Illia Polosukhin: Building Near’s Blockchain, Embedding AI

Illia Polosukhin didn’t set out to build a blockchain protocol. Ever since he was a 10-year-old in Ukraine, Polosukhin has been fascinated by machine learning and artificial intelligence – he thinks it might have started with seeing the movies “The Matrix” and the extremely underrated “A.I.”

And Polosukhin was smart. Even brilliant. So in 2014 he joined Google Research, where he worked on developing the models and tools that would eventually feed into artificial intelligence. He co-authored papers and he moved the science forward. But he wanted to do more. “Google has a lot of benefits, and it’s a great place to learn, but it moves really slowly,” says Polosukhin. “I wanted to move faster. I wanted to build a startup.”

Illia Polosukhin is a speaker at this year’s Consensus festival in Austin.

So in 2017, along with Alexander Skidanov he co-founded NEAR.AI, which, in the beginning, was completely focused on artificial intelligence. It had nothing to do with blockchain or crypto or non-fungible tokens (NFT) or HODLing.

That was then. Now, of course, Near is one of the most prominent blockchain protocols in the space, and claims to have 25 million accounts and 446,000 daily transactions and a flourishing ecosystem of decentralized applications (dapp) and projects. It is one of several smart contract projects trying to make a dent in Ethereum’s current supremacy. (Solana, Polkadot are others.)

But maybe this is just table stakes. Polosukhin has far bigger plans. “We want to hit one billion users,” says Polosukhin, who opens up about the journey that brought him from artificial intelligence (AI) to crypto, why Near is launching a “Blockchain Operating System,” and how AI once again could be part of Near’s future. Bonus? He shares why, in his view, blockchain and AI just might save humanity.

Interview has been condensed and lightly edited for clarity.

So when you started, Near had nothing to do with blockchain.

Exactly. It’s purely AI. And as we were doing our AI thing we tried a lot of stuff, but nothing worked particularly well. We didn’t have hundreds of millions of dollars of hardware to train models on, like some companies right now. But what we did have was a crowdsourcing system.

It was a system that allowed us to find students, specifically computer science students around the world. From China, Russia, Ukraine, Poland. Countries where students are from humble means, and they do work for us for cents, pretty much. So we would need to pay them. And it was actually hard to send them money. It’s really hard to do through a traditional financial system. And that’s where the blockchain came in.

This was before the Lightning Network, before a cheap way to send transactions, right?

Yeah. The things that were around were IOTA and EOS and stuff like that. We looked at all that stuff. And we’re systems engineers. With IOTA, we actually had a way to break it. And EOS just didn’t work. EOS had a lot of great ideas and had complete utter [crap] implementation.

So we ended up brainstorming, how would we solve this problem? And we looked at Polkadot and Cosmos as well, and they’re too complex. And so we thought, OK, well, we can try to solve this. And so that’s how NEAR.AI became Near Protocol.

How big was the original team?

We went from three people doing AI to nine people doing blockchain in the span of a week.

Then you guys grew like crazy. What do you pinpoint as the reasons for your growth?

We did a few things very differently. For one, we talked with about 300 blockchain projects. We literally just talked with them and asked, what are the problems you have? What are the challenges to your growth? So we had a very good understanding around the product side of blockchain.

And we actually built a developer environment first. We called it the Dev Node. And the first app on Near was built in December 2018, in an internal hackathon by our team.

Part of our team started focusing on the developer experience right away. They didn’t need to wait for everything to be done. There was already an environment you could build and iterate on. We ended up implementing five different consensuses and three different charting designs. We were trying different things, implementing them and seeing how they behave.

So you could experiment and kind of throw different things at the wall and see what sticks?

Exactly. This gave us some feedback and we could improve things. We weren’t just sitting in the lab, trying to develop. We were at the conferences, in the hackathons.

Almost like beta testing.

Yeah, like proper alpha testing. We actually had people coming to our office to build on this stuff, and we gave them pizza.

What would you say was the next inflection point for your growth?

I mean, the next milestone was that we raised actual money, which was interesting timing because this was in March 2020.

Right as the world was melting down!

Yeah. Then in 2020 we started what we called the Stake Wars. This was a way to engage validators and build the validator community. A bunch of projects actually came from that interaction. We pretty much got validators to start running our code in production, getting out all the kinks. Then, in parallel, we set up what we call the guilds. These were micro-communities where we would innovate, so we kind of decentralized the community. And all of that got us into when we started running mainnet.

And then you guys went on a tear. But what has the past year been like? Realistically, what are the challenges of crypto winter?

I think kind of high-level, right? Let’s say we have $100 million in Near. If the price is $10 or if the price is $2, that makes a difference in how you allocate capital.

You cannot say it does not make a difference. It does. And so at the same time we did get a lot of momentum from all of the efforts we did. And to be honest, a bunch of them have failed since then, right? Some of the companies closed, they didn’t raise money, some of our hubs actually closed as well. So it’s not like 100% of everything worked out. Not everything is a unicorn.

Sure. You think of it as a portfolio of projects, right?

Yeah. It was a portfolio approach. The other thing is that we were shifting strategy. The way I look at it, the Near Foundation needs to shift strategy every six months.

That seems frequent. Why?

Part of it is because crypto’s changing all the time. But also part of it is that we’re kind of going into different stages of the ecosystem, right? In the beginning, you have just the initial team that’s building stuff and nobody else.

