Eric von Hippel was my doctoral advisor at MIT. I didn’t finish the Ph.D. – I dropped out – but Eric stuck with me through all of it, more as a mentor than an academic.
Around 1978, Eric said something that was radical at the time. “Innovation comes from users, not manufacturers.”
Today that sounds obvious. You read it and think “yeah, uh huh.” In 1978 it was heresy. The entire model of how products got made assumed the manufacturer was the source of the idea and the user was the recipient. Eric spent the next 48 years building the academic foundation underneath the opposite claim, along with a worldwide research community to go with it.
His Lead User work is the foundation for a wide body of research that came after. Lead users are the people who hit a need before the rest of the market does, and who build their own solution because nothing on the shelf does the job yet. They aren’t waiting for a vendor. They’re already solving the problem for themselves. If you want to know where a product category is going, find the lead users and watch what they’re hacking together.
His later work on sticky information is the part that matters most for what I’m working on. The insight is that the information you need to solve a problem is expensive to move. It’s sticky. It lives where the need lives, which is with the user, not with the manufacturer. So the place where innovation actually happens shifts to wherever the sticky information already sits. That’s usually the user.
Free and open source software is the cleanest example of all of it. The people who needed the software wrote the software. The user and the manufacturer collapsed into the same person. Eric pointed at FOSS for years as the canonical proof that his framework was real and not just a quirk of scientific instruments and surgical equipment. His book Democratizing Innovation is full of examples, and he released it as a free PDF under a Creative Commons license, which is about as on-brand as a theory of user-driven innovation can get.
My own research under Eric was on the boundary between the user and the manufacturer. Where does one stop and the other begin? When does a user become a manufacturer? I’ve been chewing on that boundary for almost 40 years.
Here’s my premise.
All future end-user software innovation will come from the user. Not most of it. All of it. And it should.
The first half of this is just Eric’s thesis, extended to a world where software is ubiquitous. The second half is the new part. The machines should be able to build it in a fully automated way. The user has the need and the sticky information. The machine has the ability to write, test, and ship the software. The boundary I’ve been pondering in different ways for 40 years is about to disappear, because the manufacturer is becoming a machine that the user operates.
When I built a small plugin earlier this year, I described it through a framework I’ve started calling The First User, which is my own riff on Eric’s Lead User. I didn’t build it because I thought it would sell or become a product. I built it because I needed it. I was the user, the machine was the manufacturer, and the boundary between us was a prompt.
A lot of people are circling this idea right now, talking about a bunch of overlapping concepts. One of my favorite thinkers here is Dan Shapiro. He’s been writing about the Software Factory and the Dark Factory – a set of levels that run from spicy autocomplete up to a fully automated factory where no engineer writes the code and no one is in the loop.
Instead of theorizing about this, I decided to build it.
For the past six months I’ve been building multiple machines that build and run machines. They’re all software machines. I deliberately use the word “machine” instead of “agent.” Agent has become a meaningless word – it gets attached to everything and means nothing. A machine is a fully defined, functional thing with inputs, outputs, and a job. Using the word forces me to actually define the thing rather than wave my hands at it.
Years ago I wrote about The Three Machines – the Product machine, the Customer machine, and the Company machine – as a way to think about how a company is organized. It’s still one of my favorite posts, as it helped me simplify how to think about a startup in a new way, and then use Dr. Goldratt’s Theory of Constraints, from his novel The Goal, to figure out where to focus energy in the early stages of a company.
I’m using this theory of constraints to figure out whether any of this works. Find the constraint, fix the constraint, find the next one. The constraints keep moving, which is the whole point.
Right now this is a project that I’m working on with my brother Daniel. Maybe it turns into a company. Maybe it fails. Maybe something else happens with it. The future of this one is non-deterministic (see what I did there?)
What I do know is that I’m learning a shitload. I’m also having a ton of fun working with my brother, which is a great joy to me.
While I’m paying a lot of attention to what other people are doing, I’m using a philosophy many of the CEOs I’ve worked with in the past have heard. “Be obsessed about your competitors’ products and what they do and then completely ignore them and play your own game.”
I’m in some online groups with people who are a lot smarter than me or are extremely experienced software developers. They are openly sharing what they’re figuring out, creating, and discovering. However, based on the approach I’m taking, a good chunk of the time I feel like I’m in a parallel universe, because some of my premises are just different from theirs.
That’s a fun and interesting place to be.