in some sort of recommendation, there's sort of this uncanny product that you click into. They're like, oh, yeah, I actually do want that. I actually do want that. You know, they know you well enough to know what you might want to buy. And behind the scenes, that's all powered by e-commerce.

machine learning and artificial intelligence, just in a very sort of harmless and frankly friendly way. But all of that is about how do you understand the product super duper well? You know, when you're in a store, the salesperson is a really good recommender of products because they know their products in and out. As they talk to you, they start to understand what do you like, what do you not like, what are the things you're looking for, et cetera. And that's sort of the same challenge that you have in these e-commerce experiences with the machine learning algorithms they have to build.

But the main thing that we help these companies with is how do you, A, build this encyclopedic understanding of the actual products that you're selling? And what's all the data that you need to feed in to power that sort of deep understanding of the products? And then how do you build effectively sort of the digital salesperson? How do you build these algorithms that are going to be able to take in what little information you give them and guide you through the sort of buying experience and buying process?

And the last one is in medical care. And kind of, as I mentioned before, you know, globally speaking, there's a massive shortage of doctors. I think roughly it's sort of like a 10x shortage of doctors globally. And the medical universities just cannot train up enough doctors to sort of keep pace with the sort of growing global population. And so the only solution this is going to have is going to be in building algorithms that are able to

to automatically identify illnesses or other things in medical imagery and medical data so that you can start automating a lot of that to be able to enable us to scale to provide medical care to, frankly, a lot of people who don't have that medical care. So we did a bunch of research with,

MIT Media Lab as well as Stanford in analyzing dermatological data and basically using that to help identify skin conditions automatically using algorithms. So they're fairly different, each one of them, which I think shows the power of the technology because it can be applied in almost every field. There's some really interesting and transformational use case of

artificial intelligence, machine learning. But none of these use cases are a big bad AI that's going to take over the world. They're all, frankly, pretty sort of near future use cases. Yeah, no, it's funny. 300 years ago in the industrial revolution, these big machines arrived. But wait a second, that's what humans are doing. Why are machines doing it now?

But in the long term, the productivity gains have led to the last 300 years of prosperity. Is that how you think about being able to lift humans into more ability to progress? Yeah, 100%. And I think that like,

One of the major misconceptions about artificial intelligence today, and it's a very small nuance, but I'll explain why it's so important, is that the modern AI systems are intelligence machines, not learning machines. What I mean by that is, if you think about each of these

sort of AI systems. And I think now many of us have probably seen fantastical things on the internet done by Dolly 2 or some of these image generation AI or GPT-3 and some of these text generation AIs. If you look at all of them, it's true they've gotten to a certain level of intelligence that makes it so that they can do some pretty impressive things. But they actually hit a sort of cap in getting there. It's very hard for these systems to get beyond their limitations today.

And if you take that in contrast with like a baby or a child, you know, a baby or a child, you know, they're actually learning machines. They don't start out knowing really anything. But with this incredible velocity and speed, they start learning more and more about the world around them such that they get to a point where all of a sudden they can reason about not almost anything.

this subtle but very key difference is really important because if you look at how the algorithms of today work, they do not learn particularly quickly. In fact, to get them to learn, we throw the biggest, biggest data centers and just massive amounts of energy at the problem to get them to learn at this relatively slow rate, actually. And if you compare that to the efficiency of a child, it's just like

Many, many, many orders of magnitude different. And the reason I think this is particularly relevant to the Industrial Revolution example is that in the Industrial Revolution, we didn't build general purpose machines. We built very specific machines that it's true, a machine that assembles a car is expensive.

a lot faster and better at assembling that car than any person is. But as a person, you can do all sorts of other things that that machine is never going to be able to do. The same is true with the artificial intelligence systems of today, which is, yes, they're going to get to a point where they can do very, very interesting and incredible things, but they're not going to be able to do all the things that humans can do. And so the way to think about that is like, how do we harness

that ability and how do you harness these sort of intelligence machines to then enable us to get to the next level of creativity and the next level of

enabling humanity to accomplish more. One great example of this that I think is really in the near future is around actually content production and sort of thinking about making a movie. Right now, sure, there's some movies that are made low budget and low budget in the movies content is still hundreds of thousands of dollars. It's still quite a bit of money or low digit millions of dollars. It's still quite a lot of money to make a movie. And then there's all the way up to sort of the big Marvel movies that cost literally hundreds of millions of dollars to produce.

