Last September, OpenAI CEO Sam Altman shared this provocative idea during an interview with Reddit cofounder Alexis Ohanian:
“In my little group chat with my tech CEO friends there’s this betting pool for the first year that there is a one-person billion-dollar company, which would have been unimaginable without AI and now will happen.”
Whether you agree or disagree, a one-person billion-dollar company is a bold statement worthy of debate. But what’s not debatable is that technology innovation is continuing to happen at a staggering rate, leading to huge transformations and benefits for all of our lives. And one of the biggest beneficiaries of that innovation is technology itself.
Technology Prosperity
Back in 2012, I wrote a blog post about a concept I called App Prosperity, observing that technology itself was making new tech products easier and easier to build. In his book “The Rational Optimist”, author Matt Ridley defines prosperity as “the increase in the amount of goods or services you can earn with the same amount of work” and outlined some inspiring examples for how prosperous the world has become. In the 1800s, one hour of work would earn you enough wage to buy 10 minutes of reading light. Today, one hour of work can buy 7,200 hours of reading light.
The same is true for technology. 20 years ago, a billion dollar software product like Microsoft SQL Server 2000 took 2 years and hundreds of engineers to build. Just one decade later, three people built Instagram in 8 weeks because of advancements in technology. Technology has invented new building blocks that have made subsequent technology faster and cheaper to create than ever before.
The Internet is one example of a transformative tech building block. Before the Internet, engineers had to manufacture, package, and distribute physical media to ship software. Now you can release and update software instantaneously through the Internet. Cloud computing is another transformative building block that has eliminated the need to buy, rack and maintain physical servers. Using virtual machines as a building block, writing new code is now easier for engineers. Using open source products as building blocks, engineers don’t even have to write new code because they can reuse someone else’s.
Building blocks have made engineers more prosperous than ever before. And building is about to get even easier and more prosperous with the introduction of perhaps the most disruptive tech building block ever invented: AI.
Technical expertise required
Over the past 2 decades, the Internet, cloud computing, virtual machines, open source and other building blocks have made technology products faster and cheaper to create. Companies need less time with fewer engineers to produce great products. But you still need engineers to do the building. In fact, engineers have arguably been the most valuable resource for tech companies since the turn of the century with companies like Google and Facebook fighting a compensation and benefits arms race (covered extensively in news articles like this) to recruit and retain engineers.
The startup world is no different as VCs have clamored to invest in whatever idea an ex-Google or ex-Facebook engineer left to pursue. When it came to starting new companies, technical talent was what mattered most because engineers were still the only ones who could build products. Everything else — business understanding, industry experience, go to market strategy — was secondary because it was assumed it could be learned later.
A technical founding team revolutionized food delivery (DoorDash), not a team of restaurant professionals. A technical founding team revolutionized music (Spotify), not a team of record label professionals. Same for hospitality (Airbnb), groceries (Instacart), payroll (Gusto), corporate spending (Brex), and more where the technical talents of the team were touted front and center. Being technical was both necessary and sufficient. Technical founders and their teams got investments to build their companies, and non-technical entrepreneurs got feedback to find a technical co-founder.
For startups, the rule was technical expertise required, domain expertise optional. And then came AI.
Domain expertise required
We’ve all seen the amazing tasks that AI can perform from summarizing documents to generating media to even writing code. But a key aspect of the AI advancements that have been popularized over the past few years is that the most common interface to access this breakthrough technology is simple natural language prompts. You don’t have to use a query or markup or some other programming language. You just provide a prompt — a plain English request — and the AI system goes to work from there. AI’s not just an incredible tool or technology building block, but it’s also the first one that you don’t need to be technical to build with.
As AI continues to improve while still being accessed by prompts, imagine giving AI a request to build something and getting not just code in return, but entire services or products in return. Just like cloud computing made physical servers an implementation detail that engineers don’t have to worry about, AI could make implementation itself a detail that no one has to worry about. Technology isn’t just accessible at that point, it becomes automatic with AI.
And that’s how you get to an entirely different type of startup in the future, perhaps not as small as Sam Altman’s hypothetical one-person company, but still a very different company nonetheless. Because when AI can make anyone a builder, engineering talent no longer is the most critical resource for startups. Building with AI in the future won’t be about who can write the best code, but rather who can craft the best prompt. And the person best equipped to do that is the person with the most domain expertise, not technical expertise.
VC investors will have to focus on different types of entrepreneurs beyond the ex-Facebook, ex-Google, ex-Amazon engineers because the next disruptive food technology startup could be founded by a restaurant owner, the next disruptive HR platform could be built by a benefits manager, the next popular music app created by a musician, and so on. With AI building blocks, the best founders going forward may be those with the most understanding of the problem to be solved, instead of those most qualified to build the solution.
So the new rule for startups in the age of AI? Domain expertise required, technical expertise optional.
