podcasts

Building on the Shoulders of Bitcoin, Mark Wilcox of 21e8

podcasts

Building on the Shoulders of Bitcoin, Mark Wilcox of 21e8

August 2020

Posted by

Jamie Burke

CEO and Founder

As an early investor in Bitcoin and Ether Jamie went ‘all in’ during 2013 founding Outlier Ventures, Europe’s 1st venture fund and platform dedicated to blockchain and Web 3....read more

Jamie speaks to Mark Wilcox, Founder of 21e8 (The Magic Number Company) the pioneering ecosystem of computational data markets – competitive systems that combine real-time content creation with distributed data exchange. They discuss everything from building businesses in crypto, what the computational data market is, and what Mark looks for in his community.

Posted by Jamie Burke - August 2020

August 2020

Posted by

Jamie Burke

CEO and Founder

As an early investor in Bitcoin and Ether Jamie went ‘all in’ during 2013 founding Outlier Ventures, Europe’s 1st venture fund and platform dedicated to blockchain and Web 3....read more

Key Themes:

• What is 21e8?
• How do you create outputs? The process of figuring out the answer, is to take every piece of data and combine it with every other piece of data.
• The real energy costs from computation
• Energy efficiency graph optimisation vs physical infrastructure problems
• Why you can’t build traditional businesses in crypto
• How Proof of Work creates demand for future content by signalling demand
• What is a computational data market?
• Building communities not competitors

Listen on iTunes

Transcript:

Jamie Burke
Welcome to the founders of web three series by Li ventures and me your host, Jamie Burke. Together, we’re going to meet the entrepreneurs, that backers and the leading policymakers that are shaping web three. Together, we’re going to try to define what is web three, explore its nuances and understand the mission and purpose the drive its founders. If you enjoy what you hear, please do subscribe, rate and share your feedback to help us reach as many people as possible with the important mission that is web three.

Okay, so today we’ve got Mark Wilcox of 21 he ate the magic number company working on computational data markets. I know that begs the question, what the hell does that mean? So we’re going to kind of delve into that. Actually, I’ve known Mark virtually for some time now, probably almost as long as I’ve been in crypto and blockchain actually only got to meet in person physically in London recently just before the corona outbreak, and will probably never get to me ever again. So, it’s good to have you on the podcast.

Mark Wilcox
Yeah, it’s, uh, was good to see you, you know, a few weeks ago. And it’s an interesting time we’re in. But yeah, I’ve been kind of based down in New Zealand, kind of working on the side on on a bunch of technology. And the timing is just interesting, as it’s kind of coming to market and coming to fruition and the world is changing a lot. And it seems like there’s a lot of benefits that these types of ideas can can help us with.

Jamie Burke
Well, yeah, I mean, especially if you think you know, Bitcoin came about during the last financial crash crashing crisis. So let’s see what comes out of this one. Also. Yes, you know, you are based in the place where all the billionaires are flying in their private jets to escape Armageddon. So you’re probably doing all right. So by way of background, actually, I knew you when you working at Neary aid, also based in New Zealand, and I’ve always struggled to explain what that company did. It was wonderfully complex, but I believe was born out of a spin out of a global distributed satellite dish system, right? So so I know you’ve got you’ve got some pretty good chops when it comes to solving very complex technical problems. And now that’s being applied in this new startup. So, in this series, we’re interviewing the founders of web three, and I’ve always kind of had a keen eye on seeing what you would do when you left in every aid, and that’s now born out as 28. So maybe before we go into, you know what What is the magic number company and how that can be applied to computational data markets? It’d be just good to understand your background, as I understand it’s a mixture of gaming media and supercomputing, but maybe if you could kind of give the listeners a high level summary of those skill sets, and then how they kind of are applied in your new startup.