And that makes making movies very inaccessible to everyone in the world and actually really limits, I would argue, limits the creativity of the movie industry because it's so expensive to make a movie. Only very particular movies that sort of like studio execs like end up getting made.

If you look at the AI technologies that are being made today, it's really in the near future, we're going to be able to make movies that are just as visually compelling as the movies of today for a small fraction of the cost. Because we're going to have good AI that can automate visual effects, that can automate a lot of the editing, automate a lot of the special effects, and really enable people to be far more creative in the process of building, of making these movies. And that's going to be this

huge, huge enabler for unleashing the creativity of the global population. You know, now a very small percentage of people in the world have any access to the process or have any hand to play in the process of making a movie. But what if it could be the case that like, you know, almost anyone in the world could, with the help of AI, make movies? Wouldn't that be just this like incredible world that enables sort of this onrush of creativity?

to your point alex that and i love this sort of differentiation because i think everyone's going to understand it between sort of intelligence machines are really no match for humans right which is kind of like oh my god yeah but they can compute so much faster than me but computing something is just one element right so in your example i'm going to go back to your movies thing it's not the absence of humans you're not going to hit an algorithm it's going to make a movie

But the creativity kind of goes up to being able to think about how to put all those pieces together and storytell and how and where to use that AI rather than all of the hard work that it would create and the millions of dollars it tends to create. And therefore, we only get certain types of movies getting made, right? And going in the movie theaters.

So it changes the dimension of the input for humans, right? Which is back to the Industrial Revolution. It wasn't that people stopped working, right? There were other jobs to be done. And maybe they worked in the factories and they worked the machines and then they slowly moved up or they moved off to do other things. And, you know, I don't know. We're at this moment now of like massive full employment, right? And everybody's been worrying about this moment would be all about Wally and, right?

If you remember the movie where humans had to leave the planet and we're all sitting overweight at 500 pounds on a movable machine slurping slurpees because there's just nothing for humans to do. This is not the dystopian version of the future that you're describing because what I'm hearing from you is

and not to diminish it, super important, but quite rote, intelligence, repetitious stuff that we should give to machines so that humans can progress, move forward. You know, someone told me, like, we were working with a company that was kind of figuring out that every third fish was sort of in the wrong place and got thrown away, right? Why? Because they haven't got machine learning to figure out where the fish need to be at what time and stored here and be over there and right.

They're all in the wrong place. A third of the fish are thrown away. If you could just apply machine learning to fish, to the supply chain, to every company, you could maybe be a third more efficient because, and maybe it's higher than that. And so that pushes humans up to do greater and better things. Is that right? Totally. Yeah, yeah. There's two examples that just in you talking about it really struck me. This example that you mentioned about, yeah, every third fish goes to waste. That is happening in microcosms across...

across the entire US economy. Most businesses in the world sell physical things. And any business that sells a physical thing, there's this immense problem of like, do you know who's going to want to buy that physical thing, where, and how do you make sure that you get your product in the right spot to enable that as efficiently as possible? And that's just this massive, massive problem

for the world. If you look at how Amazon, who's sort of at the forefront of solving all of these problems, if you look at how they did it, they had this massive project around how do you build accurate forecasting technology? You can have a very good sense of when are people going to want this particular thing and how do you make sure you get the product in position so when the person clicks the button,

the thing is like as close as possible to their doorstep. Delivered within two hours, which is always a, seems like a piece of human magic, even today. Exactly. Getting to that point, it actually is like, we can build algorithms, we can build systems that get you to that level of efficacy. It needs to be rolled out throughout the,

entire economy. That's kind of one. And then two, to make this sort of like, what do I mean by these planes of creativity, right? And what does that mean in practice? I'll give a really concrete example, going back to the sort of movie studio concept. So if you think about Kevin Feige and the Marvel movies, one of the ways in which the Marvel movies were allowed to be creative

in a way that is not really present in any other sort of movie universe, is that because they were making so many movies, there was this element of creativity behind the scenes, which is like, how do all the characters in all the movies tie together over like a 20 movie arc, right? Mm-hmm.

