There was actually a one-week period where I more or less shut myself in a very small hotel room in Tokyo. So came up with like an alternative concept that was what was to become Polkadot.
Today we launch. Okay. Today we launch.
Taking all of those ideas and then applying them to leading edge distributed decentralized systems. Awesome. People want to be able to do things.
Bring in the normal people. Make everybody able to like run your stuff on the decentralized computer.
Here is some construction material. Well, here is some knowhow. Build what you think is going to be the best thing that it can be.
If this is the scope of web three, crypto blockchain, not only compete within it, but try to like go beyond it and see what else is possible.
This freedom is well earned and we should use it as best as we can.
Welcome everyone. We're here to celebrate the 10th anniversary of Parity. We're going to chat about history and also see what was the background and the story of how we got where we are today. Gavin, tell us about how Parity was founded which was originally called ETH Core and tell us about the early days.
Well, it sort of began conceptually, I suppose it began when we were in we'd launched Ethereum, so it was like autumn 2015. And you know, it seems odd now, but back then ETH wasn't worth very much. It was down to like 50 cents and the Ethereum Foundation didn't have any real capital. And its ETH reserves didn't seem to be so extensive. So Vitalik and I sort of and a bit Jeff but mostly Vitalik and I were like let's try and do something like that maybe can get some investment that can continue this down this path. This was sort of the basis of the initial meeting that we had in London that sort of put down the ideas for what was to become Parity.
Vitalik after this meeting at some point a few weeks later sort of said he didn't want to be founder of the company but he would be an adviser. So that's where sort of he fit in. And we sort of knew that the Ethereum Foundation chapter was sort of coming to a close and that it made a lot of sense to sort of go off and try and create this Ethereum sort of spin-off that potentially could receive investment going into products. That was the sort of premise that we began with. Putting a bit of seed capital, I think we sort of raised between ourselves a few tens of K and we set about like seeing if there was because there were a lot of people that were working in EF DEV that were uncertain of the future within EF DEV. Again, foundation didn't have so much money so it couldn't really guarantee anyone's future. And that's basically where we sort of decided let's create this Rust client for Ethereum. That was our first sort of technical project. And so two of the coders that were sort of very prominent in the C++ team in my part of Ethereum, Ark and Marek were very keen on sort of pushing this forward. So they became sort of two of the first employees, founding employees.
And then at what point the idea of Polkadot and its paper came to be?
So yeah, we focused on building this Rust client as EF Core, we called the Rust client Parity, for a few reasons. We found that there was no obvious way of moving forward within the Ethereum ecosystem for us as core developers. Of course, there were people doing application layer stuff and launching their own tokens and all the rest of it, but we didn't really want to do this because that wasn't our core competency. We've always been quite low-level systems coders.
Around summer 2016, late summer 2016, Marek and I were in San Francisco. I can't remember what we were doing there, maybe we were meeting a VC or something. And we were just sort of discussing what because Ethereum hadn't made much of a move towards Ethereum 2 which was what we were all about. We were like right we're going to be the coders for ETH2. We're going to be the best ETH2 client. And they really haven't. I mean, they still haven't 10 years later.
But they had the execution sharding model that they...
Precisely. And so, you know, at that point, it's like we didn't want to step on that turf. So, I didn't want to come up with an alternative to ETH2, but then I also didn't want to just sit around doing the same ETH1 stuff. So, I came up with like an alternative concept, and that was what was to become Polkadot. Which was basically this idea of having not a homogeneously execution-sharded chain but heterogeneously sharded chain and really pushed the idea of these execution shards being their own thing with their own token economies and their own program code and domain-specific functionality but also base it all on WASM so everything was easily checkable.
And a very early version of Polkadot before it sort of came to the conclusion that WASM was totally practical to be used, the assumption was that the actual relay chain client would include hard-coded versions of all of the parachains.
And so if you wanted to launch a new parachain type, you would actually have to get the relay chain code maintainers to sort of include your logic, which you know was going to be a lot more horrible. But thankfully that's not what turned out, which was a lot better.
Was WebAssembly known at Parity at this point?
No one had thought of it at this point. I mean it was relatively new in summer 2016. Like it wasn't long. I think it was around, but I think it had only just been sort of standardized. And the big sort of cool stuff was still like asm.js.
