They say Earth is home to over 10,000 AI startups. There are more of them than cheetahs. They outnumber the redwoods at dawn. This number is, of course, a guess – startups come, startups go. But last year, more than 2,000 of them received their first round of funding. As investors pour billions into artificial intelligence, it’s worth asking: What are all these boomers doing?
I decided to contact as many recent AI creators as possible. The goal was not to select winners, but to see what it looks like in practice to create AI products – how AI tools have changed the nature of their work; how scary it is to compete in a crowded field. It all sounded a bit like trying to tap dance on the turbulent surface of the sun. OpenAI introduces an update, and an avalanche of posts on X announces the slaughter of a hundred startups. Brutal!
Is this a revolution that will end with the burnt legs of so many engineers? Certainly – not everyone can survive. A startup is an experiment, and most experiments fail. But run thousands of them across the economic landscape and you just might find out what the near future holds.
Navvye Anand is co-founder of a company called Bindwell. When we connected on a video call, he spoke with a half-smile and somewhat elegant manner, telling me how he develops pesticides using custom artificial intelligence models. Bindwell’s website once described these models as “incredibly fast” and claimed they could predict in “mere seconds” the results of experiments that would take days. Listening to Anand explain how he’s bringing AI drug discovery principles to crops, it was straightforward to forget that he was 19 years elderly.
Anand grew up in India reading Hacker News with his dad, and halfway through high school he was building his own gigantic language models. Before he graduated, he, his co-founder (now 18), and two other summer camp friends published a book paper on bioRxiv, about the LLM they built to predict an aspect of protein behavior. This caused scientists to focus on X. The article was like this quoted in a respected magazine. They decided they should try building a startup, brainstormed, and settled on protein-based pesticides. Then, as the fairy tale continues, a wood spirit (sorry, venture capitalist) contacted him on LinkedIn and offered them $750,000 to drop out of high school and college and work full-time at the company. They agreed and started. The teenagers had no idea about agribusiness. This was in December last year.
Five months later, Anand and his co-founder opened their first biotesting lab in the San Francisco Bay Area, then moved to another one where they personally squeeze drops of promising molecules into petite vials. (The protein compound may, as the theory goes, target locusts or aphids more precisely, and not also destroy humans, earthworms and bees.) I asked him how he acquired the skills to work in a moist lab. “I hired a friend,” he said cheerfully. A friend coached him all summer before he returned to college in the fall. “Now I can run some biochemical tests,” says Anand. “It’s not about a whole range of tests, but about basic validation of our models in the wet lab.”
Hmm, I thought. That a few teenagers had, in a matter of months, built their own LLM, learned the biochemistry of pest control, used their models to identify candidate molecules, and were now pipetting in their own lab didn’t seem shabby. In fact, when I added up everything they did, it seemed completely absurd. I expected to hear that AI tools were speeding up the stages of building a business, but I only had a vague sense of the scale of their impact. And so in another interview with the co-founders of a 14-month-old startup called Roundabout technologiesI went straight to this: write down what has changed and by how much.