Mark Wilcox
Sure, yes. So like my back in originally was, like Communications and Media. I, you know, worked on when I was in university, I did like a lot of Esports stuff, producing content and the community you know, working with people who built websites and built mods and, and ran tournaments and so on. And so I got kind of really a really good feeling for how the gaming business works, how players and content creators added value and, and, you know, created network effects that could actually be sustainable businesses, not just for the sort of, you know, the game developer, but lots of different people in the communities. And there was around the time that Twitch, you know, came about. So it used to be that people would, they would do shoutcast, and they would kind of stream the audio and you would have to do complicated things to sync up the, the, the stream audio and in the game that you’re watching, and then people would record them and they would play them back. And that turned into live streaming as the internet really improved. And then after I finished university, I basically realised that, you know, the whole media industry was just kind of falling apart that didn’t look like a very sustainable type of profession to go into. So I went and started working with startups, this is around the time that you know, and then that was around the same time. That I got into bitcoin. And it was also when GPU mining and FPGA mining were happening. And aetherium was really, really kind of coming onto the scene. And that first startup that I worked at, we were doing kind of location based augmented reality. And we’re trying to deal with a lot of issues around lighting and ray tracing and things in the scene that require a lot of real time data sources, but also like offloaded compute. And the time effect felt like there was, you know, there were all these people kind of putting GPUs in the arm down the street, they should have been a way to call out to them and and offload some, some sort of process. And you know, what one of the, the early kind of use cases that people were talking about with a theorem was why you should be able to rent out your Wi Fi. So as you walk down the street, you could just buy Wi Fi from people and I felt like that was a really, you know, physical opportunity, hey, you have some GPUs and in somehow sleepwalking by you should be able to just, you know, get some low latency compute from them. And it was around this time that I met Alex St. JOHN. Alex was the original creator of Direct X back in the 90s, and Microsoft. And so directives was really the the API platform that led to the creation of the GPU. And the same way that that, you know, Bitcoin led to the creation of schatten 56 H asecs. How TensorFlow API has led to the creation of all these new AI processes. And he taught me a lot about economics, the economics of compute, and how kind of insatiable demand for real time you know, highly, like really, really good looking graphics. And interactivity, was really like the driver of most computing hardware and real time software and So he, when I met him, he was basically in the process of he just moved to New Zealand. He was involved in this project called the Square Kilometre Array, which was a radio telescope and that they’re building in, in Australia. Yes. And essentially, it’s

Jamie Burke
Africa as well, right? I believe.

Mark Wilcox
Yeah. So so it’s a, so it’s basically two telescopes. They operate at different frequency bands. And essentially, they, they have two supercomputers on site where they basically take a huge amount of raw signal data. And then they put that into a big GPU FPGA supercomputer and do a fast Fourier transform on it to take all the little independent signals and produce a, you know, a four, four dimensional, you know, like a three dimensional data set over time. And then they kind of can Model Model back to the the Big Bang, they’re trying to search for echoes in the universe. And the amount of data that they had to process through that system is more than that goes through the Internet and it’s just basically normalised. But anyway, so they had a really difficult infrastructure problem, which was essentially, they had to build this telescope and run it for essentially, you know, decades, like 50 years, and the data rates are so enormous and they’re gonna keep growing that telescope, that the idea of having to move, you know, having to do your computation, and then also move it over to storage, and then load it back over again, was just looking completely infeasible. And so at the time, you know, people were talking about putting GPUs like closer to storage and everything. And we realised that you know, you needed a storage architecture, where the data, the place the data was being processed, and the place that it was being stored, were as close together as physically possible. But while still being able to have all the error tolerance and upgradability and everything you would, that you would get from storage. So rather than having one supercomputer in one big storage system, you would have compute and storage close together constantly working off of one another.