Um, and I think there was another one. I think asm.js was in Firefox and I think Chrome had some other thing.
And then the two teams sort of collaborated to standardize WebAssembly. And then it became clear over time that WebAssembly was going to be fast enough to be able to run the on-chain logic.
Then eventually Parity built Substrate which is also interesting to talk about. It's used both within Polkadot and also outside. What was the motivation behind it?
So I mean the basis of Substrate was just Polkadot. They wanted to make the Polkadot relay chain.
And so if you actually go to the GitHub repository of Substrate and you sort of I don't know if GitHub records this, but the repository began being called Polkadot because it was Polkadot, right? It was just a Polkadot chain and then we renamed the Polkadot chain to Substrate and took the Polkadot specific stuff out and put that in the actual Polkadot repository. And you know, whether this was such a good idea in hindsight is arguable. I mean, it did probably create a fertile ground for Substrate to be utilized by other projects.
I think as a technological contribution to the broader blockchain web space, I think Substrate is something that I'm happy I've contributed to. I think it has served a lot of purposes. That's a good thing we've achieved.
This brings us to about like mid 2018, mid 2019, which is where Parity eventually hired the rest of us in this table. So, I want to now talk about that. You two can each say how you found Parity, why you applied, and what was the first days and weeks like?
I can go first. I think I was hired before he.
Um, October 2018. Then it was early then.
Um yeah, so I was just I happened to be living in Germany at the time. And I kind of needed a job as I was broke and I wanted to work at blockchain because I'd been like into blockchain since 2014. A lot of my friends were in the open source movement and were into this stuff hacking on it and things like that. I said, 'Okay, I'll go to Berlin Blockchain Week, see if there's anything good.' And I met some Parity employees at a place that was unrelated to Parity. It was just there was free beer. And we just kind of got to talking and they were really cool people and I said I was looking for a job and they said you should apply and I applied and I got the job.
Yeah. For me it was, yeah, I was also searching for a job. I was working for a startup before and I was on like parental leave in April 2018 and when I came back it was like yeah there's not that much more to do. So I was like getting another like 5 months of I was still getting paid but I was just doing some open source work and then realized okay it's time to find a new job and I wanted to continue doing Rust. I was doing Rust before. So like yeah I don't know I found somewhere that Parity was searching for people and I applied.
It was also like I was talking to I think first Marek and then to Arkady and then someone else I don't know.
Yeah you were also but I think before you there was like someone else.
After every step I thought okay I'm out. No, because I was just talking. I was just talking to Marek. I was just talking to Arkady and it was like I just talked. And yeah, with you it was kind of funny because I think we needed three turns until we finally met. It was like I got scheduled two times and he was not appearing where I was waiting there. Well, there was still this appear.in or whatever where just a link at one point I joined and someone else was already there and said, 'Ah, okay. Then I'm probably leaving, but I'm already waiting here for half an hour.' And yeah, and then in the third one, we met near the Berlin office and then I met Kevin, but it was also just talking. Then I got the job.
At the time, our interviewing was different. You used to have a call with all candidates at the time who were joining. Why was that? You wanted to like get to know everyone or?
A bit, and initially it was everyone. Like when Parity was sort of 10-15 people it was everyone. When we got to 30 or 40 it became just engineers, just technical people. We grew so much in the 2020 sort of time that I was unable to do that, it would take too much time. But it was mainly for two reasons. One of them it's like what do they say? It's like a bad hire can cost you six good hires. So it was very important to me that Parity was a small company and I didn't want to I've had experience with bad hires in the Ethereum days and I didn't want to risk that. And but by the time people got to me, I think it was maybe just one guy in my entire history of 10 years that I said no. So it didn't happen very often thankfully, but I was very keen to really make sure and partly it's because I wanted to get an idea for you know you can get a feel for someone's level of curiosity and intellect in half an hour or an hour just chatting about random stuff. And I wanted to make sure that the ones that I thought were going to be very capable of working with me I knew who they were so in time I would be able to assign them.
