DFINITY: How It Got Started. Talk with Dominic WILLIAMS, Part 1 of 3

This is Part 1 of my conversation with Dominic Williams. Dom and I sat down to talk about how DFINITY got started, how he came up with the name, and most importantly: where we’re heading to. Join our conversation about both the milestones that we’ve reached and the ones that lay in front of us!

Video Transcript:

Hello and welcome to another episode of Inside DFINITY. I’m still in Palo Alto, at our main office so far. These are super exciting times now: we have a few people joining every week by now: so a rapid growth and all the fun and troubles come with that. Super exciting. But today I want to talk about how it all started, specifically I’m going to be talking with Dominic Williams – our founder, chief scientist, the person it all started with. Dom and I first got in touch roughly two years ago. It was Q2 2016 and shortly after we met here in Palo Alto for the first time. Back then it was not at our current location but a previous hacker house that we rented and worked from. And so I’m even more excited now two years later take a bit of time with him to look back and also look forward and hear his vision for DFINITY.  So without further ado, please enjoy my conversation with Dom.

Origins of DFINITY

Cédric: All right. Welcome Dominic to Inside DFINITY. Dom, pleasure to have you here. Of course, everyone watching this already knows you. You’re the founder and Chief Scientist of DFINITY.  But maybe to kick things off one thing that people might not be so aware of is, “When did you come up with the idea of DFINITY? When was that? And how was it?

Dominic: There was quite a lot of prehistory; at the end of 2013 I started researching ledgers that could be tweaked to create a cryptocurrency for the games industry. And my interest stemmed from having created massively online multiplayer game for the groups of millions of users, so I was familiar with the needs of the industry. I thought it wouldn’t be great if people could trade in the virtual goods in one game for a cryptocurrency to take it to another game, and so on. The games could make money by taking commission on the purchases and sales of virtual goods. So I began digging in and I won’t mention names but I looked at some cryptocurrencies, all coins and all around them that claimed to be much faster.

Idea of a Completely New Cryptocurrency

Dominic: After a while I realized that they weren’t all that and there were major problems with the architectures and the crypto and distributed computing protocols they were using. By February-March 2014, I resolved to create a completely new cryptocurrency. My original approach was to deep dive into traditional Byzantine fault tolerant consensus protocols; particularly this type of consensus protocol called fully asynchronously free synchronous protocols. Try to find ways to repurpose those protocols for the decentralized network setting because decentralized networks have their own special needs. So I spent four months studying traditional Byzantine fault tolerant consensus protocols from 10 to 12 hours a day. It takes a long time, it’s a complex area of computer science.


Dominic: Then, started evolving my own protocols using the existing science.  The cryptocurrency was going to be called Pebble and I teamed up with Artia, who’s now Chief Operating Officer. We had some really interesting ideas, for example, we wanted to have on-chain automatic payments. So the idea was that Internet giants, like Facebook, could offer to remove advertising in exchange for automatic cryptocurrencies – the micropayments basically. We cast around and tried to raise money for the project.

Back then, some fashionable beliefs about the only possible consensus protocol being the proof-of-work, and if you proposed alternative consensus protocols, which nobody could understand,
right? Most people didn’t even really understand a proof-of-work and they just had this idea that Satoshi solved the Byzantine General’s Problem, which, of course, is nonsense; I mean there being consensus protocols around for decades. And if you proposed alternative protocols especially the ones that were much faster, you’re almost considered a fraud.

So on the one hand, you had the traditional kind of Bitcoin community, who were extremely skeptical and even angry that somebody could propose different consensus protocol because only Satoshi could do that; and on the other hand, the venture capital community were completely new to this and they either imagined that network effects would mean Bitcoin would dominate or they weren’t comfortable getting into this space – they didn’t feel like they understood enough. So towards the end of 2014, we disbanded the Pebble effort, we made gallant attempts to raise money after a while it became true that it wasn’t going to happen.

Consensus in a Decentralized Network

Dominic: Then, I kept on working on the technical side – early 2015, hanging off the skin or something like that, I had an idea about a different way you could reach consensus in a decentralized network. And essentially I saw that you could create this sequence of random numbers using this protocol called the Threshold Relay. And the remarkable thing was that you could create these randoms extremely efficiently in a way there was totally unmanipulable – no attacker could actually manipulate the sequence of random numbers, pretty much unstoppable, it’s highly fault tolerant and unpredictable.

