Vitalik Buterin has closed his series on cryptographic obfuscation with an experimental technology called local mixing. The Ethereum co-founder considers this new line of research especially promising because it could make it possible to protect programs and build new encryption systems resistant to quantum computers. However, it remains to be proven that it is truly secure.
The Ethereum co-founder once again puts the focus on one of the great challenges of cryptography: how to protect a program's secrets without preventing other people from being able to use it. Vitalik has today published the third and final part of his series devoted to cryptographic obfuscation, a technology he himself has previously described as one of the great frontiers still pending in cryptography.
Vitalik and quantum computers
This time he focuses on a technique called local mixing, which proposes a different way of tackling the problem and which, according to Buterin, has the advantage that its computational cost is viable. The big unknown would no longer be just getting it to work, but proving that no one can break it.
The basic idea of obfuscation is relatively simple to understand. Imagine a program that contains a formula, a key or secret instructions. We want anyone to be able to use it and obtain the correct result, but without being able to open it and discover what secrets it contains or how it works internally. Obfuscation seeks precisely to hide the code, not necessarily the data it processes. Buterin explains that a program could be transformed into a kind of encrypted version that continues to work correctly even though its internal structure remains hidden.
Local mixing attempts to achieve this by repeatedly modifying the program's internal structure without altering what it does. That is, the program keeps giving the same answer, but each time it becomes harder to discover how it arrives at it.
Verifying security
The importance of this research goes beyond hiding programs. Buterin suggests that, if researchers manage to prove the security of certain constructions based on local mixing, they could be used to develop new public-key cryptography systems potentially resistant to quantum computers. This is especially relevant for Ethereum and the rest of the blockchain industry, which have spent years studying how to protect their systems against the future possibility of quantum computers powerful enough to compromise certain current cryptographic schemes.
However, Buterin himself stresses that the fundamental question is verifying the security of this new cryptographic family.
In the message with which he presented his article, the developer notes that he recommends closely following the work of the researchers behind this line because they plan to publish more results soon. The reason for his interest is that local mixing follows a very different philosophy from other approaches to obfuscation and does not depend on the same lattice-based cryptography framework that characterizes other constructions.
Breaking it again and again
Precisely the current approaches to obfuscation still have enormous practical problems. In June, when he started this series, Buterin warned that some rigorous iO constructions had computational costs so enormous that he described them as literally galactic.
NaoX Protocol, a project focused on post-quantum technology, responded to Buterin's publication by recalling the enormous distance between a novel cryptographic construction and a technology sufficiently proven to be used in real systems. NaoX compares the situation with ML-KEM and ML-DSA, post-quantum algorithms that went through years of public cryptanalysis before their standardization by NIST.
Its argument is that new cryptography may seem mathematically promising and work at a reasonable computational cost, but that does not yet prove that it is secure. Researchers need to try to break it again and again. Only after years of attacks, analysis and review can there begin to be enough confidence to use a cryptographic technology in systems where money or sensitive information is at stake.
Vitalik talks about using artificial intelligence to speed up the process. Historically, a new cryptographic family may need many years of analysis before reaching a high level of confidence. Buterin suggests that AI could help in cryptanalysis and in the search for vulnerabilities, potentially reducing the time needed to test new constructions.