The cryptography professor Matthew Green makes a stark prediction:
I think we might lose public key cryptography.
https://x.com/matthew_d_green/status/2107992246360649879
The question asked by this market is simple, will we? I will define this as all existing public key cryptography algorithms in widespread use being undermined to the point of immediate deprecation by public advancements in mathematics. To get specific that means:
RSA
Diffie–Hellman
Elliptic-curve Diffie–Hellman
ECDSA
EdDSA
ML-KEM
Streamlined NTRU Prime (sntrup761)
It doesn't matter if the deprecation is caused by AI or not, nor does it matter if new algorithms are designed that immediately replace these and are considered safe. By "public" I mean that there exists a public demonstration which convinces experts that existing algorithms need to be immediately deprecated and people behave as though this is true (e.g. Chrome changes their crypto, or we temporarily shut down all e-commerce because it can't be guaranteed safe, things like this), the actual method that breaks the algorithms doesn't need to be public just that a party is in possession of one and the algorithms are now considered deprecated by experts.
If this happens by 2028 I resolve the question YES, if it doesn't happen I resolve NO at the end of 2028.
Update 2026-10-08 (PST) (AI summary of creator comment): The market will resolve YES if we lose public key encryption, even if public key signing (such as hash-based signing) remains secure.
From the algorithms listed, I assume this is intended to be about public key encryption as Matthew Green specifies in his follow-up tweet, meaning this resolves YES if we lose public key encryption but still have public key signing. (Hash-based public key signing algorithms like SPHINCS+ only rely on similar assumptions to symmetric cryptography, so they're much less likely to get broken than any other public key cryptography.)
Just to get ahead of this question now, obviously it doesn't have to be all at once. If all of these are deprecated by EoY 2028 I will mark it YES.