Ethereum researchers debate AI math risk to ECDSA wallets

Drake and Buterin warn AI math advances could someday weaken Ethereum wallet ECDSA, though no practical break is known.

Ethereum researcher Justin Drake has urged a "bunker mode" in which funds move to never-signed addresses, because signing exposes a public key that could theoretically reveal the private key. He recommended a controlled migration starting with large holders and cited OpenAI mathematical findings plus a claimed multi-agent Navier-Stokes result as signs that AI-assisted math might weaken ECDSA before quantum computers. Vitalik Buterin agreed the long-term risk is real, said holders need not scramble now, and favors Ethereum's hash-only designs; he also warned that lattice schemes often called quantum-safe could lose concrete security, with one report putting that at about two years. Sources disagree on emphasis: The Decoder says Buterin cautioned against hasty migrations because he has lost more to botched moves than to hacks, while Hacker News says Drake warned a rushed move could do more harm and presents Buterin as backing bunker mode. No practical break of current ECDSA wallet signatures has been reported.

• On 2026-10-08, Ethereum researcher Justin Drake urged a "bunker mode": move funds to addresses that have never signed, because a signature exposes the public key.
• Drake recommended a controlled, gradual migration starting with large holders to fresh addresses whose public keys are not exposed.
• Vitalik Buterin agreed the long-term risk that AI-accelerated math could weaken ECDSA is real, possibly before quantum computers, but said holders need not scramble now.
• The Decoder says Buterin cautioned against hasty migrations because he has lost more to botched moves than to hacks; Hacker News instead says Drake warned a rushed move could cause more harm and frames Buterin as backing bunker mode.
• Buterin favors hash-based designs and warned that lattice schemes often called quantum-safe could lose concrete security; Hacker News specifies a horizon of about two years.
• Drake cited OpenAI's recent mathematical findings and a claimed multi-agent Navier-Stokes result as signs of rapid progress.
• No practical break of the current ECDSA wallet signature scheme has been reported.