And that's one of the most rewarding parts of watching the Marvel movies is that there's sort of this very tightly woven sort of tapestry between all the different storylines and arcs and they come together and they go apart and come together, go apart. And literally only Marvel is allowed to sort of like think that big because it turns out it costs billions and billions and billions of dollars to make 20 movies that are that effective. Well, if instead, because of new AI technologies, all of a sudden it costs, you know, one one hundredth of that to make

create the same level and quality of cinematic experience, then all of a sudden many of us can become sort of like kind of Kevin Feige like masters behind the scenes and like carefully architecting these broad story arcs because all of a sudden it's sort of, it's within reach for more and more creatives all around the world. And I think that that's an incredibly powerful thing is like how do we unleash the creative ceiling of people and humans as effectively as possible through AI supercharged creative tools?

And that's a world I'm really excited about. When I met you, I'm guessing you were 22 and you were going about building a company. And as you, if you don't mind the pun, scaled scale, you have some very strong views about how you get big without getting bad. And what are your beliefs of maintaining innovation as you grow?

So yeah, now we're about 700 people. Wow. Yeah, it's a lot of people. Yeah. There's companies that are way bigger than us as well. I think Amazon has over a million people. But now that we're 700 people, I love the way you framed it, getting big without getting bad. This has been something that I've always really thought about. You know, one of the things, one of the facts that is really shocking is that the initial iPod at Apple sort of took, I think it was like 6%.

six months to release or thereabouts. You think about it and you're like, how is that possible that you go from the idea of building a music device to getting that onto store shelves within six months? That's just astronomically insane.

And I remember in the early days of scale, sort of the electricity in the room and the sort of the energy in the room of the startup where it's sort of like everyone is so focused on just building something incredible. Everybody is dreaming big. Everybody is like...

thinking about how do we make something really impactful, really amazing, there's sort of this electricity. One of the analogies that I like to use is like, it's almost like you're reaching into a nuclear reactor. And it's sort of like, it's this strange otherworldly energy that empowers everything about the building. And you don't really know exactly quite what it is, but you know that it's there and it's very powerful.

So that's something I've been obsessed about. And I've written a little bit about this and I have a bunch of beliefs on this. The number one thing is to hire people who give a shit. You know, it's a simple thing, but I think that like an unfortunate fact about most big companies or mildly big companies is that you look around and people don't even seem to care.

about what the company's doing anymore. And it's like, it's not even clear if they ever cared about the work that they did. And this isn't a judgmental claim, but I think it's impossible to build something nearly impossible, or it's impossible to make magic if nobody around you cares about what they're doing. And I think that this problem arises because

There's a subtle change that happens at a lot of companies, which is when you're small, the only people you'd be able to get your company are people who give a shit. Because frankly, you're this small little company. Nobody's going to care about you unless they happen to really, really care. And maybe they're slightly crazy in the process, but they really, really care. That's the only way you attract people. It works up until certain size. Then you get to be bigger, successful, more notable company. And a

And then I think what ends up happening is the recruiting team turns into almost like a college admissions office. And it turns into sort of this, you know, we have all these applicants coming in. Let's select the most impressive, diverse and sort of like interesting group of people coming in rather than who lives and dies for the mission and who really, really cares. And this small difference of sort of like, how do you bring on sort of missionaries versus mercenaries is a really, really important

important thing for companies to get right. And ultimately, you know, the company is made up of the people that you bring into your company, your foundations are your people. And so it's really important that you get that right. And you have to hire people who just

are fervent. A, they're just the kind of people who just like care so much about what they do and care so much about their work and care so much about having a dent on the universe. And you get people who are so passionate about what your company is doing that it's sort of like it naturally becomes the sort of inspiring thing that drives their life. I've heard you describe that as optimism or attitude-driven.

defines reality? Totally. Yeah, yeah. So we have this belief, optimism shapes reality. We also call it ambition shapes reality. I think we all know this in certain ways. So one example is just this funny phenomenon where people end up more or less accomplishing the magnitude of their optimism or the magnitude of their ambition. And it's a really funny thing. It's a really funny phenomenon. I'll give a few examples of it happening in a totally different area, and then I'll sort of like apply it to building great products.

