Looking Past the Unicorn Hunt
Why trying to find the “killer app” for robots is a waste of time
It sounds like such a simple premise: “just build the killer app for robots and you’ll sell a zillion”, but it doesn’t work like that.
If I had a dime for every time an investor has said to me over the past 12 months, “you need to build the killer app for your Misty robot and then let’s talk” we wouldn’t need to be venture backed.
The venture industry is highly compensated for finding the quick buck, sure-fire, can’t-miss unicorn. So it’s no wonder that when they see a more foundational and disruptive technology such as multi-purpose personal robots, they look at it from the perspective of their core game plan.
I don’t blame them, or others, for thinking this way. It’s where your mind goes when you think of a tool that can perform a million different uses for thousands of different customer archetypes. It’s the shortcut, pass-Go-and-collect-$200, direct path to enormous success.
My usual response: “If Steve Jobs had known, beyond a shadow of a doubt (or even with 80% probability) that spreadsheets would sell 100x more of his product, don’t you think he would have poured everything into spreadsheets?” Because while I didn’t know Steve, I did spend a decade at his company, and I’m pretty sure he had a nose for commercial opportunity.
So why didn’t one of the greatest tech visionaries ever put all he had, right out of the gate, into spreadsheets? Because nobody knows where inspiration will come from.
Necessity is the mother of invention. But one person’s necessity is another person’s blind spot. Preparation and opportunity are also parents of invention, by the way. And Dan Bricklin and Bob Frankston, the inventors of the original killer app — VisiCalc —were definitely prepared. They had been working on spreadsheet-like concepts for a while but they had no viable outlet for their concepts, until the opportunity of affordable, powerful, and easy-to-program computers came along. How many other people in the late 70s were also inventing software for personal computers? Millions. Yet it was these two people who unlocked the secret.
This is a needle-in-a-haystack problem. How do a few dozen people at a small startup (or a few hundred inside a large corporation) find the “killer app” needle? Sure, they go out and do market research, create focus groups. They pore over Amazon listings and Apple store information and comments sections across the Internet to understand today’s pain. They hypothesize and analyze. But what are the odds that these few dozen people with their biases, blindspots, worldview, geography, and experience can compete to find the killer app with the million(s) who will start inventing, once given developer-level access to the very same platform?
Slim.
As in winning-the-lottery slim.
It turns out that modern-day Apple, not Apple-II-Apple, doesn’t help matters either. I could buy a car with those same dimes for every time I’ve heard “Apple proved that killer apps on a phone could work as a platform.”
My response to the Apple example is usually, “If you give me $100B of cash in the bank, the world’s most recognizable brand, one of the planet’s most powerful supply chains, and a few more years, I’m sure we can come a lot closer to finding the killer app for robots.” Basically, if one can amass vast resources, then one might find the one (or three) golden tickets among them.
Still, one company, however large, can’t build a hundred or a thousand consumer-grade use cases. It simply can’t have the diversity of thought and influences that are found in the world at large, from which all that varied creativity has to emerge.
And all that urging to just “go build the killer app” is also a red herring because it’s not clear that even platforms like the personal computer or the web took off with a single “killer app”. As far as I can tell, they took off with multiple killer apps. For personal computers it wasn’t just the spreadsheet. It was the fact that the same computer could be used to run word processors, databases, manufacturing shop floor systems, graphics programs, and eventually communications (email). For the web, if we had only Amazon and didn’t have search engines, or news sites, or social media it’s not clear to me that all of humanity would have flocked to the web just for the shopping. It was the fact that those single, flexible platforms enabled developers to do a hundred things — and do them pretty well.
This is all why at Misty Robotics we’re not investing our time and energy in developing robot skills (a.k.a. apps). Instead, we’re focused on providing the kind of flexible, rich foundation upon which the Dan Bricklins and Bob Frankstons of the robot world can build. If we do our job of making a great platform, they can do their job of pulling the needle out of the haystack. Then humanity as a whole, as it did with computers and the web, wins.
