Cardano founder Charles Hoskinson on October 9 rejected Vitalik Buterin's warning that AI-accelerated mathematics could compromise lattice-based cryptography, challenging the mathematical reasoning behind the Ethereum co-founder's concerns.
In a post on X, Hoskinson disputed Buterin's argument that lattice cryptography faces risks comparable to those that undermined integer factorization through the General Number Field Sieve (GNFS). Buterin had suggested that AI could discover similar shortcuts for lattice problems, requiring developers to substantially increase cryptographic parameters.
Hoskinson countered that lattice-based systems have withstood decades of cryptanalytic attacks, including the LLL attack in 1982 and subsequent lattice sieving optimizations. He noted that post-quantum cryptographic standards ML-KEM and ML-DSA have incorporated these historical attack vectors into their security parameter design.
The Cardano founder also challenged Buterin's assertion that hash-based cryptography offers inherent safety advantages. Hoskinson pointed out that MD5 and SHA-1 were both broken through attacks on their internal structures, and that hash functions like Poseidon and Poseidon2 used in Ethereum research are built from algebraic operations subject to cryptanalysis.
Hoskinson rejected Buterin's recommendation to multiply lattice-based key sizes by ten, arguing that sound parameter selection requires specific attack-cost calculations rather than speculative multipliers applied without concrete threats.
The dispute reflects a broader disagreement over cryptographic strategy as both blockchains prepare for potential quantum computing threats. Buterin has indicated Ethereum's roadmap increasingly favors hash-based signatures including WOTS and SPHINCS+, while acknowledging that public-key encryption remains necessary for secure communications and privacy protocols. Hoskinson contended that abandoning lattice-based systems before a demonstrated attack emerges contradicts established security research and standardization practices.
