Interview with Matter Labs (zkSync) Co-founder Alex Gluchowski

Last Friday, we had Alex Gluchowski join us on Wechat for an Ask-Me-Anything session. We translated the questions and Alex’ answers below. Enjoy!

<Begin transcript>

Nour (imToken): Before we share the theme, let the two guests briefly introduce themselves.

Alex: Hi everyone, my name is Alex Gluchowski. I’m CEO and co-founder of Matter Labs and we are creators of zkSync. Our mission is to help scale blockchains, and help to make it accessible in everyday life for millions of people. I’m really excited to be here. I’m especially happy to have an AMA with the Chinese community. I’ve spent a very long time in HongKong. Really happy to be here.

What do you think of the ETH 2.0 and Layer2 scaling solutions? Which one do you think will succeed in scaling faster? What are the problems with the other one?

Alex: When Vitalik Buterin presented the rollup-centric Ethereum roadmap back in October, he already knew that L2 would arrive much sooner than ETH 2.0. This is indeed what’s happening: it looks like most EMV-compatible rollups will launch this Summer — including zkSync 2.0.

There are a number of Layer2 solutions available to project developers, and this can lead to a problem of DeFi composability. Do you have a preferred solution for Cross Layer2 Bridge?

Bridges will indeed be useful, especially to move small liquidity volumes between rollups faster. Connext looks especially promising to me. But they won’t solve the problem of composability. For protocols to be truly composable in the way we’re used to on Ethereum, they must all reside on the same L2.

The Optimistic Rollup mainnet has not been released yet. However, the zkRollup (zkSync, Starkware) mainnet has been running for a long time. What do you think is the reason for Optimistic Rollup’s lag?

People tend to underestimate the complexity of optimistic rollups. While ZK rollups are fully secure when run by a single operator, optimistic ones require multiple validators that work very reliably. Personally, I would only trust an entirely new architecture with a lot of money once it’s been running for at least half a year, to make sure that there is no fundamental design flaw.

What was the reason Matter Labs chose the Layer2 track to develop zkSync in the first place?

We realized that Ethereum will need to scale very soon, and that it doesn’t make sense to try building another L1 that would compete with it. People who put a lot of money on chain will always pick the platform that offers the highest security. ETH is second only to BTC in this regard. For the same reason we believe that everyone will end up on the ZK rollup, since it’s the only L2 that offers exactly the same security guarantees as mainnet.

Centralized sequencers are very efficient in submitting batches but that will lead to the problem: single point of failure. How are you going to solve it?

Fully agree. This is why we’re working on a L2 consensus mechanism to decentralize validators.

What is the difficulty of zkEVM? Why did it not appear in the past, but recently both zkSync and Starkware started to announce support for generic contract capabilities on zkRollup?

Practical portability of smart contracts developed for EVM to ZK is harder than just general computation. The reason for this is that EVM operations are very heterogeneous, i.e. they take variable amount of work, so it’s hard to put them in a rigid zero-knowledge circuit without incurring massive overhead. The trick that allowed us to finally solve it was recursion in PLONK, which was first implemented by us in Summer 2020. Recursion allows us to combine together many different specialized circuits optimized for different EVM operations. Starkware is not EVM-compatible because heterogeneous recursion is currently impossible with STARKs.

The cost of zkSync recharge and withdrawal is high. How does zkSync see this problem and will it provide a solution in the future? What are the user scenarios for zkSync?

L1 operations will always be inherently expensive. The solution we see is to integrate directly with exchanges and on-off-ramp services, so that there is no need to ever go on L1 at all. We’re currently working on these integrations.

At the end of March zkSync released its 2.0 Roadmap. What do you think users should expect from the zkSync 2.0 upgrade?

EVM-compatible smart contracts in a ZK rollup!

Optional ultra-cheap zkPorter accounts. In my opinion, this is the biggest breakthrough in L2 scaling since the invention of rollups. We will soon publish a detailed post about it.

Many users have recently been concerned about zkSync airdrops. Does zkSync have any plans for the native token mechanism and distribution?

We believe that decentralization of the initial token is critical for any crypto protocol. zkSync intends to distribute the largest portion of its token to the public in a fair and permissionless way. We are still figuring out the details of how exactly this will happen.

Second part — Community questions

{Here, the community started to ask questions. Alex answered.}

Fan-1: The token on the zkSync network and the token issued on the ETH mainnet maintain a fixed mapping relationship, and this mapping relationship is managed manually. Currently, the mapping of decentralized arbitrary tokens is not supported. If you want to map an unknown mainnet token to the zkSync network, you need to manually submit the token information to the zkSync team. This is also a relatively large limitation, when zkSync can support the decentralized token mapping?

Alex: This will change very soon in the v1.x.Token listing will be fully permissionless.

Foreseeable that in the near future, with the gradual and large-scale application of L2, hackers will inevitably launch attacks on the zkSync wallet. Are you prepared for counter measures?

Yes, we encourage everyone to use native wallets, not the web-based one!

Mobile-based wallets such as imToken are a lot more secure, of course.

In the long run, what should Eth2 pay attention to? How do L2 like Loopring and Zksync compete with ETH2.0?

We won’t compete, we will build on top and extend ETH2.0! I believe that ETH 2 is a lot more about PoS than scalability. Rollups will make it scalable. Maybe the sharding plans will even be cancelled.

What do you think of the relationship between zksync and op rollup? cooperator or competitor?

I frankly don’t see why you need optimistic rollups anymore, after ZK rollups can now support fully Turing-complete, EVM-compatible smart contracts. Especially with the extension of zkPorter, which is only possible on top of a ZK rollup.

Many people regard layer2 as the transition between Ethereum 1.0 and 2.0, what do you think? can zksync play an important role in Ethereum 2.0?

Yes, as I said above, I think that ETH 2.0 is primarily about the move to PoS. Scalability is the job of rollups.

Will zksync and l2’s old project Loopring be a competitive or cooperative relationship?

Definitely cooperation. We’re looking forward to Loopring launching on zkSync. Our focus is on the R&D side, to create a platform, and Loopring is more about the community and user-facing functionality.

After the launch of the ZKSync mainnet, will the cost of some existing ecosystems such as DEX loan migration be high?

The migration of code will be easy because we now support Solidity. The migration of liquidity is easy as well. I don’t see any issues.

Will it rise like eth?

zkRollup is at least 100x more scalable than ETH, and zkPorter is practically infinitely scalable. We have a lot of room before fees will start rising strongly, especially on the zkPorter side.

I personally deposit some SNX on op L2, took me very long time to withdrawal, it’s annoying

Yes, optimistic rollups have a problem of a 1-week withdrawal period. zkRollups are better in this regard, it only takes minutes to withdraw.

Can ZKSync support more Chinese users, and make a Chinese translation convenient for users to use, after all, China is huge, 1.4 billion people

We don’t focus on frontend, but many great Chinese wallets are going to support zkSync.

Hey Alex,is it possible that it could be interchangeable for ZK Rollup chains. Say zksync and Aztec chain?

It would be possible, but rather within ETH ecosystem than between different L2s.

Has zk considered making a bridge tool by itself, and what projects(bridge) are currently doing it with zk?

We’re focused on zkSync v2.0, bridges are currently not in focus.

Thank you for your active participation and this is the end of this AMA.

谢谢!!! Thanks everyone!

<End of transcript>

Thanks again to Alex for taking the time! 😃