Jamie Burke
And would it be fair to say at the edge, that the way that, you

Mark Wilcox
know, well, this is a weird thing. So it’s not just about the edge or the cloud, it’s a whole system. So the edge was the, you know, the actual radio antenna in the outback, right? And they would attach little mobile processor GPUs with FPGAs on them to that single signal, but then they would have to aggregate all of that. And then once it was in it, once it’s in the data centre, it’s still a, it’s still like a complex graph problem. It’s not just like, well, we put it in the Big Data Centre thing, and we have it there. It’s just as difficult and complex once you get into the system. So in that light, high in order to have it really, really, really high performance, reliable core, they actually have you know, you had to have a parallel like massively parallel distributed system. In order to have the kind of density between nodes that like instead of having one system that is at risk, and then having to back it up to a completely other system, you would have just a single, you know, really large cluster, essentially, that had all the data and it was doing all the compute, but you could take a note offline, you could put it, you could upgrade it, you know, this type of thing. And what was interesting was the actual that, Frank, that accurate data problem is very interesting. So in the blockchain world, people can think like, we Okay, we have the data, and then we need to distribute it. So and then the more you’re computing something on a single node, the harder it is to distribute. But in the radio telescope, you know, the signals that coming from a distributed system, right? And then the process of figuring out the answer is to take every single piece of data and then combine it with every other piece of data, and then have your output right and because the data was distributed in the first place, as long as you cashed the new code, you didn’t have to worry about the final results being lost because you could just recalculate it and you would just catch, catch the results. You know, as you went through the system like that. What are you? What are you optimising for you’re not optimising presumably for security or latency. It’s, it’s something different. So you’re, you’re optimising just for pure power efficiency, essentially, like, Edison scale, like a supercomputing. And this is, you know, this is what everyone in supercomputing is dealing with right now, anyway, is that the energy costs is not from the computation and you cost us from moving the data, you know, over a wire. So you literally want to get as in when you’re actually doing the hard work to come to a conclusion that is deal with the huge volumes of raw data. And so you need to adjust as few ways as possible. And so there are people even now there are people building gigantic chips that are, you know, like a like a Pizza size GPU purely in order in order to, you know, not have any cables go back and forth. And so that was a big driver really say like, hey, the cost of computing is not the power efficiency of chip cost of computing is the is the network and and the the risk associated with you know, running something on a on a really big, expensive chip, you know, the fault tolerance thing you need, you need to be able to get out. And so once you started thinking in that way about how to solve infrastructure problem, like it was not a physical infrastructure problem anymore. It was a energy efficiency graph optimization problem, then suddenly like even to think about how to run that type of system, you need a GPU, right? That’s the like, please graph optimization problems are what GPUs are designed to solve. That’s like a GPU as an asset for these kinds of graph optimization problems. The GPU is really a nice, that’s what it’s like really kind of Designed to do extremely well as really high throughput, large volumes of parallel processing, and recognising patterns and quickly connecting them. And that’s why, you know, that’s why they use for, you know, all artificial intelligence, rendering,

you know, large kind of real time analysis work. And so when you think about these ideas of routing problems of, of scalability, and centralization and decentralisation kind of tensions, then at this at the scale of, you know, an economy, it’s really impossible to even reason about these things without applying the without using the GPU without using the same system to understand itself in the world of HPC. You know, these questions of scalability theoretical academic architecture design problems, they’re hard problems with numbers to say like, Hey, we have a supercomputer where we need to design a more powerful one. And these are all of the constraints that that we have on the system. So what’s happened in the last part, you know, a few years is essentially in order to continue to scale because the scaling constraint of a supercomputer is not the amount of compute, you know, you can just buy as many servers. It’s just a construct of money, when you’re actually trying to solve a problem and a supercomputer that constraint is the is the IO is the communication between all the different nodes. And then the, the, the way that you know, you structure you design your programme, and the type of problem, you know, the inherent limits to the problem that you’re trying to solve. But, but especially, you know, practically a lot of it is how you actually structure the code such that the data is as independent as possible. So when you’re when you’re not needing to move data over a slow network like in between nodes, then you can and you can stay in memory. But you can be confident that the whole, you know, workload is doing its job properly and staying in sync, then you can, then you can achieve pretty good performance numbers. But what hell has happened as, as the amount of data that is being used in these systems, and the amount of data that has been generated, like within a simulation has just exploded, is that in order to keep data off of the network, you basically need to continuously train a model on those data sets, then you can understand what is the topology, what are the upcoming data dependencies that you need to be aware of so that then you can prioritise the traffic.

Jamie Burke
So obviously, you work On a lot of this stuff in parallel to, I guess a personal interest in, in Bitcoin, crypto generally your work there wasn’t kind of directly linked or was it What was the what was the relationship between your interest in, in in Bitcoin and crypto and, and your day job?

Mark Wilcox
So, like, I mean, as I was saying, when I got into it is that, you know, it felt like there was a both a physical, you know, computing energy latency problem to solve that needed some GPUs over a network and needed interactivity and this type of thing, but also some sort of market economics problem to actually make it possible to connect things up and at the, you know, when you’re dealing in supercomputing, where, you know, a single system designed to solve a very specific problems could cost 10s of millions dollars. It is just the economics problem of how do you wire things together such that you know, you You can get the answers that you’re looking for, with while using as little energy and getting as much simulations of jobs through the system as you can possibly get for your money. So it is it once you’re at a certain scale computing problems or just economics problems, and we really had to model that we had to understand it in order to even make it possible to run these types of systems. So I was thinking a lot about, you know, it felt like there was an obvious connection, it made sense that if you thought about it, about these systems and these type economic systems, you know, that could have their own accounting systems and have their own way to be resilient and account for error. Then you could that could be very useful on a wider scale, you know, like we weren’t trying to build a company that just sold one thing to to one supercomputing business. We were trying to build a general purpose, technology and I think what ended up happening is Why in the early days, I expected the, you know, the crypto market to evolve in this direction naturally, I think we kind of got distracted by a lot of like financialization and Icos that really took away from what was happening and in around like 2013 and 2014, with, with mining, because it became just not about mining at all. There was obviously, like a mining boom with GPUs. But there wasn’t really drive that wasn’t a driver of the network value. That was just people playing catch up to, you know, the whole whole kind of bubble and 2017. So what we saw was, you know, a whole bunch of interesting exciting projects and ideas that we didn’t have the right technology, architecture for couldn’t scale, and the infrastructure and everything that went into this generation of, of infrastructure. Have you know of the network and the people involved hasn’t really created a new platform for people to build these applications on? We have we have a bunch of interesting projects and concepts that have now been turned into code, that there’s an opportunity to get to get running again, you know, get running on a live network