Then you seed maybe 10 developers. They build some interesting things, but they’re not yet production-ready. Then you seed entrepreneurs, and they build something that can actually raise money. Then you start seeding funds that invest into entrepreneurs. Then you start seeding a product lab to build more. So for each one you’re kind of changing.

We had a lot of brands that we were seeding [through grants], up to May. We used grants for integrations [with other projects], such as, how do we get the tax tool to integrate with Near? We need Near to be supported with tax tools so we can file taxes, for example. That’s all about building out the basics, like wallets.

But in the next stage it’s about bringing in use cases with real world stuff. And this is where we switch from grants to investments. The next stage is about, how do we bring enterprises [that] can actually build something for a bigger user base?

Stepping back even further, what do you see as the endgame for Near? What’s your ultimate objective?

Well, let’s not say “endgame” goal, because the endgame goal is all the internet runs on Near. [Laughs.]

Okay, maybe not “endgame,” but what’s your realistic goal?

We want to hit one billion users. That’s a very concrete goal. A billion active users.

To clarify, do you mean one billion users on Near, or one billion users in Web3?

So that’s an interesting question. And this gets to something that we’ll be releasing soon. But in general, how do we give control of assets, data and power of governance to all the people?

How do we give that? So that’s Web3 more in general. How do we get there? Well, we need infrastructure that can scale. We need it to be easy to use. We need it simple. We need to make sure people don’t lose their money. Things don’t get broken all the time. You need to solve for all that. Simple, secure, scalable. So that’s part one.

Yeah. Should be straightforward, right? [Laughs.] And you obviously need people to be building on this infrastructure. You need use cases and applications. This is why we partner with projects like SweatCoin, for example, which has 120 million installs. Right now they’re running at over 700,000 monthly active users, and they’re bringing more and more users from their Web2 user base. So that’s a use case.

I’m still confused if you meant 1 billion Near users or Web3 users?

We set out to build something that’s usable and enables Web3 for people. And the first part was to build a protocol, because from our perspective all the other protocols don’t satisfy at least one of these three properties. [Simple, secure, scalable.] And none of them satisfy simple.

Then we built all the developer tooling. We have JavaScript. We have Rust [another programming language]. We have EVM [Ethereum Virtual Machine] run as a smart contract on Near called Aurora. So you can build really complicated applications and just run it on Near. Like, you can actually run Bitcoin on Near.

OK, so that’s done. Then, over the last part of the year, we started thinking about user-focused products. Then one of our core original engineers builds this really cool technology that allows you to do composable, decentralized front ends.

The idea is that you can have pieces of front ends, the source code for which is stored on a blockchain. You can compose it on the client side inside your browser on an app dynamically based on what you’re looking at. This enables you to create this new presentation layer for the blockchain, a front end for a blockchain.

There’s no centralized server. It all serves directly from the blockchain. It can be linked to the smart contract. It can be composed from different pieces.

This allows us to not just provide a protocol for smart contracts, but we can actually have a product for end users that we present, that has all the applications inside it. This creates easy user flows. And inside it we can build all the common features that you need, right? You have search, you have notifications, you have all of the pieces for onboarding that you need from a product side. You can have social.

What’s the timeline on this?

Exciting. What’s it called?

Well, so the whole concept we’re calling the “blockchain operating system.”

So we’re moving away from just being a blockchain to being a blockchain operating system that allows you to kind of use any blockchain. Underneath Near you can use an Ethereum app or you can use a Canto app or whatever. So we’ll have the [Near] presentation layer, but on the back end the smart contracts can be on any chain. And then we are going to be linking everything through middleware to make the experience unified.

To bring this full circle, your career started with AI and Near started with AI and now AI is all the rage again. You must have thoughts!

AI is a really powerful force, but what we don’t want is it to be controlled and biased by a single company. That’s not a good thing for anyone.

What we need is some form of community coordination and governance and incentives to continue building this in the open and de-bias it over time. So what you need is neutral crowdsourcing.

Let’s say I’m an open AI. I dictate what needs to be crowdsourced and how this needs to be done, right? Naturally there will be biases. So we need a crowdsourcing that’s on-chain. You need model architectures that are also open source, and people can collaborate on them and work together.

I’m guessing you might have a solution for this?

My co-founder, Alex, he’s actually been looking back into AI. And so in December he started a company called Yasna to focus on AI powered by blockchain. It’s using blockchain for this community coordination. We’re going to be creating a data set using these approaches for alignment of large language models, with Yasna kind of running the process.

What excites you about AI and blockchain? Why is it more than just hype?

I think AI right now will become a tool to break a lot of the current societal systems. So you can [distributed denial-of-service] the legal system with frivolous lawsuits. You can create fake candidates and try to get them elected. You can create fake companies and raise money for them.

I’m not even talking about the more sophisticated stuff. And so our plan is to organize a hackathon to get people to break all the stuff and then suggest solutions. And the reality is most of the solutions will use Web3 because you need authentication, you need identity, you need traceability, you need all those pieces embedded across the whole spectrum of our society.

Like even like the articles you guys write, you need to have cryptographic signatures on the articles to understand that they have not been modified by somebody else. Or like me cosigning it, that you wrote the article and I agree with it. Stuff like that.

So basically the hackathon would be to “white hat” the problems?

We’re gonna get DARPA to fund it or something like that. We’ll have government support. We’re talking with a few other folks to do it in White Hat mode, but it’s almost like red-teaming.

Can’t wait to see what you cook up next. And see you at Consensus!