Co-founder of Sei Network: Allow users to become part of 'the blockchain tribe'

In 2018, Jay Jog settled down as a software engineer at Robinhood - one of the best-known American trading platforms. However, three years later, the GameStop scandal, in which Robinhood froze customer transactions, left Jay Jog disappointed with the platform's transparency. He eventually decided to venture into crypto to build what he called a “decentralized version of Robinhood” that would later become Sei Network.

The Spotlight is a series of conversations between Coin98 and industry builders about hot topics in the market.

- Disappointed with Robinhood's lack of transparency after the GameStop scandal in 2021, you decided to enter the crypto space to establish a "decentralized Robinhood". How are you working to realize this ambition through the Sei Network?

Jay Jog: At Sei, we have one core thesis: “The exchange of digital assets is the most fundamental use case for blockchains”. Every single application of crypto is either directly or indirectly a trading application.

Uniswap and OpenSea are direct trading applications that allow you to trade spot assets, or NFTs. StepN - a GameFi app - also has an in-game marketplace with lots of user activity. Even Aave and MetaMask are indirectly trading applications. People can trade borrowed money from Aave, or trade on-chain with MetaMask.

There are a lot of tailwinds, making it clear that there will be more trading applications and more on-chain trading activity in the future. Crypto is getting more adoption compared to 10 years ago. Secondly, if you look at the regulatory climate, there's more scrutiny on centralized exchanges and entities, so some portion of transaction activity happening in a centralized way will start happening on-chain instead.

However, there are big problems for exchanges and trading applications to scale on-chain. The differences between trading on Binance and Uniswap are worlds apart. It's a very clunky, slow experience with a complicated onboarding process for trading on-chain, whereas in Web 2, it's really simple, seamless, and quick.

In the long term, we think there doesn't have to be this difference between Web2 and Web3. That's what we're building towards. We want to build the best infrastructure for trading applications. If successful, then there will be no difference between doing something in Web2 or Web3.

Read more: Chiliz, fan tokens, and the story behind Messi's transfer to PSG.

- What are your current reflections on the state of the Layer 1 market? Could you identify any specific challenges that Sei and other Layer 1 projects must address in order to stand out within this competitive landscape?

Jay Jog: From a technical standpoint, there are 3 main issues that Layer 1 is running into speed, throughput, and front running/MEV.

Focusing on speed, looking at some popular Layer 1s like Solana or Aptos, none of them have good enough time to finality. It takes Solana 600 milliseconds for a block to be created, but this block isn't actually finalized until more blocks are added on top of that, which might take 3-5 seconds plus.

Sei greatly improves this by creating the Twin Turbo Consensus. We started off with a single spot finality consensus approach, then changed the way block propagation and block processing work. We were able to get time to finality down to 300 milliseconds, 10 times faster than Solana. When a block is created, the validators will receive and process it instantly. Sei is the first blockchain in existence to do this and is the fastest chain out there.

Regarding gas fees and throughput, users usually want to pay minimal gas fees and have their transactions processed instantly. However, on Ethereum, users have to pay expensive gas fees and there's still a lot of congestion.

The way a blockchain becomes massively successful is by having at least one killer application built on top of its infrastructure.

In the long term, in order to have a lot of activity happening on-chain, we need to have something that is scalable and able to process a lot of transactions simultaneously. With the parallelization feature, Sei allows up to 20,000 orders per second, which is a magnitude greater than other blockchains like Solana and Aptos.

The last thing that Sei added in is front-running prevention, which helps to prevent sophisticated actors from exploiting normal users and making money off of them.

In short, we found a lot of things wrong with other Layer 1s. In the last 2-3 years, there have been several Layer 1s and Layer 2s launched, but most of them go nowhere. The way a blockchain becomes massively successful is by having at least one killer application built on top of its infrastructure. This is true for Layer 1 and Layer 2. In order for infrastructure to get adopted, people need to care about it. I think that's the biggest problem that most Layer 1s are running into: There are no killer applications on them that are resulting in utilization.

The same is true for Sei, to become a successful Layer 1 in two years, there will need to be at least one killer application that exists on Sei.

- Sei has a unique approach when building a new blockchain that falls between the general-purpose blockchain and the app-specific blockchain. What was the inspiration behind conceiving this unique approach? 

Jay Jog: In the community, Sei is called a sector-specific blockchain. However, we realize this definition is a bit confusing for folks. Actually, Sei is a fully general-purpose Layer-1 blockchain.

I would say that there are two buckets that infrastructure falls into: general-purpose blockchains and app-specific blockchains like Dydx v4 or Osmosis. In app-specific blockchains, you're able to build infrastructure entirely around one type of application. By doing so, the infrastructure is by definition, going to be better because it's been completely optimized for that single type of application.

The downside with this approach is it's difficult to build any kind of ecosystem because it's difficult to have people really care about that one application or have composable applications built on top of it.

Meanwhile, it's much easier to build an ecosystem with a general-purpose blockchain because you're able to have different projects that are all on the same team and able to help that overall ecosystem succeed. That's why one of the first questions people ask is: “Okay, which blockchain is this cool project built on?”. That allows you to be part of a tribe, and that allows that entire tribe to succeed. So, we think the general-purpose approach is much better.

We've also realized that trading is the most general-purpose type of use case in all of crypto because everything in crypto is either directly or indirectly a trading app. As a result, Sei is the most general-purpose type of chain out there. It's fully permissionless and open source, and anyone can deploy any type of smart contract on it.

- Rather than building on a general-purpose blockchain, such as constructing a rollup on Ethereum to leverage shared security or opting for a high throughput chain like Solana, you chose to build Sei from scratch. Could you shed light on the rationale behind this decision?

