Lucy Guo: The woman training AI

0:05It's a warm spring day in San Francisco 2014. A 19-year-old student exhausted but buzzing with adrenaline steps outside an office building looking for a quiet place to make a call. She's just been accepted into the Teal Fellowship, a golden ticket for tech entrepreneurs, but there's a catch. She has to drop out of college. She calls her parents expecting cheers, but she gets a blunt verdict instead. You're an idiot. For her parents who had sacrificed everything, quitting school feels like a slap in the face. She knows what they'll say, no more health insurance, no more phone bills paid, no more financial lifeline.

0:39But she bets on herself anyway, a gamble that will one day lead to AI billions. Welcome to Good Bad Billionaire from the BBC World Service. Each episode, we pick a billionaire and we find out how they made their money. We take them from zero to their first million and then from a million onto a billion. My name is Zing Zing and I'm a journalist, author and podcaster. And I'm Simon Jack. I'm the BBC's business editor. And today we're talking about Lucy Guo, who in 2025 was named by Forbes as the youngest self-made female billionaire at just 30 years old. Although she's recently been overtaken by a younger billionaire.

1:12She made a fortune from her stake in Scale AI. This is an artificial intelligence company she co-founded when she was just 21. Now, Scale AI might not be a household name, but it powers giants like OpenAI's ChatGPT, Google, Microsoft, even General Motors' self-driving cars. And today she is worth$1.3 billion, but you wouldn't know it from her lifestyle. She skateboards to work, she still hunts for buy one, get one free deals, and her motto is act broke, stay rich. I act broke, but I'm not rich. But for now, let's scale back to zero and debug, if you like, how Lucy Guo coded her way to that first million.

1:53Lucy Guo was born in the Bay Area of California, Northern California, on October 14th, 1994. Her parents were both electrical engineers who had emigrated to the US from China. She said her parents worked hard to climb from middle class to upper middle class, but both were laid off from their jobs briefly when she was young. So despite their success, they lived below their means often and instilled this into Lucy and her younger brother. Sometimes this frugality was too much. Guo said she was bullied at school for wearing plain, inexpensive clothes instead of trendy branded clothing, yet it did give her the drive to make money.

2:27She told the BBC, I basically became obsessed as a little kid. I was like, how do I make as much money as possible? She's not the first billionaire to become obsessed with money when she was a little kid. Like many of our entrepreneurs, she mowed lawns, she sold things to her classmates. This included coloured pencils and Pokemon cards. Must have made her very popular in school. Sometimes this got her sent to the principal's office. She'd hide the cash she made in her room, but it would often disappear. Her parents knew that the best way to discipline her was to take away her money. Now, Guo, understandably, wanted to keep that cash.

2:59So she went to a local Home Depot store and managed to get a Visa debit card to set up her own PayPal account. This was, according to Guo, when she was just eight or nine years old. I really struggle to think of an eight or nine year old right now who would know what PayPal even is. Or getting a Visa debit card. I mean, having said that, she has repeated this kind of origin story for her wealth over time. And I think in one case, she says that she claims she sold the Pokemon cards in kindergarten. Right. which would make her a very mathematically advanced child. Let's have a look at her early background.

3:30Her parents were pushing academic excellence. Her brother skipped grades, accelerated through school. Lucy was signed up for an abacus competition. And she told CNBC, because I wasn't really cool or allowed to be social, I pretty much spent all my time on computers growing up. The only fun I would have was on a computer. At one point, Lucy's parents put a key logger on her computer. And this is a programme that kind of tracks your keystrokes, so you know what people are up to on their computer. but that motivated Lucy to learn more about computing. At age nine, she worked out how to get the key logger off her computer.

4:02Soon, she started playing online games such as Neopets, which is where users can care for virtual pets. It was really big at the time. I was an old Neopets fan. Really? Oh, yeah. And this, I think, is actually quite genius. Is this like the digital version of Tamagotchi or whatever? Exactly. But the difference is you're playing with other people in an ecosystem with money or Neopoints, as it's called, the in-game currency. So sometimes you can get these people who are trading neopoints for actual money on third-party sites. So basically what Lucy did was teach herself to code these bots to automate the tasks and games so you could earn more neopoints.