Jamie Burke
with the billions that have gone into this new infrastructure around crypto generally, do you think any of that is is useful, or will be useful has the potential to be useful in the context of some of the problems that you were trying to optimise for in your previous role? Or do you actually think most of it is kind of largely just this experimental sandbox? There’s some good conceptual learnings but there’s, there’s no real tangible technology that could be leveraged.

Mark Wilcox
So these two real pieces of infrastructure Right, well, maybe three, so like who can just put the financial life tooling, speculative market infrastructure, which is useful for valuing things, we can put that aside, let’s look at the hot infrastructure. Right. So in terms of consensus, right, we have a six, we have dedicated machines that are listed there, who their function is to solve, you know, proof of work problems. And and that is that you can take that infrastructure and really use it for anything that’s not consensus, you have the definition of like the use of reasonableness of proof of work is in building consensus. The other half, you know, side of the infrastructures in this general purpose compute that has been running other other blockchains. So proof of work for a theorem, also, like people using GPUs to then like, you know, run compute jobs and things have come from a theory because the theory was, you know, being a decent way to write some general purpose thing, put it on the blockchain, get people involved, you know, there’s a decent amount of infrastructure out there, but that those systems don’t have the Same economic driver, that proof of work has that, you know, ends up with millions of GPUs are on a theorem in the first place. So although there’s been a lot of, you know, because these things are extremely, you know, like valued at a really high number, there is a lot of infrastructure that is, has, you know, been put in place, but most of it is all there in the consensus layer. So you need to essentially, in order to figure out how to take advantage of this investment that has been put in, you need to start to figure out a way to formulate more valuable consensus problems to solve

Jamie Burke
from proof proof work.

Mark Wilcox
Well, like, proof of work is the is the way that you account for, you know, solving a consensus problem is still a human, it’s still a social economic problem. It’s just that that is the infrastructure that’s available. So in order to benefit from this whole ecosystem that you need to formulate to, to, to turn it into something more valuable, you need to find ways to continuously use that existing infrastructure and put it towards something that is even more profitable even an even better speculative investment essentially, to build consensus over then than just running in the base blockchain.

Jamie Burke
Right. So you kind of need this feedback loop in a way. So, you know, we have this financial layer, it’s it’s created enough value or driven enough speculative value, that it’s managed to kind of mobilise bottom up this this compute global net.

Mark Wilcox
So compute layer isn’t really the right way to think about it because all there is is a ledger, right? And there’s a network of computers who are building and sharing that ledger. That is just data which has all of the compute the networking, the basic hardware attached to it. And then the financial stuff is, is more institutional infrastructure. It’s, you know, systems of governance pools of capital for managing control over this ledger. But there is no real layer because that’s one is a one is a, like a social, economic, you know, idea of Finance. And the other is just a file that people are sharing through the computer. And so at the level of the actual ledger, you know, there’s no difference between money and data.

Jamie Burke
Yes, it’s all data. Yeah, it’s all information. Okay. So this, you know, led you into your new venture. And I remember saying to you, when you in your previous role, you know, let me know, when you move on to your, your next thing And if you if you create a startup, so here we are, I guess you’ve been working on this particular project 21, eight for a year or so. And I’m sure intellectually, maybe before that, but in terms of actual code being written, it’s

Mark Wilcox
Yeah, so the company was founded in August.

Jamie Burke
Okay, so even.

Mark Wilcox
And we know, we launched the first piece of it in January,

Jamie Burke
right? And it looks like it’s, it’s, it’s getting some good traction. So you refer to it as the magic number company. Can you unpack that a little bit and then we talk a little bit about the, the application or applications that are possible when you can enable this computational data market? And obviously, the linkage there is, is that you’ve been presumably thinking about how you can create even greater economic incentives and markets that can drive presumably the this next wave of speculative value.