I think I was the latest in this group to join in January 2019. I was also broke at the time. I was finishing my master's degree and halfway through it I really wanted a break like I wanted a change and I thought looking for a job and just pausing the degree would be a good idea. And I knew about Bitcoin and Ethereum before but I actually knew very little about the space. So I think when I joined and I was interviewing I definitely felt like the biggest impostor in the world like I was like this is insane to me. I don't know the language well. I don't know the technology as well. I think in like my cover letter for the job posting there was like one for Rust core engineer and there was one for JavaScript engineer and I said my skills are better for the like I know this but I don't know this and if I'm given the time I would work hard and learn it. And I also had some interviews with Marek, Arkady, and Gav and yeah I remember it was mostly talking and it was like chilled interviews.
And let's talk about like some of the things that happened in those early days of Substrate. So you invented Solidity which was the primary language for Ethereum and eventually you invented something similar for Substrate on Polkadot world the SRML as we used to call it and nowadays called Frame. Tell us about it.
Oh yes. Yeah. I mean I wrote the initial design spec for Solidity and I mean this is kind of a fun part of building systems. It's like kind of doing this like super opinionated stuff that you can really sort of use some of what you believe are important expression systems expressing meaning and the same is kind of true for Frame as well. So Frame came about or SRML, Substrate Runtime Module Library.
Yeah. Just rolls off the tongue.
It came about through I mean initially we were doing we were writing Polkadot. We're writing the relay chain and we needed some you know we had a runtime that would interpret transactions and retain some state balances and staking and so it's like okay well we want this to be modular so we make a few different modules and then all the modules are sort of like kind of hardcoded and it's not so easy to like for example make the module that may have a number of transaction entry points all like automatically kind of get collated into a general transaction format. And so it became kind of initially just a way of more easily taking the code that we had anyway and sort of attaching functionality to it like that was sort of based upon almost like metadata based upon the functions names and so forth rooting that in. And then before long it occurred to me that you know obviously others are going to want to write chains as well and they'll want different types of transactions and sometimes they'll want similar parts of the transaction like they may want a balances module or an assets module NFTs module whatever staking and some things they won't want to put in that chain's sort of transaction execution system or state system whatever you want to call it we called it runtime.
There was actually a one-week period where I more or less shut myself in a very small hotel room in Tokyo at the APA hotel for one of the Ethereum conferences there at the time and it would have been I guess in late spring 2018. And I basically figured out this like associated type system that Rust had that you could use the associated type system basically for wholly describing and configuring a set of runtime modules. So I prototyped, got the idea that this was going to work and then actually set about building it like switching out all of the code from what it was which is totally hardcoded into this like kind of softcoded but type correct configuration system and it took a week like it was a big PR. And I was at the beginning of it I was like this is never going to work but you know I couldn't see any reason why it wouldn't apart from you compile times. So yeah, and it got done and I actually skipped most of the conference because of that.
In the early days, we had this declare declare something macros. Not sure if that was you or you who made them or maybe you improved them at some point.
One of your first things that you worked on maybe.
Probably at the beginning I was working with Gav and he was giving me stuff and like yeah when I joined I had no idea about blockchain. I mean, I knew Bitcoin and Ethereum, but I knew them by name, but that was all I knew about it. It was like, yeah, it was all like a magical like black box to me. And we had kind of easy issues at this time. It was like, yeah, do that. And I was doing a lot of macros at this time. So that was also something and I think before I had written one macro or whatever before joining Parity which I copied from Stack Overflow. But then I was doing working in the runtime, manual runtime. Then one of the first things I was creating this metadata. The early metadata was actually like a JSON metadata which we then quite fast replaced with like the SCALE-coded metadata what we have today. Yeah and then I'm doing a lot of these modules of yeah FRAME or not FRAME as it was called like decl_module and decl_storage or whatever it was. I don't remember all the names we had.
The source code for those macros is a piece of history and art to look back at this like where you parse a little bit of it then you pass the rest to the other arm and then this just repeats.
I had this one macro which was generating other macros. This was really crazy but I mean like yeah debugging this stuff was really complicated. I'm really happy that we have these proc macros. I mean they came up in this time I think in 2018 they got it to stable. So yeah that's when we.
And then one of the first things I remember about you Aaron is the launch of one of our test nets actually for cumulus based chains and parachains for Rococo.
Yeah. I actually first of all joined before the first substrate testnet was launched which was called we named them in the same way there's Flaming Fir.
What is the name? Rococo? What does it mean? It's a type of art, isn't it?