4:37And then she would rack up lots of these points and then sell the account sometimes for thousands of dollars. And I must admit, this blew my mind because I was dutifully playing those games like a sucker and not making any money from it. And here comes a billionaire who makes thousands from it. I mean, it really does kind of stagger you. Yeah. Well, in high school, she continued to work on those coding skills. She learnt Photoshop so she could build websites to host adverts which could earn money through Google AdSense. She told Business Insider, I remember I go to Fry's Electronics, a kind of high street store, and just sit and read books there all day on how to do HTML and CSS.

5:13These are coding languages. And by the end of high school, she claims she was earning nearly five figures. What is that? $10 ,000? Yeah, easy. She also says she made and lost her first$1 million in high school. In an interview with Fortune, Lucy recalled how she bought around$5 ,000 worth of Bitcoin during high school from money she'd made on the internet. The Bitcoin she bought, she said, was theoretically worth several million, but she left it on the exchange and never put it in her wallet. So she didn't actually make that first million. Do you know what? The internet is replete with stories about people who forgot to take, you know, lost a phone with Bitcoin on it.

5:48But, you know, in the early days, there was one guy who spent like 100 Bitcoin on a couple of pizzas, which would be worth hundreds of millions now. Lily Allen, the singer, famously turned down payment for a private gig in the form of Bitcoin. Right. And funnily enough, my brothers tell me that the pizza day comes around every year and all the Bitcoin bros celebrate this because that was basically the world's most expensive ever pizza. Well, I hope it was worth it. Let's look at where Guo was at the time. She went to high school in the late 2000s, early 2010s in Fremont, which is across the bay from Silicon Valley.

6:22Sometimes referred to this area as the hardware side of the bay, where the factories, companies including Apple and Tesla, who actually make sort of, you know, things rather than software are based. Bitcoin was invented in 2008 and it first traded for cash in 2010. Probably didn't hit the mainstream until a bit later. But, you know, growing up in a tech literate community like Fremont meant she could well have been an early adopter. Now, worth noting, you don't need to be a coder to actually get Bitcoin, but her coding skills would have made it easier to mine, to secure, to experiment with the cryptocurrency.

6:53In 2011, it was ranging in value between$5 to$30. So depending on the price when she purchased it, five grand would be worth between tens of millions to$100 million right now. Yeah, I think a lot of people do their daydreaming. If only I'd put$5 ,000 in Bitcoin in 2010. If only I hadn't paid for a pepperoni pizza in 2010 with Bitcoin. So, you know, she's on the ground. She's around the action in Fremont. And in 2012, Groh started a double major in computer sciences and human computer interactions. Interesting course at Carnegie Mellon, one of the top schools for computer science in America. And whilst there, she started competing.

7:34And we've come across this word before, hackathons. And these are competitions where computer programmers try to code the best solution, often against the clock to a specific problem. They can last for a day or 48 hours. Prizes range from iPads they give away and some cash prizes as well, like$35 ,000 in some cases. Gwors said she won a lot of iPads, but hackathons were also an introduction to the tech startup world, which was much more appealing than just straight coding. She blogged in 2017, I got to experience the whole spectrum, product, coding and design. It's where I realised code was my tool to unleash creativity.

8:10but I loved product the most. And her hackathon prowess also piqued the interest of the Peter Thiel Fellowship team. Now, Peter Thiel, one of our billionaires, you can go back and listen to our episode on him. The Thiel Fellowship was founded in 2011. It's a two-year mentoring scheme. And back then, it was awarding 20 people under 22 years of age,$100 ,000 scholarship, which is now worth 200 grand, to build their own business. But there's a very important and interesting catch to this. fellows could only be accepted into the programme if they agreed to drop out of university. So in 2014, aged 19, Guo did just that.

8:47Dropped out of university and joined the programme. She was selected from a pool of applicants from 44 different countries, but that came with a family cost. Oh yes, definitely. Now dropping out of university caused a huge rift with her parents. They wanted her to pursue traditional education and this is exactly where we found her at the start of this episode. These days, Lucy says she understands that her choice was taken as a sign of disrespect from her parents. She told CNBC in 2025, I think my parents viewed that as a sign that I didn't love them. And they weren't very happy with it when it was just me making a bet on myself and choosing to optimise for what I thought would be a better future for myself.