If you look at a histogram of marathon running time, so how long it's taken every marathon runner ever to run a marathon, you might expect it to just be this sort of like normal distribution, this sort of Gaussian curve. You know, some people, it's sort of like it's a very smooth curve, but it's not like that at all. It's actually like this very jagged line where basically at every hour boundary, there's a lot of people who try to get right under that hour boundary.

So under four hours, a bunch of people try to get under four hours. Under four and a half hours, a lot of people try to get under four and a half hours. A lot of people try to get under five hours, five and a half hours, et cetera. So you notice it's this very jagged line where people more or less really accomplish sort of whatever they set their mind to. It's this really weird thing. Another one that's kind of interesting is when we broke the four-minute mile,

It was, it was hypothesis for a long time that nobody could ever break the four minute mile and everybody struggled with it. And then as soon as somebody broke it, all of a sudden, everybody, not everyone, but like, you know, all the people who cared enough, a lot of people ended up breaking the four minute mile. Wow.

And it's just one of these things that sort of like humans are incredible. And by the way, start conscious to AI, but we'll get to that later, I guess. Humans are incredible at sort of accomplishing the sort of scale of their ambition. And when you think about that in an organization, it's so important that you create a culture where people dream really big.

you know, where people have incredible optimism for what you're going to accomplish. Because if they're not thinking big, then you sort of have no hope. You know, another way to put it is that thinking small is a self-fulfilling prophecy. If you think small, then maybe you accomplish those small things, but you're never going to accomplish the big things thinking small.

And so I think it's important that like the nuclear reactor analogy is actually so great for what a company should feel like, because it needs to feel like this almost beehive where, where people are dreaming big, people are really excited. People care a lot. And people are also really good. People are really competent and they like,

there's a really high density of just like incredible people. And if you can accomplish that, then it's just like insane what an organization can do. I think Apple, again, not to keep going back to it, but it's such an incredible example of this where very, very small teams built the products

that shaped the future of humanity. And it's incredibly disproportionate. You know, maybe you would expect that it would take tens of thousands of people to build the first iPhone or maybe hundreds of thousands of people to build the first iPhone, given how important that was going to be for, you know, literally the future of humanity. But no.

is a pretty small team and it came from a willingness to dream about the future of computing. And I think that that kind of hiring people to give a shit, letting ambition shape reality, building the future, you know, these are so important for actually having a shot at

at building the platforms of the future. Yeah. It's fascinating. You mentioned Apple. I think Steve's pretty famous for saying, expect great things from great people. So I'm just trying to imagine, you know, a fast company listener and what can I take from this? I'm, I'm not in a stop. I'm in a category that could be innovating. How should I think about that? What kind of instruction might you give to someone that's buried in a company that isn't pushing innovation, but how might I get that going? I have a few thoughts on this. Um,

It's a really good prompt. One of the things that I tell my team a bunch too is there's nothing more seductive than the invitation to think bigger. And what I mean by that is like, let's say you're anyone in any organization, you're meeting with anyone, give them the roadmap as to how they could be thinking way, way bigger. I'll give an example, which is imagine we're talking to

a large agriculture company. It's this critically important piece of infrastructure for the world. And you're sort of like, how do we think bigger? And it's like the global challenge is that we're going to run into a food crisis or a food shortage. There are so many in the global population is not going to stop growing. And

We need food to be able to feed everyone in the world. And frankly, we need to get way, way more efficient with how we produce food. If you take that problem and you think about it, it's like, how do we use, you know, satellite data and better technology and better robotics and, you know, all these better technologies to actually enable us to meet that, you know, impending challenge that's going to face, frankly, all of humanity over the next 20, 30, 40 years. And just

Framing in that way and creating this sort of like permission to think big about the problem, you're going to be impressed and shocked and frankly surprised by the level of innovation and the level of sort of ingenuity that comes in sort of thinking about the problem in these like very big experiences.

expansive ways. The other thing I was going to mention was that, you know, I think people think about their careers in the wrong way, which is, I think a lot of people think about their careers in this form of like, how do I build the best resume possible? Like, if you think about the output of your career as your resume, that's a little backwards, isn't it? Because it

It's not about what company to work at for how long and then how did you get promoted over time? Then what other company did you jump to and what other, what company to jump to next? That's the wrong thing. The question is, what is the thing? What are the things you actually built? What are the things that you actually contributed to the world? What are the, what are the products that you helped will into existence that would not have existed if you weren't there now?

