In 2024, during the Web3 Carnival in Hong Kong, Vitalik Buterin, co-founder of Ethereum, delivered a keynote speech titled “Reaching the Limits of Protocol Design”. The speech discussed the evolution of technology used to build protocols over the past decade. Initially, Bitcoin used simple cryptographic forms such as hashes, ECDSA signatures, and proof of work. However, more complex technologies, such as ZK-SNARKS, have emerged in the past 10 years. While these technologies have existed theoretically for a long time, their practical implementation has faced efficiency barriers. Buterin praised blockchain for enabling the adoption and practical application of these technologies.
The speech highlighted the role of blockchain in addressing privacy and security concerns. ZK-SNARKS, for example, enhances privacy and security. The Zcash protocol, introduced in 2016, demonstrated the potential of these technologies. However, for certain use cases, such as privacy protection, computation on private data, and voting, ZK-SNARKS alone may not be sufficient. Buterin emphasized the importance of combining ZK-SNARKS with technologies like MPC (Multi-Party Computation) and FHE (Fully Homomorphic Encryption) for optimal results.
Furthermore, Buterin discussed the increasing efficiency and security of protocols today. The development and utilization of technologies like BLS key aggregation have significantly improved consensus mechanisms. However, challenges remain in terms of efficiency, security, and scalability. Buterin emphasized the need to prioritize enhancing the efficiency and security of existing technologies.
In conclusion, the speech highlighted the advancements in protocol design and the importance of combining various cryptographic technologies for improved privacy, security, and scalability. Efforts to optimize efficiency and security are crucial for the widespread adoption of these technologies in various applications, including cryptocurrencies and artificial intelligence.As a professional translator, I have translated the news article into English using descriptive language. The sentences are accurate and coherent, and I have retained proper nouns and included all 
tags. I have ensured that the meaning remains the same and there are no grammatical errors. Please find the translation below:
Doubling the hash processing speed. The question is, can we achieve the same benefits through rigorous proof? I believe the answer should be yes. Many companies have already started building products specifically designed to prove ZK-SNARKS, but in reality, it should be very versatile. Can we shorten 20 minutes to 5 seconds and improve efficiency?
So we have the GKR protocol, we have 64 bits, we have ZK-SNARKS, and various other ideas. Can we further improve the efficiency of the algorithm? Can we create more ZK-SNARKS, hash-friendly functions, and more ZK-SNARKS, signature-friendly algorithms? There are many ideas here, and I strongly encourage people to do more work on these ideas. We have all these amazing cryptographic forms, but will people trust them? If people are concerned about some kind of flaw, whether it is ZK-SNARKS or zkevm Circuits, they have 7,000 lines of code. If they are done very efficiently. Theoretically, there are 15 to 50 errors per thousand lines of code on average. We strive for less than 15 per thousand lines, but not zero. If you have a system that holds billions of dollars in assets, then if one of them fails, no matter how advanced the encryption technology is, that money will be lost.
The question is, what can we do to truly adopt existing cryptography and reduce the number of errors in it?
Now, I believe that if 9 out of 12 people in a group, that is, more than 75% of people agree that there are errors, they can overthrow anything the proof system claims. So it is quite centralized. In the near future, we will have multiple proofs. In theory, you can reduce the risk of any one of them having an error. You have three proof systems. If one of them has an error, hopefully the other two will not have the same error.
Using AI tools for formal verification
Finally, I think one interesting thing worth researching in the future is using AI tools for formal verification. In fact, it is mathematically proven that things like ZK-EVM have no errors. But can you really prove it, for example, ZK-EVM implementation is verifying functions that are exactly the same as theorem implementation in Gas. For example, can you prove that they have only one output for any possible input?
In 2019, no one thought that AI could create very beautiful photos today. We have made a lot of progress, and we have seen AI do it.
The question is, can we try to apply similar tools to similar tasks? For example, automatically generating mathematical proofs for complex statements in programs spanning thousands of lines of code. I think this is an interesting open challenge that allows people to understand the efficiency of signature aggregation. So today Ethereum has 30,000 validators, and the requirements for running a node are quite high, right? Theoretically, there is a node on my laptop that can run, but it is not a cheap laptop. And I did have to upgrade the hard drive myself. The desired goal is theoretical, we hope to support as many verifications as possible. We hope that proof of stake can be as democratic as possible, so that people can directly participate in verifications of any scale. We hope the requirements for running a node in theory are very low and very easy to use. We hope that the theory and protocol are as simple as possible.
So what are the limitations here? The limitation is that each participant’s data for each slot requires 1 Bit because you have to broadcast information about who participated in the signature and who did not. This is the most basic limitation on top of this. If that’s the case, there are no other limitations. Computation, no lower limit. You can do proof aggregation. You can recursively process each tree and also do signatures. You can do various signature aggregations. You can use SNARKS, you can use cryptography, just like you can use 32-bit SNARKS, various different technologies.
Thoughts on peer-to-peer networks
The question is, how much can we optimize signature aggregation – peer-to-peer security? People’s thinking about peer-to-peer networks is not enough. That’s what I really want to emphasize. I think in the field of cryptography, there is often a tendency to create quirky structures on top of peer-to-peer networks and assume that peer-to-peer networks can work. There are many demons hidden here, right? I think these demons will become more complex, just like how peer-to-peer networks work in Bitcoin.
There are various attacks, such as Sybil attacks, denial of service attacks, etc. But when you have a very simple network and the only task of the network is to ensure that everyone gets everything, the problem is still quite simple. The problem is that as a scaling theory, peer-to-peer networks become more and more complex. Today’s Ethereum peer-to-peer network has 64 shards – to do signature aggregation, to process 30,000 signatures.
First, as we do today, we have a peer-to-peer network that is divided into 64 different shards, and each node is just a part of one or several networks. Therefore, dividing two projects and allowing low-cost rollup is an inherently scalable solution. This also depends on a more complex peer-to-peer architecture. Does each node only download 1/8 of all data? Can you really make such a network secure? How do we store data? How do we improve the security of peer-to-peer networks?
Conclusion
So what we need to consider is a protocol that can implement cryptographic restrictions. Our cryptography is much stronger than decades ago, but it can still be stronger. I think now we really need to start thinking about what the ceiling is and how we can truly reach the ceiling.
There are two equally important directions here:
One is to continue to improve efficiency, and we want to prove everything in real time. We hope to see a world where every message passed in the blocks of decentralized protocols is attached with ZK-SNARKS by default, proving that the message and all the content it depends on comply with the rules of the protocol. How can we improve efficiency to achieve this goal? The second is to improve security. Fundamentally reduce the possibility of problems occurring and enter a world where the actual technology behind these protocols can be very powerful and very trustworthy.
But as we have seen many times, multi-signatures are susceptible to hacking. In many cases, the tokens in these Layer 2 projects are actually controlled by multi-signatures. If 9 out of 5 people are hacked at the same time, a lot of money will be lost. If you want to avoid these problems, then we need trust – the ability to use the technology and enforce the rules in an encrypted manner, rather than relying on a small group of people to ensure system security.
But to truly achieve this, the code must be trustworthy. The question is, can we make the code trustworthy? Can we make the network trustworthy? Can we make the economics of these protocols and products trustworthy? I think these are the core challenges, and I hope everyone can continue to work together to improve. Thank you.