9:27I mean, how many of our internet and tech billionaires dropped out of university? I would say at least half. More than half. I would say the majority. It's almost kind of a requirement to drop out. Gates, Elon Musk, you know, it's a long list. Zuckerberg. Zuckerberg. I mean, it's almost like if you don't drop out of university to pursue a tech dream, you're sort of several years behind. Well, also, I think that the community sees you as an establishment, cautious person. And that's not the qualities they look for. They want someone who's, in a way, a bit punk about stuff. A little bit of an edgy radical.

10:03Edgy radical, two fingers up to the establishment. I don't need to go through these hoops that everyone else goes through. I don't need to get my diploma or graduate or shake anyone's hand to do my own thing. So Guo had some business ideas, but some of them were pretty short-lived. She had an app called Nomit, which was a homemade food delivery app, a bit like DoorDash or Deliveroo, but homemade. She quickly realised there would be a litany of legal issues, including liability in case she poisons anyone. But that failure made her realise she still had a lot to learn. So she sought out work experience at some successful startups.

10:36She'd already been a summer intern on Facebook's iOS Messenger team and at Endless Mobile while at college. But her first full-time job was as a product designer at Quora. Now, this is a question and answer discussion platform founded by a Facebook alumnus. Less than a year later, she took a job at Snapchat, which is founded by another of our billionaires, Evan Spiegel. And at Snapchat, she was the first ever female product designer. and she worked on projects like Localised Stories, which is kind of similar to Snap Maps, where you can see where your followers are. Right, OK. So she's getting a good smattering of experience.

11:08And in 2016, after two years as an employee, she quits her day job to work on ideas for a new start-up with someone called Alexander Wang, a maths prodigy she'd met when they were both working at Quora. When they started, she was 21, he was 19, and, of course, he quit his degree at MIT. Now, their first idea was a subscription service for access to nightclubs, which Groh said was a terrible idea. Her words. The people who went to nightclubs regularly knew how to get into the best ones already. So within weeks, they moved on to a health care app called Ava that would match patients to doctors. And that was inspired by Lucy's own experiences after breaking her jaw in an electric skateboard accident.

11:49Of course, how very Californian. Something I find so interesting about tech billionaires is that they aren't really guided by specific interest. You know what I mean? You know, she's just interested in setting up businesses. Well, yeah, she's just kind of trying anything, just sort of sticking her finger in the air and see which way the wind's blowing and see whether, I suppose, you know, you throw enough mud at a wall and you hope that some of it sticks. Well, Guo and Wang used that healthcare app idea to apply for a startup accelerator called Y Combinator for the 2016 summer intake. Now, Y Combinator.

12:17We've come across that before, haven't we? I mean, it's kind of huge in the tech world. You know, it's run by another one of our billionaires, Sam Altman, who's now the CEO of OpenAI. And it's kind of like getting into the ox bridge of startup accelerator funds, right? It's like sort of the finishing school. Basically, you get picked like that's a good idea. Here's some money. Here's some expertise. Here are some ideas. I suppose what they would call it an accelerator, wouldn't they? It's kind of a melting pot. But importantly, it's a melting pot that also comes with cash to start the businesses and concede them.

12:50So it's not just theoretical. There's real hard money available for those people who, you know, rise to the top of that programme. Interestingly, after they joined the scheme, Guo and Wang realised their healthcare app was actually flawed. They wouldn't have enough return customers. Elderly patients would struggle to use it. So they had to leave that one behind as well. OK, so while sleeping on air mattresses in the Y Combinator shared house in Silicon Valley, I can see this picture quite clearly in my mind, they searched for some new ideas and our roommate suggested API for humans. Now, API is an important acronym in the tech world.

13:23It means application programming interface, and it basically lets different software applications talk to each other. And the idea of an API for humans meant treating human tasks as if they could be requested and executed, like, if you like, software commands. Some would see that as dehumanizing. So, for example, you could ask for a real-world task, like identifying something in an image, through a simple software command. and a human on the other end were completed as if they were part of the code, as if they were part of the programme. So in a way, it's a bit like thinking, you know, when TV people used to think there were little people inside of it.

