Ethereum researcher Justin Drake is urging the crypto community to start planning for what he calls "bunker mode", as he claims the ever-growing leaps in AI might threaten Bitcoin, Ethereum, and other key blockchain networks.
The researcher is encouraging big and sophisticated crypto players to start thinking about a controlled movement of funds to new addresses with public keys never seen on-chain.
"Today I call upon the blockchain industry to calmly start planning for ‘bunker mode," wrote Drake.
He has proposed such a plan as a way to limit the possible fallout of a theoretical breakthrough that would make it possible to get private keys from visible public keys.
Notably, Drake asked users not to panic and move in a rush. Instead, he framed this measure as a prevention that can be done today with current wallets and crypto infrastructure.
Will AI break ECDSA?
The issue raised by Drake pertains to the Elliptic Curve Digital Signature Algorithm (ECDSA), a common way to approve crypto transactions.
The Ethereum researcher says that it is now time to think about a theoretical extreme case. That is, ECDSA could be broken by classical computers before we reach the power of quantum computers (usually called "Q-day").
In the worst-case scenario, this kind of hypothetical breakthrough would allow us to extract a private key from a public key in approximately 1 week, with existing hardware.
At the moment, there is no publicly known attack that can do that on the curves adopted in key cryptos. This is not a warning to react to a compromise. It is a warning to prepare for a theoretical future breakthrough.
In Drake's view, as AI gets more advanced, we may see an explosive speed of mathematical discoveries, with the possibility of finding algorithms that may escape human eyes.
In support of his thesis, he mentions a few recent mathematical advancements that shattered long-time assumptions. It shows that even well-established conjectures and complexity bounds can be unexpectedly broken.
Drake also mentions OpenAI's recent developments in mathematical reasoning. He says that "mathematical superintelligence" could offer a never-before-seen speed of mathematical discoveries.
"Could our magic 64-byte ECDSA signatures be too good to be true?", asks the researcher.
Nonetheless, Drake always insisted that at this stage it's a matter of preparation, not panic. There is no evidence that AI has broken ECDSA, RSA, or the crypto fundamentals of Bitcoin and Ethereum.
