Matt Dalio spent three years building affordable computers for people in Rio's favelas who'd never used a PC. He cycled through Android, Linux desktop, and GNOME before learning the real barrier wasn't hardware cost but internet access. His small team included a user researcher, a favela field operative, and top Linux engineers like Christian Hergert.
Chapters
It's amazing how ignorant we are. I'll speak for myself. It's amazing how ignorant I was. I had all these ideas about the way people wanted to use computers, what they would understand, what they wouldn't understand, the paradigms you could break. You're going to people who've never used computers before. You can reimagine everything. And it turns out that it's actually really easy to understand what people want and how they understand stuff. Just put it in front of them. So employees number 3, 4, and 5 were, one, a user researcher who is now my girlfriend. Number 2 is someone who we hired to take her into the field and literally just bring her, and this is someone who used to work in the favelas of Rio when the favelas were bad, so her job was to walk up for a nonprofit, walk up to a guy with 3 machine guns on him, one in front, one on back, one over the shoulder, and say, I am coming into your favela for your people. You are going to let me in. And he's like, OK, you're good? We're good? OK, go. How do you find these people? I don't know. Fate, right? And so she took us into the favelas to do user research. And then the third was a designer whose sole job was just to iterate and prototype. The other person I didn't mention was a friend of mine who another fate of everything, who's an awesome designer, awesome marketing guy, and just an awesome friend, who basically simultaneously rolled his eyes like, here's Matt again doing this, and was like, I'm going to roll up my sleeves, dig in, and help you get this thing started out of friendship. And so here it is. That's the triumphant crew that's going to change the face of technology and get computers in the hands of billions. And— Bless you. But it really, it felt like this mission, like this is this thing, like we are like the torchbearers of humanity, like this is a mission. Something that we felt painfully every single day. So fast forward about a year from the time that I had started that Android app, and we realized that Android was not fundamentally a desktop operating system. We had to throw it away. At that point, we were basically throwing away everything we had, capital that had gone into that included. Uh-oh. So we started on Linux, and Linux is this very complicated operating system that's very powerful, and what we realized is that we could make all of those simple designs actually much more powerful as an engine and much simpler and much more smooth and much more— all the things that we wanted, we could actually do that on this Linux base. So we spent a year, over a year, building a Linux OS with a whole interface from scratch, and like, we're onward. And we had endeavors to install this into 5,000 computers in Rio, and this great journey of like— you could see the users' reactions at every point and how much they believed in this. This jacked dude in the favela of Rio, we're getting an açaí, this dude walks up and is basically kind of like, what are you doing in my favela? favela, and we had someone from the favela we'd hired, and he starts telling this guy about this. And at some point— at this point, I just wanted to know what do people want, what do people need. So, okay, tell me what you want. You're interested in music. Oh my God, we're going to make a music app in this thing, and you're going to be able to make music, and you're going to be able to share it with your friends. And at a certain point, he just looked off and teared up. And you could see this person who had been hardened by the realities of life feeling something. And at each point, it was that that was leading us to the next place. And so, yeah, this Fluxbox Linux GTK shitty operating system we'd built, we were on the right track because we could see it. Except that technically we weren't on the right track. Naritzi, the person I told you about, found someone who had helped build GTK, the framework in upstream Linux that we had built, found one of the people who helped build the thing and I had a coffee with him, 3 hours of tea, I had like 18 teas in that time, jacked up by the end. I managed to convince him to quit his job just not to join us. Well, Christian Hergert. And he managed to convince us that we were built upon the entirely wrong stack. So here it is, a year of Android, threw it away. A year of Fluxbox GTK, threw it away. Started from scratch based on GNOME. Who here knows anything about Linux? Does GNOME ring a bell for anyone? Okay. So GNOME is one of the primary— there's sort of Ubuntu and GNOME and those are the two. There's Canonical and Red Hat, Red Hat GNOME, Fedora, some of the things that may ring a bell for you. So at this point we had like a tiny, tiny team of engineers, tiny, tiny team of— it took a year and a half to find— to convince our first engineer to join us. This was someone who, in retrospect, was totally crazy. He moved his whole family across the country from Rochester, New York, like left their dogs and cats behind and like moved out here, him and his 3 kids, to join us. And yet we didn't have anyone who had Linux expertise, and so we hired this Linux contractor, which is the number one Linux contractor in the world, so like people who— Samsung, Google, when Google has Linux stuff they can't do, this is who they go to, to give you sort of an idea. And the CTO there at the time sort of joked, oh, if I wasn't CTO here, he was the guy who founded the thing, I'd be joining you guys, I'd be asking you for a CTO role. And so moments like this kept us excited, excited and moving on, except that him, his number one engineer, and his number one business development guy, who was also one of the really important engineers in the Linux space, basically told me, Matt, this is impossible. I just want to be really transparent with you. They didn't use the word impossible. They used— we went there for about 3 days and spent 3 days with them— about 3 days' worth of words to convince us that that this was not a wise endeavor technically. You won't be able to get a desktop operating system running on an ARM processor. Do you have any idea how hard that is? You won't be able to get OEMs to ship this. Do you have any idea what sort of processes they have in place in order to be able to ship this and that you have to fit into that? And it was a list of these things. It's too important. There is no question. This is too important. Put a brick wall in front of me, and I will figure out a way to batter my way through it. There are 3 billion people sitting on the other side of us figuring this out. So we started from scratch, throwing everything else we had away again. So here I am 3 years into this and basically don't have a line of code. 2 years into this. And so this is— let me just do a quick time check. This is where things started to get really on the right track. We were able to wrangle a really sharp professional operating system. GNOME is a really quality piece of technology in the underlying stack and in the shell, and we were able to make that be this very seamless, simple thing. An incredible amount of work, but we were able to do it. And what we discovered was 2 things. One is They were right. Getting this running on an ARM processor is basically impossible. Number 2 is even if you make it affordable, even if you make it simple enough so that they can use it, if they don't have internet, what can they do? And internet is actually the real barrier, it turns out. Just to sort of play a trump card on that, but just to give you sort of an example. So I got a chance to spend quite a bit of time talking with Bill Gates about this. his answer. I was saying, why aren't you making ARM computers cheaper for emerging markets? And he says, that's not the problem, Matt. Internet's the problem.