like in the trenches, driving the product, driving the vision, driving execution and making magic happen. I think because it's so easy to look at someone's resume, we totally ignore this fact around like what exists in the world because you worked on it. When you're at the end of your career and you're looking back, are you happy with your life's work? It struck me when I sort of like read about it in the biography about Steve Jobs, sort of like our promise to the people who work at Apple is that you have the opportunity to do

your life's work. It's such a powerful idea because the really, really sad fact is that most people do not end up leaving a dent in the universe. Most people do not end up doing work that has the ability to have a dent. And part of it is because they're not even thinking about it in those terms. They're not thinking about how do I make sure that my career, the product of my career are products and innovations or changes that, that possibly benefit the world. You think about it in the form of resume.

Anything that's really unfortunate. Yeah. Inspired leadership can set a tone for people, even in more mundane occupations or jobs, seemingly. One of the ways I like to think about that at Founder is maintaining the conditions for innovation. Johnny Ives' favorite quote, ideas are fragile. It matters who's in the room as to whether they live or die. And so you have to create the conditions for innovation. It doesn't just happen.

How do you raise the bar? I think Einstein's definition of madness was doing the same thing over and over again and expecting a different result, right? There must be a lesson here. What do you think about that? Yeah, so I think there's a few pieces to this. I think that the first thing is ensuring that everyone sort of has the right mindset, which is about building upside or building something amazing rather than maintaining. You know, I've actually noticed versions of this problem as we've grown.

You know, when you're a startup, you have no choice. You know, you have to build something from nothing. So you're only going to be thinking about building. But then if you're as you get to a large and larger company and you have a pretty successful core business or a pretty successful product, there's a massive temptation to just think about incrementally maintaining or incrementally improving that core instead of also thinking big about what are other ideas that you could do.

One part is that, and it's really easy to get stuck in the mindset, which is like, why would we think about that small little thing over there that's never going to be as big as the big core of what we're doing? You know, that's the process of invention. Invention is considering and thinking about all the potential opportunities and working on those things. So that's one major part of it. I think another part of it is to your point around sort of like the fragility of ideas in certain rooms.

It's so difficult to explain because it's almost like smoke in the wind or whatnot. It's so fragile and so chaotic. But there are definitely, and I know it's even for myself, there's certain people where when you bounce ideas with them or in a small group of people, you notice that there's sort of this reverberation.

and there's this resonance that happens where it's sort of like, I have this idea and then someone adds something to it and someone adds something to it and someone else has this crazy idea and sort of it builds upon itself like a snowball rolling down a hill. And that process of this building of energy and idea and sort of like all that stuff is one of the most powerful forces in the world, frankly.

That only happens if you get the preconditions just right. People feel safe. People feel excited. You know, people are excited about what they're doing. You know it when you see it. That's as best as you can do. So you sort of need to like keep trying to find it. And then when you find it, never let go of it. That's great advice.

I'm going to zoom out and bring us home here today. And obviously felt great when you told us that even very intelligent machines are no match for humans. So let's just imagine the future and the dent that you might make in the universe. Five, 10 years from now, what world do you think we can build with human progress, with AI at full pace, or at least doing as much as you think you can do at scale?

Our whole goal is how do we make AI a reality and how do we make AI actually drive progress against the world's biggest problems? And so to explain in greater detail, I think five to 10 years from now, we're not going to have these sort of terminator like AI that take over jobs and take over the world and sort of and do all these horrible things. But I think what we are going to have is

we're going to have made meaningful progress and actually have started to get a handle on the world's biggest problems, whether that's the food crisis, like we talked about before, or climate change and sort of sustainability of the global world, or whether that's scaling the medical care, like we talked about,

or it's enabling geopolitical stability and enabling geopolitical peace, or it's doing simple things to help us all find what we're looking for on the internet more easily. Basically, driving forward on all of these major world problems

and actually starting to drive action and technology that makes a difference, rather than, I think, the unfortunate state that humanity has been in for the past few decades, where we all know that these problems are there. We all know that they're big world problems, but it's been really hard to make progress on them.

I'm a big optimist. I think that AI is a technology that actually gives us a really meaningful chance against all the world's biggest problems. And I think then the next five to 10 years, we're going to be able to prove that out and drive a lot of that change and solve many of the sort of problems that plague humanity today.