Mark Wilcox
Yeah, so you know, magic number company. It sounds like cryptic and, and, and, and things and you know, people often put that on onto the idea, which is natural, like with with crypto, it’s you know, things are encrypted, but a magic number and programming is basically just a protocol identifier, you know, when you have a file on your on your file system, it will have a special flag at the, at the front of the file in the metadata. This is you know, hey, this is a JPEG or when you have a networking packet comes through it will, you know, say hey, this is this type of networking packet and, you know, in, in your code, you will you have this kind of constant that you replace with some name that makes sense, or, or or in a in an in a UI, then you would, you know, represent that and in a way that is human meaningful. And that’s a that’s known as sukkos triangle, which is basically a tension between a machine readable ID And the human meaningful name. So how does kind of when it comes to crypto, this is actually really important because what we’ve what we’ve seen where the theorem is, you know, people will have a new idea for a for solving some sort of domain specific problem. And so they’ll create a new protocol. They’ll create a new contract, and issue new tokens. And essentially, you know, fork the community and lose the network effect that came from people solving all these other problems before. So in order to really be able to build higher level, you know, but more specialised systems that retain the economics and the network effect and the shared infrastructure as Bitcoin, we essentially need to be able to start to create new protocols that share the same new systems that share the same underlying, you know, protocol. But can have different interfaces and different names and different brands and so on, attached to the same system. So when we say so the company is called 21 eight because essentially, protocol identifier for the technology. And it also means that as a business, we can go out there and we can create a whole suite of like other businesses or applications or a community that is using the same infrastructure without constraining the kind of the way that they think about the network. So it provides this nice bridge between the basics of the underlying actually that exactly how the underlying data structure works and, and the opportunity that then emerges from turning that into something human meaningful.

Jamie Burke
So you’ve created this technology as I understand it, you’ve kind of released it into the wild and It’s kind of getting some organic pickup. And its success looks like it’s in its simplicity. So as it stands right now, there isn’t a commercial entity necessarily that you can invest in. You’ve open sourced this technology. And already markets are starting to form. And the idea is that, you know, there could be several marketplaces that form around computational data, leveraging proof of work. Can you talk about the, what you’ve released, how it’s starting to be used and potentially the vision long term vision for that?

Mark Wilcox
Yeah. So essentially, so there’s a couple things to your question. The first one is basically just you know, what is it? So what we’ve done is we’ve sort of generalised proof of work. We’ve said that okay, we actually have, you know, multiple Bitcoins. We have lots of different blockchains out there. Can we start to think of like that whole market just as one, you know, one ecosystem because it really kind of is there’s lots of people trading between different ages. And when it comes down to it, the the real driver and the basis for the whole ecosystem is proof of work. It’s this, you know, huge infrastructure, this huge capital investment, that there are people who, you know, a lot of other people out there who are putting in just as much time and energy and money as you and that means that it’s a it’s a market that kind of actually exists that you can rely on to, to invest then to try to, you know, build companies and, and so forth. But we’ve had a real hard time and in turning that into something useful, because we end up kind of having a false start every time we try to build a business on and that’s because the way that we build businesses and in the software world is from trying to like you know, build a startup have equity build a platform, build a brand, get users and somehow turn that into a company that is more, you know, been more more valuable than just its than just its revenue. And that really kind of doesn’t work. And in crypto because the base infrastructure you’re relying on is the thing that is that has the network effect that had the, you know, users are really stuck to. So if you have, if you try to kind of capture your users in an assess platform kind of thing, then they actually just have a really easy way to exit. So we have this tension in the market between centralised wallets and centralised exchanges where they don’t really want they want to keep the network effect on their system and not have a network effect emerge around the blockchain. So you’ve had real kind of constraints and restrictions. Put on these industry players to, you know, scare you into not not leaving your coins on chain or say like, hey, you’re gonna get hacked and this kind of thing. Or they’ll say like, Hey, you have to keep your coins in a cold wallet, don’t put them online, don’t put them on the network, don’t spend them because that empowers these companies to, to keep to keep trading. So that’s a business model and economic incentive thing and attention between, you know, the Fiat world, the Fiat, kind of, you know, equity world and the blockchain world. But that thing that they really are still relying on is that same underlying proof of work network effect, like without the, without the miners who were there to attest to the currencies, you know, none of these guys would have any business because the networks just wouldn’t exist, you wouldn’t even be able to deposit coins into an exchange. And so in order to really think about, you know, how to how to build more advanced useful stuff, you really need to defend and build out that core infrastructure and that core thing that, that lets the market exist in the first place. And that is really proof of work. So proof of work kind of solves a really interesting problem, which is, if you if you take, like a content addressable storage system, you know, you know, we’ll call it a data market like and, and basic version is that you have a little discovery problem, you know, you have to know what content has already been stored in order to, to ask for it. And, and all and you also need to, if you if you’re storing something new, you need other people to know that that content exists, right? And so there’s something like ipfs it can only really take it can only really handle like a cold. We’re not a cold data set, but but a historical data set something that already exists, and proof of work what it really did, was it Created demand for future content. And it did that by signalling demand. So essentially a miner, you know who’s doing a lot of proof of work, they’re saying, they’re trying to really put out a strong message to say, like, Hey, I’ll pay for some service or some content that doesn’t exist in the market at the moment to pop up. And, and when you look at the actual use of a checksum, and the use of proof of work, to game it, it becomes a really interesting, it just naturally forms this information economy. So you get people who are able to understand that there’s some upcoming content, some upcoming data in the future, and you should pay attention to it. And you should maybe participate in the process of producing it, because you can see that it’s really, you know, valuable to these people who are Who wasting a lot of compute on it. And so if you really treat compute as the underlying currency which Bitcoin does, then you’re able to understand you can connect the dots between a actual piece of addressable content and the future and the amount of compute on the present. So, that is kind of at when you when you take that concept, and you generalise it, that’s what I call a computational data market. Because it’s a data market for content that is being computed in real time. So you don’t know the