14:01It's basically you're putting questions into a computer, thinking the computer's answering, but in fact you've got little humans beavering away trying to help with the answer to this request. So in a way, it's quite non-digital in its own way. Yeah, it's sort of like what producers do on radio, really. You know, when the DJ is like, well, play the next song. It's not a little magic spell to make the next song come on. It's actually a producer doing the hard work of pressing the buttons like ours does. OK, so the line between computers and humans really quite blurred in this example, which I find quite interesting.

14:31Exactly. Because on demand human work wasn't new? No, it wasn't. So companies like Google, Facebook and Quora, where both Wang and Guo had worked, already used large in-house teams of workers for tasks like content moderation, you know, like scanning through what people have been uploading to take down things that violate the terms and conditions. And there was also Mechanical Turk, which is Amazon's marketplace for hiring people to do jobs that computers struggle with, like writing product descriptions or identifying images. So Guo and Wang could see that more companies than just these online giants were using AI for things like chatbots, and they needed humans to train these chatbots and AI programs.

15:08Got it. OK. So in June 2016, Wang and Guo launched something called Scale AI. In the first few days, Guo was responding to some of the API requests herself. But the Wall Street Journal reported that she recruited some of Scale's first contractors by joining remote work groups in the Philippines on Facebook by sharing a quiz that she created. And Scale soon recruited hundreds of contractors this way through these online chat groups, mostly, as I say, in the Philippines. We don't know why Guo chose the Philippines, but it was probably because by 2016, it was already one of the world's major destinations for outsourced digital work for tech companies.

15:46Now, this is thanks largely to its big population of English speaking, internet savvy workers. That makes recruitment of data labelling freelancers scalable so you can hire people en masse and also relatively inexpensive compared to labour markets in the West. She also probably picked Facebook groups because by that time, many Filipinos were using Facebook as their primary online platform for connecting with others and for information and job searching. It's a bit like how Europe used and the UK in particular used India as an outsourcing. Again, quite a computer literate, English speaking, well educated workforce who were prepared to work for less.

16:23And with the UK and India, there was a historical relationship as there was between the Philippines and the US. so to get clients Guo and Wang mass emailed every company involved with Y Combinator which at that time was 1400 companies in that little ecosystem Scale AI claimed they were different to existing competitor Mechanical Turk, the Amazon product as they vetted their workers and they also spotted the need for labelling footage for self-driving cars, so identifying vehicles whether it's a pedestrian, a cyclist a cyclist pushing a bike, a cyclist on a bike Like these little distinctions that computers found it hard to do, they would use a human to be able to tell the one thing from the other.

17:03And as you say, train the AI to notice that difference. So, you know, Uber Technologies, for example, became a client. They added Procter & Gamble, Google's holding company, Alphabet. So they're getting some blue chip customers. What I find interesting about the great AI revolution or whatever you want to call it is that a lot of people, I think, you know, regular people like, you know, my mum, for instance, uses ChatGPT. might think of AI as just a sort of magical spell that, you know, the computer just knows exactly what's going on and what to do. But actually, it's just a lot of it is powered by humans training these AI programs.

17:36Yeah, certainly in the early days. And now it trains itself in many ways by sort of harvesting lots of data. But, you know, in the early days, that was definitely true. But now we get to the point where we start turning this into a real company and real money. Yeah, exactly. The Y Combinator Fellowship clearly was really useful for Globe. It opened doors. you know, it gave her a kind of seal of approval. In July 2017, Scale AI announced their Series A funding of four and a half million dollars from investors, including Axel Partners, Y Combinator and Arena Ventures. According to Bloomberg sources, Axel, which led the round, reportedly took a 25 % stake and it valued Scale AI at$15 million.

18:14Now, we don't know Guo's exact stake at this point, but she's co-founder, so it's likely she had a considerable stake in the company. So Late 2017 is a good place to say Lucy Guo is a millionaire. But she didn't start spending that cash immediately. In fact, taking a leaf out of her parents' books, she lived an extremely frugal lifestyle until she hit the$10 million mark a few years later. Back to our thing about the fact we'd have been out long before. I've been on a super yacht by now. Exactly. She said through her early 20s she continued to couch surf and get free food from venture capital apps.

