Most people believe that AI will disrupt jobs, but it will also create hundreds of new startup opportunities — only if one knows where to look.
Andrew Bosworth, CTO of Meta, recently offered a compelling vision for how AI agents could revolutionize the startup landscape. He envisions a future where AI’s limitations become the very foundation for innovative new businesses, fueled by a constant stream of user needs that AI cannot currently fulfill. His perspective highlights a more organic, user-driven evolution of AI-integrated services, moving away from top-down app development to a demand-driven marketplace.
Bozworth explains: “The stronger we get at agentic reasoning and capabilities, the more I can rely on my AI to do things in my absence. At first it will be knowledge work, and that’s fine. But once you have a flow of consumers coming through here, what you’re gonna find is that they’re gonna have a bunch of dead ends now where they’re gonna ask the AI, “hey can you do this thing for me”, and the AI will say “No I can’t”.”
He then pinpoints the opportunity: “That’s the gold mine that you take to developers. You’re like hey, I’ve got 100,000 people a day trying to solve this problem — trying to use your app. Here’s the query stream, here’s what’s coming through, and we’re having to say “no” today. If you build these hooks, you’ve got 100,000 people clamoring to use your service. And it’s totally fine for our AI to go back and say, hey you have to pay for this. There’s a guy who does this for you, but you’ve got to pay for it.”
Bozworth emphasizes this extends far beyond traditional apps: “And by the way, I’m not just talking about apps. It’s like a plumber. There’s something like a Marketplace here that emerges over time. That’s how I see it playing out. I don’t see it playing out as someone goes into a dark room, and comes up with this app platform. What’s gonna happen is, there’s gonna be a query stream of people using AI to do things, and that AI will fail repeatedly in certain areas because that’s a type of functionality that is currently behind some kind of app wall, and there’s no bridge that’s been built. Everyone wants to build the bridges,” he said. Bozworth said that he believed that these bunches of AIs performing different functions will become the way applications are built and used in the future. “I think over time, that becomes the primary interface for humans interacting with software, as opposed to picking from the garden of applications,” he said.
This vision differs sharply from the current trend of AI developers focusing on broad capabilities and solving stand-along problems. Bozworth’s perspective shifts the focus to identifying specific pain points within AI workflows. Instead of trying to create a universal AI that does everything, entrepreneurs should focus on filling the gaps where AI falls short, creating specialized tools and services that integrate seamlessly with existing AI agents.
The implications of this are significant. It suggests a future where entrepreneurs don’t need to compete against AI, but rather with AI. By analyzing the “query stream” of user interactions with AI, entrepreneurs can identify unmet needs and build solutions that AI agents can readily recommend and integrate. This also implies a shift in the funding landscape, where startups with a clear understanding of AI limitations and a focus on specific user needs are more likely to attract investment. Furthermore, this creates an opportunity for a marketplace to emerge, where AI agents can connect users with the specialized services they require, much like an app store, but driven by real-time user needs.