15 young founders are rethinking everything from artificial intelligence to zero carbon, from sustainable fashion to...pizza!

Image credit: Ben Rasmussen

It sweats $7.3 billion worth of tiny details.

As a student at the Massachusetts Institute of , Alexandr Wang fell victim to a wickedness that plagues college campuses around the world: His roommates stole his food. Specifically, they were taking his yogurts. "Yogurt is one of the easiest foods to steal because it's so perfectly contained," he says. "It's the perfect crime." So Wang placed a camera inside his refrigerator that he hoped would not only record the events of the refrigerator so he could review them later, but know - for himself - that yoghurt had disappeared. He soon realized, however, that such a system would take a lot of learning to do what, for a human sentry, would be a relatively simple task. On the one hand, the system should learn what yogurt looks like - and, just as importantly, what it doesn't look like. After all, a lot of other things are also perfectly contained. One would also need to learn the many places yogurt might be in the fridge and the many configurations of surrounding items. To achieve this, his system would not only need a multitude of images from which to learn; important parts of these images should be labeled. This idea of ​​labeling is essential. Only then could the AI ​​learn the nature of the relevant items. Only then could the data become useful. Indeed, only then could his system detect the theft of yogurt.

Wang is guided by one fundamental principle: We don't give a damn. And as pervasive as they are, refrigerator infractions wouldn't have fueled him forever. But Wang had studied artificial intelligence and had even been accepted into Y Combinator's prestigious accelerator program with an AI company in mind. His experience on the fridge made him realize just how essential – and underutilized – labeling is in turning the wealth of business data into beneficial uses. Labeling could empower projects that could help solve much bigger problems, like climate change and expanding access to medical care - i.e. solutions worth striving for. damn. After his freshman year, he dropped out of college, and at age 19 co-founded Scale AI focusing on data labeling.

Six years later, Scale is valued at over $7.3 billion, leading Forbes to declare him "the world's youngest self-made billionaire" (on paper, in any case). But he prefers to talk about another number in the billions: Scale provided 7.7 billion labels to various forms of data. He's partnered with clients ranging from Airbnb and Lyft to Pinterest, Instacart and the US Air Force. Its systems have scoured supply chain documents, tracked structural damage sustained during the war in Ukraine, and sorted images of people's skin to help improve the detection of dermatological diseases, among many other applications. /p>

Related: This 25-Year-Old Owns 5 Restaurants, $6 Million in Revenue, and a Simple Slogan: "Don't Be an Asshole"

To begin with, however, Scale focused not on the countless projects it could serve, but on one space: autonomous vehicles. "One of the most valuable things you can do is concentrate very, very finely at first," says Wang. By doing so, you can focus on fully understanding your customers' needs. In fact, Wang initially mispredicted what autonomous vehicles would need Scale to solve. "If we hadn't focused on this set of customers, we might not have talked to them enough to realize we were initially focused on the wrong set of issues," he says.

Wang believes in big breakthroughs and keeping the pace high, refusing to succumb to incrementalism. He originally planned to expand beyond the automotive industry after a year or even six months. This remained Scale's sole focus for three years. There was simply too much to know, too much to master. It's only once Scale has grown big enough that an entire team doesn't have to split their attention, but can focus entirely on a whole other set of ideas, that they are moved on to another industry - and equally unpredictable problems their customers wanted to solve. .

Three years ago, Scale had 100 employees working at its headquarters. Today it has 600. That begged a question: How do you determine who doesn't care? It's easier to see a small, restless crew not care . But what about hundreds and hundreds of people? In fact, Wang is looking for two forms of crap - on work in general and on Scale in particular. For form...