I think AI is going to be the conduit to accomplish that. And that's what we're passionate about at scale. Yeah, no, that's great. And it's almost like we can't see the wood for the trees because we're in the middle of the forest. But when you zoom out, you see that actually there's a solution because you can see the patterns. And what I'm hearing from you is a belief that technology can catch up

with the world problems today and can solve many of the ones that honestly I think most people are worrying about, you know, the horrific weather around the world. And can we solve the energy crisis and carbon emissions? I get great solace when I speak with someone like you that believes, and I'm a true believer also, that there is incredible optimism in innovation and innovation founded on technology that frankly just didn't exist 10 years ago.

And to hear you say that humans progress through the advancement of AI and don't regress through the advancement of AI is incredibly optimistic and satisfying to hear someone with your deep domain expertise make a statement like that. So Alex, I just want to thank you so much for today. It's been terrific to see you again and hopefully it won't be another three years before I do.

Thanks for coming on the podcast and taking the time. It's been awesome. Yeah, it's been so wonderful. We covered so many interesting topics. And thanks again for having me. So that was a terrific conversation with Alex Wang from Scale.ai, the 19-year-old. I guess he didn't finish MIT. I don't know. I think he was too busy building scale. But, you know, he is a billionaire that doesn't follow the bro type approach.

which having watched recrashed recently was relieved to meet someone doing incredibly well, but being incredibly thoughtful, a great conscience about the role of AI. I really thought the audience will understand and agree and learn from the perspective that no matter how intelligent machines are, they are no match for humans.

And I thought that was terrific that actually his approach to AI is to take one menial task and rote task at a time to take waste out of systems, to make things happen quicker, to take friction out. And that just slowly reduced the inefficiencies, increased the ability to do things. But that there is this line where intelligent machines aren't thinking machines and that humans are thinkers.

And so that dystopian view of the future where everybody becomes a cyborg was somewhat dismissed. And I found that incredibly uplifting and optimistic. Alex also had some incredible things to share about innovation. So here's a guy, he started a company when he was 19. And here they are today, 700 people and a $7.6 billion company. And Alex had some really interesting tips about how to keep innovation going, how to get big,

without getting bad. He talked a lot also about optimism shapes reality, which I thought was fascinating that, you know, you set a high bar, you get people excited and you create an environment, create conditions where people just are expected to innovate and are rewarded for innovating. And of course, you know, again, that's still easier to say even in a 700 person startup than it is in a very, very large company. But

I think some of our conversation was just around an approach, an attitude that would make you see, well, if the same thing happens three or four times, then maybe there's a problem there and we should go fix it. And that just general sense that either as humans or as AI, things can get better and better. And I think the end was just so uplifting in terms of his belief that in the next five to 10 years, a dent in the universe that AI can make was really that we

we can solve some of the world's largest problems, whether it was an ability to identify disease through understanding masses amount of data and labeling it and then getting an algorithm that figures out. And so we can conquer these massive things, sustainability, carbon outputs. And so for the guy who's done so tremendously well, convincing one customer after another, tackling one customer after another, and his story about the information he gets

in the Ukraine looking at bombed out cities and then

figuring out where resources, human resources were needed because of where the damage had been was just an incredible illustration. No one had ever done that before. And clearly the US government approached Alex at scale and said, can you solve this? And indeed he went off and did so. So yeah, I thought on multiple levels, Alex blew away some myths about what it takes to be a billionaire, which I thought was lovely.

and also just gave us some really good codes for innovation and some optimism for the future.

All right, that's all for this episode. If you're a new listener, be sure to subscribe to Most Innovative Companies wherever you listen. And if you like this episode, please leave us a rating and a review on Apple Podcasts. I know we also want to hear from you. So let us know what you'd like to hear more of. Send us an email at podcasts at fastcompany.com or tweet us at hashtag most innovative companies.

Most Innovative Companies is a production of Fast Company in partnership with founder FNDR. We couldn't afford the vowels. Our executive producer is Joshua Christensen. Our sound design is Nicholas Torres. Writing is Matias Sanchez. Alex Webster and Nikki Checkley helped with the production. This podcast was done in collaboration with my wonderful partners at founder, Stephen Butler, Becca Jeffries, and Nick Barham.