18:47I'm sorry, but if my millionaire friend texted me and said, hey, I'm in town, can I sleep on your couch? I'd say, I'm not putting you up on my sofa. You're a millionaire.

19:06Let's take her then from a million to a billion. In November 2017, Guo, then aged 23, and Wang, aged 20, made the Forbes 30 under 30 list. Start-up founders want to be on this list. It is not a guarantee of long-term success by any means, but it's prestigious. Everyone knows what the Forbes 30 under 30 list is. It can also crucially boost your reputation and lead to networking opportunities. And fundraising opportunities. It's quite a big deal, the 30 under 30. I know people kill to be on that list. But while they're smiling in the photos, the relationship between Guo and Wang, who was the company's CEO at this time, was beginning to sour.

19:41Guo told Time magazine in 2025 they clashed over various ways the company was being run, including the late payment of contract workers, for example. Scale AI had long used those remote workers we talked about in the Philippines. By 2017, the company was hiring through third-party platforms like Remote Asks. And this subsidiary company has since been the subject itself of media and labour rights reporting concerning pay and working conditions. So it was not uncontroversial using this third-party platform. The differences of opinions between the co-founders eventually hit a breaking point. It resulted in Lucy leaving Scale AI in 2018, just two years after the company was started.

20:20The New York Times says she was fired. Other sources say she'd left. Guo told the Aspire with Emma Geed podcast in 2025 that she was already planning to leave when things started going south, but that she did feel betrayed at the end of the experience. She added that she believes the situation might have ended differently if Alexander wasn't the CEO. Now, Wang has declined to speak about the split. but when Guo left the company she held on to an equity stake in scale AI which is estimated to be around five or six percent or around 47 million class A shares. Okay so that's key because that's the key source we'll see to her will.

20:54A year after leaving Guo's life on social media looks like a lot of fun, rounds of partying, traveling, skydiving. She actually has a skydiving A license which means she can jump without an instructor. In 2019 Guo founded a venture capital firm called Backend Capital. That firm raised 8 million for its first fund, 40 million for its second fund, and they made early stage investments in a bunch of start-ups, some of which reached the coveted unicorn status, where we've used a lot on this podcast. That means a company worth over a billion dollars. In a way, it's quite a predictable move. You've had some success.

21:29Scali is recognised. You've been through the Peter Thiel and Sam Altman kind of factory. So you've got their kite mark of approval, their stamp of approval, that gives you some clout to be able to go and raise money to invest in other ideas. Because if you had that good one, maybe you'll have some other good ones and investors can trust you to know which ones are good and which ones are bad. It's funny that there seems to be such a route to gaining approval and kind of a thumbs up from the tech world. Even though all of these people are supposedly edgy radicals who dropped out of Stanford and MIT, there is still that well-trodden path to success.

22:05Yeah, you've got to burnish your credentials with a slightly different kind of establishment than that one. So she's out there raiding a fund to invest in other companies. Yeah. And remember, she's only 25 years older at this point. And how irritating. I know how deeply irritating for both of us. One of her best investments has actually been a fintech company called Ramp, because in 2020, Backend invested six figures and Ramp was most recently valued at$32 billion. So while we don't know the exact figure of that investment, it's likely to be worth tens of millions of dollars and growing. She also co-founded her own startup incubator called HFO, which is a three-month program where a group of founders live together in a house in America and focus on building their company.

22:48So she's kind of almost making kind of Peter Thiel-esque moves. Yeah. So while traveling the world, working with these startups and attending festivals along the way, she met a lot of online creators, you know, influencers, celebrities, etc., who shared frustrations about their income. They didn't think they were getting enough money most make their money from brand deals often inconsistent they're easily lost so in 2022 she launched something called passes a subscription-based sort of direct to fan platform and it lets creators earn money from their audience a bit like patreon with one-off payments as well as subscription a bit like only fans without the adult content bit of that but allowing people doing stuff online to collect more of the money from the attention economy which they're developing.