I realized that if I could generate these round numbers efficiently and quickly in a decentralized network of any size, I could use these random numbers to drive other protocols. So once I had the Threshold Relay, I began working on other ways of bringing… the network having reached agreement on a set of random numbers I was now looking at protocols that could drive the agreement on things like transactions and state. So 2015 was a very productive year and I came up with all kinds of crypto theories and so on.

Scalable Blockchain

Early on I decided that it should be possible to create a scalable decentralized network – a scalable blockchain essentially. What about creating a decentralized cloud that hosts the world’s software and data? It would have some very special properties that would make it attractive for hosting standard software systems but you could also create a completely new generation of Internet services on an open platform. And the DFINITY project was born.

Cédric: Cool. So that dates back to basically your research in 2014-2015.

String Labs’ Cooperation

Dominic: I continued working on the protocols, and in 2016 it’s still kind of like a private project and then joined String Labs. String Labs was an incubator and its aim was to incubate cryptoprojects and get a stake of the founder tokens. Initially, we were looking at a thing called Mirror Assets, which was an evolution of an idea that Nick Szabo first had, which a company called Mirror Labs pursued for a bit. Was Mirror Labs a company that you were part of as well or is that something that happened here? That was based in San Francisco I worked for it for three months, I think, in 2015, trying to help them out with their software architecture and software development processes. So in 2016 we realized in the spring that Mirror Assets was too early and there were a lot of regulatory challenges that we probably couldn’t overcome.

Dominic: We said to the team at String Labs, “Hey. Look the DFINITY project is already well-developed and we have a whole lot of science for it. It’s also well-known in the industry, because I’ve been at events talking about DFINITY and the different techniques that I was working on for some time at that stage. I said let’s just do DFINITY. And so during the summer of 2016, String Labs started putting some money in and we built out the team, we created the foundation – you and I – in fall of 2016?

Cédric: Yes, I think it was the October of 2016. Time flies.


Dominic: We had the foundation; the team by this stage had some really brilliant new people in it, like Tim Hankey, Mahnush, and so on. By the way, I should credit Tom Ding as well. He was marvelous and at that stage he was helping to build out the organization with a lot of great people working on the projects, and early 2017 we ran – well some people call the first ICO – but it was very differently conceived; it was completely a fully-automated and fully-decentralized crowdfunder, but those were kind of naive days.

We decided, “Look, we don’t need that much money at this stage. So we’re going soft cap this at 1 million Swiss francs and as soon as we’ve raised 1 million Swiss francs, we are going to stop collecting money. After some number of hours and in the end I think it raised 3.9 million Swiss francs. One of the ironic things actually about that is to date I think it’s the only kind of automated crowdfunder that was ever fully decentralized – there were no central servers, it was all done with client software and Ethereum smart contracts. So we did that and we’re very pleased with the money we raised and super excited, but of course, very quickly a new kind of ICO arose.


Dominic: A few weeks later I think Gnosis went out and raised ten million dollars for 5% of the tokens. We went out there saying, “Look. We’re going to do this responsibly and just click one million dollars. Very quickly people were saying we want to raise as much as possible, as much as the market feels we’re worth and I’m going to let this crowdfunder to run for six weeks, uncapped. Then, of course, we’re going to run crowdfunds for a year uncapped. So the whole ethos changed.

And during 2017 for the tokens we’d received value arose and so we’re well funded and we just continued scaling out the operation but we didn’t have much money compared to some of the other projects in the space. But we stood back from it, we didn’t want to run a kind of ICO and that the same vein that we saw them being around because we felt a bit uncomfortable about some of the things that were going on. So we raised money in a strategic round later that year and obviously in the year 2018.

We’ve still been fantastically well-funded and we just scaled out the organization in a quite dramatic way. Before we dive in deeper on to funding and how we’ve used those funds so far in order
to scale out the team and move the project forward, Do you remember the moment when you came up with the name DFINITY?; and maybe, you can say a few words about like what it
means for people that have never heard about it.

