Hackers and Scientists

In May 2003, Paul Graham (founder of YCombinator) published one of his most famous essays, “Hackers and Painters”:

“Hacking and painting have a lot in common. In fact, of all the different types of people I’ve known, hackers and painters are among the most alike. What hackers and painters have in common is that they’re both makers.” – Paul Graham

It was long after that I read it, probably around 2014, by the time I started working on my first company. Still, I remember that it highly resonated with me. I bet you have been there, as well: with a fresh painting coding assignment, looking at the empty canvas screen and thinking carefully (even with pleasure) of all the implications; of the different parts and how they would related to each other in the grand scheme of the composition architecture. Probably you went through that process several times, creating and destroying, and even drafting prototyping in the middle. You felt that just with your brushes and colours screen and keyboard, you could paint build anything. Rings a bell?

“The Last Supper” by Leonardo da Vinci. Restored version.

I had those thoughts in my head when I started my first company, Relendo. It started from a simple realization: we have all gathered many things that we bought during are lives, and those things are 99% of the time waiting in a drawer just accumulating dust. What if other people could use those things for a fee? Yep: we discovered rentals. Not any kind of rentals: rentals between peers. We wanted to be the Airbnb of everyday items: photo and video equipment, tools, sports equipment…

With that intent, we started building. We used all our creativity to decide how the product should work. We felt like Leonardo da Vinci imagining The Last Supper and placing mentally each of the characters in the perfect spot, with the right expression and expressing a subtle but elegant meaning that flowed directly to the viewer. We could perfectly imagine how users would react to this and that feature, and how the product would serve them swiftly like an extension of their own body. We imagined a spotless masterpiece. And then we built it. Or some part at least. Want to know what happened next? Well, it didn’t work quite actually. Our master piece, so well thought and masterfully executed wasn’t appreciated and much less achieved its goal. Only a handful of users got to experience it, and they left more confused in the best of cases. We were devastated. What had happened? What could have we done wrong in the process?

Pacific Northwest National Laboratories researchers picture from Unsplash.

Not long after that, we discovered Lean Startup, and the idea that you have to accept you know nothing (or close to nothing) of your users, so you should build, measure and learn as fast, efficiently and cheaply as possible, in iterations, to maximize your learning:

“The only way to win is to learn faster than anyone else.” – Eric Ries

So we started following Lean Startup, and discovered that it was simple, but not easy. When we were dreaming about the grandiosity of unbounded creation, Lean slapped us the face to remind us that this is not about you, or what you create (your solution), but about the problems that you are helping to solve. With Lean came many other things: like hypothesis, Minimum Viable Products (MVPs), A/B testing, metrics…7With all those tools, we felt like we were making real progress into understanding our users and helping them. We also felt that, slowly and steadily, we were becoming scientists: we laid our hypothesis about how our users would react to certain feature, and defined an experimental model based on metrics to track its impact in our Key Performance Indicators (KPIs); we then built and released it, always carefully measuring every step on the way to ensure no piece of information was lost. Finally, with all the data, we sit down and analyzed it, in an attempt to remove some of the uncertainty surrounding our understanding of our users and our own product. Sometimes, we realized that it was all for nothing, just because some of the process had failed (we were learning, as well). Some other times, we realized that the data was not conclusive, despite the long time we had been measuring it. In the end, we were just applying the good old scientific method to create a product, and a business on top of it

A Google Analytics dashboard like the ones you can use to track your experiments.

I still agree with PG in that hackers are more similar to painters than most people think. Besides, what I believe is that nowadays hackers are also more like scientists: dealing with tons of uncertainty, laying hypothesis that can be tested, building with the purpose of learning and understanding, and, in summary, in the business of learning as fast as possible.

Share on facebook
Share on twitter
Share on linkedin

Fulbrighter, MSCS/ML at Columbia University, and CTO & co-founder at FounderNest. I am fascinated by artificial intelligence, entrepreneurship and agile software development. I consider myself open minded, detailed and humanist, and I love learning and growing as a person every day.