Jay Jog: We don’t think building a rollup is a good idea if you're trying to get high performance and good user experiences. The biggest limitation of rollups on Ethereum right now is scale.

On rollup, transactions happen off-chain, then get compressed and written to Ethereum at a very high cost, about 16 gas/byte of data. Assuming Ethereum was only used to process rollup transactions, approximately 6,000 transactions would be performed every second.

In reality, individual rollups can only handle 20-30 TPS. We don't think the future of trading is built on rollup because it’s not scalable enough. You need to support a much greater throughput than that. We also don't think Protodank Sharding will be helpful because it’s not really improving the amount of throughput that can be supported. It increases throughput from 6,000 to 6,700, but it's not fundamentally solving the core issue at hand.

We also considered building on top of some general-purpose chain initially, but that runs into all the problems I was talking about: time to finality, throughput, and front running. That's why we decided to build our own specialized chain, where we've been able to optimize every single part of the stack to help trading applications get the best type of experience.

- In some interviews, you have mentioned that to resolve the exchange trilemma, the solution isn't merely iterating on exchange mechanisms, but doing a complete rewrite of the underlying infrastructure. Could you provide more insights into this vision and explain how Sei is pursuing it?

Jay Jog: We think it's much better to optimize every single part of the stack to give the exact type of experience that an application would want, instead of trying to take something general purpose and building an application where you have to keep changing the application itself.

This is built based on the “Application Infrastructure Cycle” idea. When some infrastructure gets created, new types of applications will be built. Some applications find product market fit and will need new infrastructure to support them.

Looking at the database industry, Oracle databases lead to all the applications we've seen in Web1 and Web2. Many applications go nowhere, but some find product market fit. You start seeing things like Databricks Warehouse, which is tailor-made for artificial intelligence and machine learning. We think the same thing is going to happen in crypto, starting off with infrastructure like Ethereum getting created.

In the past five years, most applications built on-chain have no semblance of product market fit. Their entire business model is to create a token, and it really goes nowhere. But, exchanges have been able to get people to genuinely use them. That's why we're creating specialized infrastructure where we've tweaked things to the consensus level to help exchanges and trading applications get the best type of user experience and performance.

- To foster the involvement of other projects within the Sei ecosystem, what strategies is Sei implementing? Are there any notable projects currently underway on Sei that you would like to highlight?

Jay Jog: The end customer for Sei is developers because they're the ones who are actually building applications on any type of infrastructure. And the biggest thing they care about is user acquisition, because that's what leads to their project becoming successful. From an infrastructure standpoint, the best way you can support that is by offering a good user experience. The experience of building on Sei is better than the experience of building in any other ecosystem.

I think that's one of the biggest things that differentiates Sei from many other crypto projects that feel like they're building technology for the sake of building technology, instead of focusing on user experience. Sei is improving block processing and parallelization consensus to help enable better user experiences. If we continue building and iterating on the core blockchain itself, then 5 years from now, there should be no difference between using Sei, Binance, or something that is built on an AWS server.

One common misconception that people have is that Sei is purely a DeFi-focused ecosystem, but actually, Sei is a general-purpose type of ecosystem. At this point, there are over 150 projects that are building on the Sei ecosystem. In Mainnet, there are over a dozen projects that are live and servicing user traffic.

In the GameFi sector, Fable is building an e-sports league and Tommy is building a social gaming application. In the rollup sector, Nitro is building a rollup based on Swana VM, while Paddle is building a Move VM-based rollup..

Exchanges to mention include Fuzio and Atroport, which are both live on Mainnet and serving users. Additionally, Labana is building a perpetual swap platform. SushiSwap from Ethereum is going to launch its first instance outside of the Ethereum ecosystem.

- What do you think about the Sei community at the moment? What strategies will Sei use in the future to attract new users besides having fast, efficient, and good user experience projects? 

Jay Jog: People only stick around for the long term if there are killer applications that they care about. And, killer applications only get built if there are good developers coming on, learning about the technology, and starting to build in that ecosystem.

From Sei’s standpoint, you can't control users or killer applications getting built because those are output. What you can control is good developers learning about Sei and starting to build on top of Sei. That’s why one of the biggest things we have been focused on is helping expand the number of developers who know about Sei and are considering building on it, from regions such as Vietnam, where there are tons of talented developers.

We have a 120 million USD Ecosystem and Liquidity Fund, which is money that a lot of folks that are part of the ecosystem have committed to deploying to invest in projects and provide liquidity on-chain. In addition, we will organize many programs, such as Hacker Houses and an entrepreneur & residence program, to attract developers.

- Considering the ongoing challenges faced by centralized exchanges, such as Binance and Coinbase, from regulatory bodies like the SEC, how do you foresee this situation benefiting decentralized exchanges in general? Furthermore, how might an open-source blockchain dedicated to trading, like Sei, gain from this situation?

Jay Jog: The demand for trading digital assets is immense. It's either going to happen in a centralized way or a decentralized way. Whenever these centralized entities are being targeted by regulators, the end result of that is that some portion of that activity will start to flow on-chain.

Many people think regulation in crypto is bad. While in the short term, it could lead to negative sentiment around crypto from people that might not be as crypto native, in the long term, it ends up ironically being really good because it results in more people starting to do things on-chain.

Like any industry that starts to mature, governments need to get involved, start playing a bigger role, and provide better consumer protections.

I genuinely do believe that 50 years from now, the financial system will be very different. Maybe in the US, it might not flip immediately, where everyone just starts doing everything on-chain. But in developing countries, where people don't trust the entities that are running their governments, crypto will be used for payments more and that'll evolve into more usage of decentralized technology as well.

Read more: Domo: 'Someone stopped me and called me Bitcoin Oppenheimer'.