General-Purpose Computation

Dominic: So I can’t remember the exact moment, but at some point I realized that it was possible to create a decentralized cloud. The Ethereum project was kicked off in 2014 and I’ve been very interested in that and stayed quite close to it. What I realized was that sure it’s a smart contracts platform, but I felt really the computer aspect was more interesting. So I thought okay: let’s focus this a bit more towards general-purpose computation – a focus on ways of massively increasing the performance and finding ways to scale out the network. My research showed that it was all possible.

Infinitely Scalable Network

Dominic: So D stands for decentralized obviously, and then the rest is infinity – Infinitely Scalable Network. So the ambition of DFINITY is to host the next generation of the world’s software and
data. And that’s why it’s called DFINITY. Actually if you go back on to the Wayback Machine, the Internet archives by the fall of 2015, It will be a very basic DFINITY website with a very embarrassing DFINITY logo that I created myself, but it really was bad when you look at the current on. The pitch is and the mission is pretty much exactly as it is now. So it’s been around for a long time.

It’s not like a pop-up ICO project. There’s a huge amount of technical work that’s gone into DFINITY. I think that’s why we’ve been able to sort of scale out the team with sort of genuine world-class talent and leaders in fields like cryptography and distributed computing, and brilliant Haskell engineers and so on. I was just mentioning this in the intro – it is a very exciting time; we basically have people joining every other week now. It’s really exciting where we’re thinking not only have we scaled out the office and operations here in Palo Alto, but also we’re thinking about opening offices in different places of the world.

Cédric: Maybe, just briefly can you give a brief overview of what happened in the last 12 to 18 months since we’ve completed our first fundraising that you mentioned? What were the milestones that DFINITY has had before we start looking forward?

DFINITY Research Centers

Dominic: Organizationally, we’ve upgraded from the kind of a hacker house that we’re in to Palo Alto Research Center, we’ve got a new research center going up in Zurich, we have people all over the world, we have people in Germany, the UK, and three people in Japan, and people all over the US. So people come and go through this office here. It’s the kind of place where people come and meet and the same will apply to the Zurich HQ. Setting up a San Francisco Research Center as well to save people the Caltrain.

So I think that two biggest basis is really a sort of like Switzerland Silicon Valley thing, but we’ve got people all over the world. So that’s brilliant and it’s enabled us to have quite a large geographical reach when looking for talent, and the very best talent in the world isn’t in one place – it’s everywhere. You’ve got to be willing to provide them with reasonable options of where they can relocate to.

Cédric:  That’s been brilliant. I mean just talk about the tech in a moment that a lot of people forget how difficult it is to scale an organization out.

DFINITY as NASA for Decentralization

Dominic: DFINITY’s aim has always been to build this thing we term the NASA for decentralization.

Cédric: NASA as a space agency.

Dominic: NASA as the space agency. DFINITY is obviously a not-for-profit foundation and NASA’s a not-for-profit organization, and we believe that if you really want to create a platform that hosts the next generation of the world’s software and data, not only does it have to be brilliant but it’s necessarily going to be very complex; and it’s also going to be the very large community of engineers and researchers supporting it, looking for security vulnerabilities, fixing problems that occur, building it out, developing it, and making it better and better. Otherwise people are never going to have the confidence to build on top.

Best Talent at DFINITY

Dominic: So we’ve always differed in that respect from some other players in the decentralisation industry in the sense that we don’t think that protocol alone is enough; we think that we have to build this huge pool of talent that will stand behind the network itself and continue developing it. That’s why we want to have this NASA for decentralization. A lot of people might imagine that if you just have lots of money as we do, it’s straightforward but in actual fact scaling organizations is really tough. You have to find people, which have the right cultural as well as sort of technical fit, you have to hire operations people like yourself. As you grow larger, there are all kinds of complexities relating to moving money around, tax, offices, and payroll.

I think that’s been one of our biggest achievements so far the fact that we know we set out to build a NASA for decentralization, we’re nowhere near that yet but we have been able to methodically
scale it out. The process is accelerating so the quality of the team is absolutely astounding; people are going to see a new wave of hires announced in a few weeks time.

I think the people will be dumbfounded by the quality of some of people that are joining us, world famous names in different branches of computer science and superb managers with incredible track records. That’s I think separately to the technology piece. The team that we’ve built and the organization we’ve built is probably one of our biggest achievements.