THE UNIVERSAL MACHINE (THE GAVIN WOOD WORLD JAM TOUR FILM)

The point of web 3 is in some sense to transcend beyond national boundaries and create a system of law, digital law fit for a global nation. You shouldn't wait for technology to be delivered to you in a product form. But kind of like with education, you need to go out and you need to seek.

The current Polkadot protocol, the expertise is really very concentrated, centralized. You might say it's my fault for not making a big enough effort to help others understand the protocol. And so one of the sort of main changes of approach with JAM is really to try and decentralize the expertise of the protocol. Technical and intellectual decentralization. It's not something you see very much in web3 protocols really at all these days. And Polkadot is not the worst of them in this regard, but it's also not quite up to the sort of Bitcoin and Ethereum level of intellectual decentralization. And with JAM, I'm really hoping to move it into that sort of family of protocols.

This is a formal specification. It's not a vision document per se in the same way that the Polkadot paper was a vision document. It didn't have very many details in. This is fully detailed and from this document it should be possible to implement this protocol. JAM is a little bit like a marriage of Polkadot and Ethereum. It takes some elements from both. In particular, it takes the sort of scaling machinery of Polkadot and marries it into more of an Ethereum-like paradigm by trying to take it to many different centers of learning around the world. I hope to really make many more people familiar with the protocol. And a core is a logical construct that does computation a little bit like a regular CPU core but spread out across a network of nodes. JAM has, for what it's worth, 341 cores. It's a formal specification which means I'm using set theory, basic formal maths. And so part of this great paper tour is to go through each formula and actually try to re-understand it with people around. But then actually the final value beta prime is informed additionally by a few other terms and see if anyone brings up any important points that need to be addressed.

Basically, we take what would normally be a single coherent piece of code and we split it up into two parts. One part is called refine, the other part is called accumulator. One of the reasons that Argentina, of all of the Latin American countries, is the one that chose to make the JAM lecture in is that people are really very interested within web3 technology for the right reasons. When I say the right reasons, I don't mean like, you know, speculation on crypto tokens or anything like this, but I mean actually trying to use the token to solve real world problems. And I think part of this is because Argentinians actually have real world problems to solve.

It's not that complicated. And I really want to make this thing implementable by coders, by regular people, not teams of 15 that have already implemented Ethereum twice. And I mean, I'm trying to structure lots of incentives to make it worth people's while. It's certainly in my mind worth Polkadot's while to try and get as many separate development teams implementing its protocol as possible. But this is why we've got the implementation prize. The JAM prize is really there to try to bring people into this environment and incentivize them for delivering important milestones in the building of an implementation. There's a 10 million DOT and 100,000 KSM. So, you know, it's in the tens of millions. And each team will get up to, I think, 500,000 DOT and 5,000 KSM. The sort of size that I would hope allow people to sort of take the risk that, you know, spend their own time actually building something. And I would hope it's also the kind of size of prize that gets some academic institutions interested as well.

I don't think there was a particularly significant amount of take up when Parity and I were here the first time. There certainly wasn't in Silicon Valley in general.

Yeah. But no, they were interested. I mean it would have been an excellent investment actually done it.

My hope is that this is still based on what I have to say. And less based on who I am. And what I've done before.

So JAM, how does it look? It's kind of this funny flower shape. So basically like a bunch of petals with this middle core. The numbers seem a bit, you know, outlandish and we're sort of talking a factor of a million. Essentially, the petals function as huge powerful beefy pre-processors that take the external data to the system from the world, pre-process it in some way and then feed it on into the middle, into this melting pot.

While big unions, you know, like the US tend to manufacture their own trust, I think increasingly in the modern world, this is becoming questioned. I hope JAM will be the blueprint protocol for what will be a very scalable trustless computation platform, web 3 platform and be part of a solution that reduces those concerns. I think we can see very hybrid systems forming in these sort of superstate entities whereby the foundation is a classic web 3 blockchain but we have very clear strong providers of last resort perhaps run by national governments or licensed institutions that can under the rules enforced by the blockchain help provide services like things like voting procedures.

The issue with the Stanford talk was that it was only 15 minutes long. Buenos Aires one was I had a lot of time and I was really able to give the JAM introduction and then really deep dive into an aspect. And I think that is the big lesson here is that I need a good 2 to 3 hours time slot to do what I want to do.

SASS is the Polkadot virtual machine execute, like invocation function, the lowest level of the Polkadot virtual machine. So just to give you some idea of its inputs and outputs it accepts some amount of code. It's a Harvard architecture machine, not a von Neumann. Unlike RISC-V, there's a bunch of alterations that we made, minor alterations which are trivial to... When you have an audience you can kind of gauge how much people are nodding and how much people are following and if you don't see those nods then you go a bit deeper into your explanations. It's also very helpful if there is time, you know, if there's time for them to ask questions. It can show you where there is the deficit in their understanding and where you need to go deeper.

Yeah, it's not far off the mark and it's not doing anything dramatically different from let's say conceptual building blocks that are out there in the industry. I think the main difference is that it's putting them together in a particular way because it's part of the same protocol. We can design them with much more efficiency in mind. ETH2 is going to be held back by its desire for stability with the pre-existing protocol. It's also a little bit hamstrung with its fascination over ZK technology and the cost implications from that. Just for scaling, we don't think it makes anywhere close to sense. You're looking at around 60 to 70 million x in cost compared to executing it in a single native instance. In our case, the cost is around 30x because we execute it about 30 times.

We're in Tokyo tomorrow. There's a lecture that I'm giving at the University of Tokyo and it's a useful sort of center of gravity, center of mass for Asia. It's a very cultured country, Japan, very cohesive. Essentially, people seem to read social situations with a lot of conscientiousness. If for whatever reason an area is designated to be quiet, you will find that people respect that. Things work here. The system works here. And I think this is necessary because people live in such close proximity. Tokyo, I believe it's the largest metropolitan area in the world. There's a lot of automation. I mean, you see vending machines in other countries, but not like hot drink vending machines on every street corner. There are restaurants you can go to where you order your thing in a machine. It gives you a little ticket, you sit down at a booth, a hand accepts your ticket, a minute later, some ramen comes out. The idea of interacting with systems and machines rather than specifically individuals is I think very, very natural. And even when individuals do interact, a lot of the time it comes with rules. It comes with a protocol. It's a system. This perhaps is where some of the interest stems in blockchain and crypto. It's an advanced technology. It provides for social consensus and near zero transaction cost and it does so in an automated fashion. And this is the most automated society I've been in.

Today's topic is SASS, the block production algorithm which is basically half of the JAM consensus algorithm. JAM splits time into six-second time slots. I mentioned before cores work on this 6-second thing. A work package is a 6-second usage of a core. This 6 seconds is very thankfully what we're doing with the gray paper is very similar to Polkadot. Most of the hard bits are already done. They're already part of Polkadot. Scope is constrained enough that we can actually deliver it in a reasonable time span. It's basically a year and then we have E prime and M prime which is just the same thing except it's for this block rather than the previous block.