Cardano founder Charles Hoskinson has publicly challenged Ethereum co-founder Vitalik Buterin's warning that artificial intelligence could expose fundamental weaknesses in lattice-based cryptography, arguing that decades of security research have already tested such systems against known attacks and that Buterin's concerns lack supporting evidence. The dispute centers on the long-term safety of several post-quantum security methods and adds to an ongoing debate over how cryptographic infrastructure should prepare for quantum-era threats.

Buterin's warning questioned whether post-quantum schemes could withstand AI-accelerated mathematical research. He suggested that AI could discover previously unknown weaknesses in lattice-based cryptography within two years, and he compared that possibility to advances in integer factoring that eventually made older RSA security assumptions less reliable. RSA's security rests on the difficulty of factoring large integers, and successive algorithmic improvements steadily eroded the key sizes once considered safe.

Hoskinson countered that the comparison does not hold, arguing that factoring breakthroughs depended on specific mathematical properties that researchers had identified and exploited, whereas the case against lattices rests on assumptions about mathematical breakthroughs rather than demonstrated weaknesses in existing cryptographic systems.

Hoskinson Rejects the Warning About Lattice Cryptography

Hoskinson laid out his position in a public post on X on October 9, 2026:

Vitalik is now trying to convince everyone that lattices are bad because he is bag-holding too much hash-based crypto research can't back out. The case against lattices is the GNFS story, a.k.a. a hunch about "structure," and a multiplier pulled out of thin air. None of it has…

— Charles Hoskinson (@IOHK_Charles) October 9, 2026

https://x.com/IOHK_Charles/status/2108377196561867242?ref_src=twsrc%5Etfw

He explained that lattice cryptography has undergone decades of research, including sustained improvements in the algorithms used to attack such systems. According to Hoskinson, the security parameters of ML-KEM and ML-DSA — lattice-based schemes used for key encapsulation and digital signatures — already account for established attacks and their computational costs.

Hoskinson also disputed Buterin's suggestion that lattice-based encryption keys might need to become ten times larger. “Multiply key sizes by ten is numerology,” he wrote, arguing that parameter changes must follow measurable improvements in attack methods. Under certain assumptions, he noted, a 20% improvement in attack efficiency could require approximately 25% larger dimensions — far short of a tenfold expansion.

He likewise defended the existing process for vetting cryptographic systems, pointing to weaknesses uncovered during past security reviews. Hoskinson cited the failures of Rainbow and SIKE, two post-quantum candidates whose vulnerabilities emerged during public evaluation. Those cases, he argued, show that researchers can spot flaws before affected cryptographic methods are widely deployed, and that new AI-assisted attacks should undergo similar testing before developers abandon established security approaches.

The exchange comes as Cardano continues its own research into quantum-resistant security. In May, Hoskinson confirmed that Cardano was preparing a governance strategy addressing future quantum computing threats.

Hoskinson Questions Ethereum's Hash-Based Security Strategy

The disagreement also concerns Ethereum's growing preference for hash-based cryptography, which Buterin considers less exposed to certain mathematical attacks. Buterin has argued that AI could accelerate mathematical research enough to weaken lattice-based systems previously considered resistant to quantum computers, and his Ethereum security roadmap places greater emphasis on hash-based signatures, including WOTS (Winternitz One-Time Signatures) and SPHINCS.

Hoskinson disputed the assumption that hash-based systems are inherently safer because they appear to contain less exploitable mathematical structure. In his warning, Buterin pointed to historical attacks against MD5 and SHA-1, arguing that hash functions have also suffered serious security failures.

Hoskinson further questioned Ethereum's reliance on Poseidon and Poseidon2, which use mathematical structures designed for efficient cryptographic proofs. He argued that these designs also require careful security analysis, because their algebraic properties could become targets for future attacks.

The criticism comes as Ethereum continues developing quantum-resistant upgrades alongside improvements to privacy and transaction verification. Buterin has also recommended caution when using older wallet addresses, although he warned against rushing into potentially unsafe migrations.

Hoskinson raised a separate concern about discouraging the adoption of lattice-based encryption before researchers identify a practical weakness. He pointed to hybrid ML-KEM deployments in browsers and messaging applications, which combine established encryption with post-quantum protection. Hybrid designs are intended to keep connections secure even if one of the two underlying schemes is eventually weakened. According to Hoskinson, unnecessary doubts about these systems could delay upgrades while organizations continue relying on older cryptographic methods.

The debate comes as Ethereum developers work toward post-quantum security upgrades planned through 2029. Hoskinson maintained that security decisions should follow tested attacks, measurable costs, and practical deployment requirements rather than predictions about AI breakthroughs. The discussion is therefore likely to remain focused on new cryptanalysis, the performance and security parameters of deployed schemes, and how Cardano and Ethereum translate their respective research positions into governance and upgrade decisions.