checksum of the actual data that you’re trying to get. You don’t know the results yet. But you can check some the state of the application you can check some the code, and if there’s enough of a strong economic signal behind it, then that can create the space for people to come into and to You know, make that prospect a reality. And so

Jamie Burke
what’s the thing that So, I mean, I get I get the principle that you know, we need everybody needs proof of work to be economically viable or more viable and that you feel that this natural extension of how GPUs used beyond you know, current mining activity towards media and the general information economy. Yeah. And so, how does how does 21 eight what sorry, what does 21 eight bring to catalyse that to make the shift for for miners to then act, it’s a kind of focus that power around your content generally.

Mark Wilcox
So if you think of, you know, anything If you have the base ledger, you have the consensus layer of the whole network. And then you have contracts that you can deploy on to it, you know, you create other higher level currencies. So 2028 is basically this idea that you can put different protocol identifies on a proof of work, and then create a sort of, you know, application specific chain, an application specific data structure, but, because you can just pay to use the underlying ledger, then you don’t need to resolve reimplemented all the consensus, like all the data availability, sorry, in data distribution, parts of the consensus, you just need to get agreement over the state of like a contract and that’s what lets us kind of do the design a blockchain if you like, like create a market To compete over doing some application specific compute. Now, if we think about the use of a GPU, you know, a GPU is designed to render graphics, it’s designed to do media processing to do, you know, huge parallel, now, you know, data, ai processing and this kind of thing. So in a market, you know, it would make sense that all of the GPUs were trying to do the thing that was most valuable for the whole market instead of designing currencies that then consumed a useful resource for the purpose of consensus when you could have a sort of own stake much more energy efficient, you know, better at, you know, more accurate measure of doing that. And so it’s not about using GPUs for proof of work. It’s about using GPUs for work, right? And then you can compute a proof of that useful work that you did in a standard measurement. way. So rather than think of, you know, 21 eight as a way to mind things by hashing them, it’s a way to think about the creation of new of new assets of content of, of a testing, you know, reputation into things that gain trust and increase quality and the network, converting those into into, you know, registered assets on chain. And then using the same proof of work economics as the underlying ledger, to then extend up to the transaction to the to the level of transactions. So miners are kind of looking to not just make the, you know, make profitable block and make a bunch of transaction fees, but actually participate to the maximum extent that they can have with their expertise and their infrastructure and they read resources to make the economy as valuable as it could possibly be. So when you get these weird, you know, situations where you have Silicon Valley funded startups who are looking to capitalise on equity, it actually is in tension with a miner who is wanting to have something sort of like an equity or or have that, you know, the same value that you get from technology, you know, in their own in their own network, in their own platform in the infrastructure and then charging, you know, Silicon Valley for

