TL; DR
Instagram had 13 employees when they sold to Facebook, now Meta.
As AI accelerates ever-more, a "one-person unicorn" appears close. But I argue that we're not there yet.
What's especially interesting is what kinds of startups to expect from its predecessor, the "two-person unicorn"– and whom you will need to staff one.
Companies have already gotten much smaller than ever before
One technological breakthrough after the other has allowed companies to do ever more with ever fewer people.
- Assembly line took the place of many artisans and helpers.
- Electricity gave individuals the "muscles" of many people.
- Computers eliminated manual calculations.
- SaaS eliminated on-premise server and IT requirements (similar for "x as a service" for many real-world tasks).
One of the most famous examples of this was Instagram having 13 employees when they sold to Facebook, as a "unicorn" company, i.e., one worth at least $1 billion.
Good or bad, AI-based automation and creation are coming to work
Now AI is increasingly making programming, much of marketing and sales, and even some operations accessible to everyone.
This will remove the last need for big groups in young growth businesses (for topics without licensing etc.): specialty knowledge. (Prior tech already removed need for physical productions, IT infrastructure, and most commerce.)
We'll get 2-person unicorns before 1-person unicorns
My friend Brendan Cahill (here and here) coined the idea and vision of a "1-person unicorn," a billion-dollar company run by just a single person, plus a bunch of tech tools.
I agree with his vision. It's just too tempting for too many people to think that each of us can now make our billion-dollar idea happen on our own and become AI-magically filthy-rich. People will work so hard for it that it WILL happen.
But that won't happen instantly.
Even my intermediary step of a 2-person unicorn likely is still in the future. We may be closer to an 8- or 6-person unicorn for now. Plus, those numbers look very different for pure-software companies vs. ones with a manifestation in the real world, e.g., ones that have to produce and ship physical products.
But still. It's good to think ahead "what must be true" for such a 2-person unicorn to exist and how it might look.
How 2-person unicorns may look: The rise of the non-techie
AI startups are still just startups
Many AI-centric startups will undoubtedly look similar to those we have today, just better/ faster/ cheaper.
They may make gobs of money. But they basically represent "incremental innovation." Better but not totally novel.
Much more exciting: What completely new classes or types of startups might appear!
There will also be a totally new kind of startup, created by people with ideas to solve problems.
Most of those people will NOT be technologists. They will come from all walks of life.
That’s because technologists already can create projects, products, and companies that earn them money today. Cloud infrastructure and payment APIs, at minimum, have made that straightforward for talented full-stack technologists.
As a result, there is only a moderate backlog of ideas. Many things that technologists could imagine were already possible to create. Sure, AI will unlock more ideas. But others have already been realized.
By contrast, non-technologists (business people, professionals from all walks of life, and amateurs across all fields) have not had this opportunity. Without access to technologists, most cloud-based businesses were impossible for them to create until today.
That changes with AI. Not only can those people create AI-centric businesses now, just like technologists. They can also on their own bring to life many ideas that “merely” require the existing generation of technology: Apps, APIs, cloud infrastructure, and the like. We will experience all those ideas now too.
With matured AI tools, these non-technologists can now create new companies without/ with minimal support from technologists.
Don't expect non-techie 2-person unicorns tomorrow yet
Bunches of barriers to entry remain that block this change in power.
- AI is still hard to use: Especially, advanced AI interfaces are still very hard for non-technologists to navigate (eg require understanding of advanced math and at least some Python, ability to install models locally and to access APIs, slogging through minimal/ poorly-designed UI and almost non-existing training).
- AI takes new thought processes: AI agents, in particular, require a new way of thinking. Where current code takes “computational thinking,” this “agentic thinking” carries over some computational aspects (eg object encapsulation, service-oriented architecture) but also requires new additions, eg process mapping and leaning out from process improvement and solution and worker choice from executive skills.
- AI agents need new "know what:" Finally, there will be a need for new baseline/ quasi-instinctual/ native knowledge (“know-what, not know-how”), eg clear understanding of the difference between different agent classes and between agents and other AI. Just like we’ve had to learn every other technology in our lives, we now are experiencing another round of such onboarding.
These preconditions will need to be met for non-techies to unleash a new class of startups that are much more use case-/ niche-specific, possibly more intensely problem- and user-centric, and initially, much more different from each other, ie likely to break design conventions that have standardized since the early days of GUIs, the Web, and Mobile technology.
All to say, there’s still a lot that must happen before our happy, shiny multi-intelligent future arrives. Meanwhile, it’ll still be a slog to make things happen.
There's still value in the transition period, and operators will unlock it
The transition period will unlock ideas
It is this “in-between” time where more people can develop their ideas for what is now "adjacent possible." Many people don't even know that they can now do much more than before.
So far, so good. But, as mentioned, they'll then only be able to implement their ideas with difficulty. So we will not suddenly go from current company size to the one-person unicorn.
Enter the operator.
Operators will help entrepreneurs not drown in chaos
Today, operations professionals appreciate their own rigor and structure and result generation. But let's be honest. To many others, especially to innovators, operators are seen as a bunch of squares.
It doesn't help that Lean Manufacturing and 6Sigma overdid things, trying to "optimize" and "efficient-ize" things that never needed it, like company culture. This caused these management philosophies to attract terrible reputations, some deserved, some not.
But bottom-line, operators are not who people think of when they imagine "innovators."
Why? The current "default" founder duo for fancy startups is 1 business person + 1 tech person. The business person also handles ops and, mostly, fund raising. The techie builds "the startup's thing."
But if "building" becomes ever more simple for non-techies, just like prior tech expertise got virtualized, i.e., eliminated from view, then the techie will only be needed for some startups anymore.
A business-centric founder can take on much of that work using freed time from other automations.
At that point, the title of "business-centric founder" also no longer makes sense. Really, they become the "domain expert founder." They know a space. They know what to build. Now they can build it. And, having built it, they become the ultimate expert to sell it.
Ah, but then the domain expert's "plate" will be quite full of work. Eventually, it'll become too much, and chaos will set in: "Good growth advice becomes bad growth advice."
Then, the domain expert needs a "company building expert." That person won't be an "operations" person in the old, or, rather, in the "steady state company operations" mold of things.
Instead, this company building expert will be a creative craftsperson who can shape a company's operations to fit the domain expert's specific vision. No templates, no limiting the vision to force it into some operational straightjacket. Company builders will build companies around the domain expert's vision that amplify how well the vision can come to life.
In other words, operators will "10x" what the domain expert can achieve, to use a Silicon Valley cliché.
The company builder, too, will have a range of technology options at their disposal for automating ever more than possible before. Those technologies won't all be AI. They will be a pragmatic mix of tools.
And, to be clear, they TOO will be experts, not generalists. They will just be experts at company building, not at the domain. Think "multi-experts." They will know how to implement professional finance, accounting, supply chain, inventory management, marketing, legal, HR, and more. Whatever function a company has, they will be able to create it professionally, not just in a half-a**ed way (which would stall growth by itself).
The question will be how many of these expert skills they can have or build themselves, or how many outside experts they will pull in to close gaps. But the number of extra people needed will diminish.
And that will put us on the path from the 2-person startup to the true 2-person unicorn.
Watch it happen.
Want to work toward the "2-person unicorn?" Say hi!
Further reading:
Cahill, Brendan. "The Age of the Full-stack Entrepreneur is Upon Us [Part 1/2]." LinkedIn (2024).
Cahill, Brendan. "The Age of the Full-stack Entrepreneur is Upon Us [Part 2/2]." LinkedIn (2024).