Jamie Burke
so how does this how do these markets form so you put this thing out in the wild now, how are you seeing it being being used? I saw that the POW market was was put up recently was that you guys century or did did the community just throw that up and and how are you seeing that used? It’s

Mark Wilcox
an example of This working in action. So, you know, the problem with a lot of the way that we’ve approached things in the past is that, you know, seeing software development or protocol development is separate from the market. And it’s really not. So if you kind of raise a lot of money, and then try to put together a team that works for a year on a protocol, then you’re really the market is not even getting the ability to contribute the ability to speculate, and the ability to use the work that has been done. So it’s much if someone in my position is trying to introduce, like something that levels up the whole ecosystem. It’s more important for me to give the opportunity to write code to other people. So I put out essentially a demonstrate I didn’t, you know, push code out to begin with what I did was I just demonstrated by mining, a few proofs of work, that there’s something interesting going on here, right? And then I was able to engage people have a conversation about, about, you know, the incentives in economics and where value comes from. And then that just cause people to kind of click, and then build something of the rock, you know that that can generate revenue. So there’s a guy who has a product called business v. And it’s essentially a website that you can purchase essentially a hybrid of, you know, Reddit and Patreon. So you would post some content and then you would have you have to pay five cents 50 cents $1 to unlock that piece of content. And he understood the benefit of being able to say, Hey, this is the most valuable content in a way that tied to proof of work. And he built that feature into his app and then built a POW market, you know, website to kind of see what was happening on chain And he was using a the mining puzzle that came from another guy in the community who understood and he has a basically a distributed uptime testing product. So he would use a blockchain to put up a reward. And then you could run the app on your phone, and you would ping, you know, top 500 website and then collect the they would aggregate all the ping data from different parts of the world. And, and, you know, sell it to any enterprise who wanted a, you know, uptime distributed, a real world distributed data centre, uptime tested website. And so he understood the benefit of having a puzzle so that then you could have, you know, people out there run an app on their phones and make a little bit of Bitcoin in order. You know, it’s a process of collecting and selling information. And so essentially, you know, those two different things in the market where there was someone who was interested in selling information and and someone who isn’t what someone who is interested in helping people sell their information. And someone who is interested in helping people collect information that basically, you know, understood different sides of market about the same time, and they were able to, to put something together.

Jamie Burke
And so as a founder, one of the things, as I mentioned, that we’re trying to explore with this podcast is really the founders and what motivates them and how they’re you going about bringing products or, you know, foundational technologies to market. So in this instance, you know, there isn’t a, you haven’t done an Ico you haven’t done any fundraising for this. There isn’t at least an hour, a commercial entity. It’s all been open sourced. So what, firstly, what motivates you to do this? And then secondly, how do you as a founder participate in the economic upside that’s generated

Mark Wilcox
so I mean, there is a conditioning into the eye of a business, right. And so this is ion. And that is, you know, that is all of the branding work and the in the community development and, and the partnerships and relationships and things that I’m building, you know, go into that company, that’s my company. But the nature of the economics of the system mean that instead of inviting other people and hiring them and raising lots of money and selling my equity, it’s, it’s much more effective for a whole ecosystem, you know, I can put in the same way that in a regular salad, you want other people involved, so that you’re the founder can capitalise off of the idea, not just you know, the amount of work that one person can do. So, instead of trying to have a bottleneck or a single entity, I just encourage other people to if they want, you know, to invest, if they want to participate in the ecosystem by building things, you know, just register your own entity, register your own business, and let’s do business together. Right. And as long as you know Like my job as a founder is to, to really paint a picture of a big ecosystem and play the role of a leader. Right? And like the CEOs job isn’t to sit there and see, you know, Steve Jobs isn’t sitting there doing the hard engineering work or baking an iPhone, you know, he has people that, that, that, do that for them and for him and get get paid to do that, right. So by being able to create a market where lots of people can participate, and and, you know, capitalise off of their own contribution, especially in a way that they might see as being more fair, rather than having to deal with like a management hierarchy or having to compete over access to equity, that actually makes the whole ecosystem work much better. Otherwise, I would get in, you know, trapped and get into the same kind of trap, where I would be taking, I would be putting my attention away from the network. I’ll be putting my you know, my into hiring people to work on site. And then I would be forced to kind of turn the network into, into a customer user base that I was trying to hold on to. Here, it’s it’s much more open. And essentially, everyone can take their own risks and make their own decisions. And when you need to move as fast as we need to be moving, there’s not really time to sit around and interview or, you know, ask for permission to go and build stuff. You know, things are popping up on GitHub all the time. We really want people to understand how this stuff works. And if there’s something that they think that they can contribute, come along and do it, and take advantage of the fact that there’s lots of other people who, who can benefit from this technology benefit from this ecosystem to capitalise things according to their real value.

Jamie Burke
Yeah, so but in terms of the structures are you structured as almost like a sole trader and you In engaging to commercial agreements with other entities, but how do you how do you participate in economic upside? I still don’t get that what is the instrument or the asset?

Mark Wilcox
Well, it’s essentially, you know, my my reputation and the network effect. So in the future, you know, if as the network grows in terms of its economic volume, right, that will be worth something. And so they’ll be, you know, my time is valuable. And as long as I’m creating value in the wider market in the wider ecosystem, I won’t have any problem getting paid for whatever it is that

Jamie Burke
is the premium on your, your time and your brand that you’re Yeah, you’re investing in by putting this thing out there. And the assumption is, at some point, there will be perhaps multiple ways that that may turn into future commercial value. But this this particular stage that isn’t your concern, your concern is building The community around the technology allowing them to apply it to their problems to have that market form. And presumably, in the absence of, or perhaps in between that gap between now and when that happens, you pay the bills, how?

Mark Wilcox
Well I mean, I had I, that’s my problem, right? But the main thing is this, you know, this value today, there’s, you know, a little bit of money that was going back and forth. So, you can obviously, you know, you can speculate and be exposed to Bitcoin, if, like, if I believe that, that this will make Bitcoin more valuable, you know, then that’s one way to think of, right,

Jamie Burke
that’s what I was trying to get, but

Mark Wilcox
it’s not really in terms of speaking Bitcoin. It’s about you know, just playing a long term game, right. So in the short term, it’s really important that I’m not trying to compete with my community, otherwise, they’re not a community, that’s my competition and they’re gonna go off and do something else. So I need to create opportunity I need to get create space for others. people to work towards the same idea. Right? And defer the same way that a startup differs. It’s payday until exit, right? Like I’m, I will just defer. Yeah, you know, Capitalising it until someone is someone can take the whole reins for me and, and and, you know, put an even bigger mission on top of this infrastructure that I’m putting together.

Jamie Burke
Yeah, I mean, I think it’s really interesting. I’ve been talking about this quite a lot recently, the idea that there is a generational shift, which I think we’ll be compounded by what’s happening now with the after effects of what’s happening now with Corona in you know, the future of work, and how people like yourselves you know, top up and coming engineering talent are unlikely to go into Corporation and increasingly unlikely to follow The classic Silicon Valley, you know, venture startup approach, but instead look to create value within within communities that could be, you know, relatively small clan like formations that you see in dows. Or something much larger, but it definitely feels feels like there is is a is a real generational shift happening there. Look, it’s been fascinating talking to you. I could go on for for another hour. I think we’ve already clocked well over 45 minutes. So it’s going to be really interesting to to track the progress. So maybe just for the viewers, how can they How can they follow progress? What are the URLs give us a Twitter handle and anything else that you think that they should be paying attention to, to see the momentum that you’ll build

Mark Wilcox
here? So look up 2028 is a hashtag on Twitter. And you can you can mine it. So there is a 28 miner puzzle. That’s a the the default miner is in JavaScript. So you’re going to have to do some work to optimise it. Other people already have more optimal ones out there and running. But it’s an information market. So the value is really in content that you’re deciding to, to get an interest in as opposed to just, you know, outcompete other people without really thinking about it. There is a kind of growing community, especially on Twitter, but and other places as well. Yeah, but the one of the weird things about an information market is that information is not free. So in order to really participate, you have to pay attention. You’ve got to think for yourself, you’re going to think about how you can contribute, and there’s not just a token that you can go into exchange and buy.

Jamie Burke
Great Well look, you know, the reason why I got you on here alongside, you know, some seasoned and like well established entrepreneurs in this space is because I have always believed you’re gonna you’re going to create some interesting things and I think it’s great to see this this first thing coming off the ramp, I’m watching eagerly and it’s testing me intellectually as to as to what’s possible, but I’m I’m really enjoying seeing the progress that you’re making. So thanks for coming on, and I’m sure we’ll speak at some point in the future. If you enjoyed today’s podcast, please make sure you subscribe, rate and share your feedback to help us reach as many people as possible with the important mission of web three.