Matthew Green gives 15% odds of losing trust in public-key encryption
A short blog post, dated 9 October 2026, quotes a statement by Matthew Green on Twitter about the future of public-key encryption.
Green opens by saying everyone is very concerned about being respectable, so he is going to be "the goofball who raises worst-case possibilities." He then gives two personal estimates: he thinks there is a 1% chance that we live in Minicrypt, and a 15% chance that we functionally lose confidence in our existing public-key encryption algorithms. The quote is excerpted, and the omission is marked with [...] in the original.
The argument that follows is about speed. Green writes that the speed of AI producing surprises and the speed of human beings replacing standards, even with the very best AI assistance, are "just orders of magnitude different." His conclusion is that you only recover from a surprise like this if you do the preparation in advance.
Minicrypt is not defined in Green's words. The blog author adds a gloss after looking it up: Minicrypt is Russell Impagliazzo's hypothetical world in which public-key encryption is impossible. So the 1% figure is a worst-case estimate of living in such a world, and the 15% figure is the broader chance that confidence in existing public-key algorithms is functionally lost.
The post is a quotation with a one-line gloss, not an analysis. It names no specific AI system, attack, algorithm or research result, and gives no timescale for the 15% estimate.
Key facts
- Matthew Green says he thinks there is a 1% chance we live in Minicrypt, a hypothetical world in which public-key encryption is impossible.
- He puts a 15% chance on functionally losing confidence in existing public-key encryption algorithms.
- His core argument: the speed of AI producing surprises and the speed of humans replacing standards are orders of magnitude different, even with the best AI assistance.
- His stated conclusion: you only recover from a surprise like this if you do the preparation in advance.
- The figures are personal estimates from a Twitter statement, quoted on a blog on 9 October 2026, and the quote is excerpted.
Why it matters
Green is not announcing a break in any algorithm. He is making a point about tempo. If AI can produce cryptographic surprises quickly, and replacing standards is slow even with strong AI help, the gap itself becomes the danger. Putting a 15% number on losing confidence in existing public-key algorithms turns a vague worry into a figure people can argue with.
Who it affects
The statement is about existing public-key encryption algorithms and the standards that govern them, so it speaks to anyone who maintains, specifies or depends on those standards. The source does not say which standards or algorithms should be replaced.
How to use it
The only actionable line is Green's own: you only recover from a surprise like this if you do the preparation in advance. The source does not say what that preparation should consist of, so treat it as a prompt to ask how quickly your own cryptographic dependencies could be swapped, not as a checklist.
How solid is it
Not very, and it does not claim to be. The 1% and 15% figures are Green's personal estimates, offered openly as worst-case possibilities, and the reasoning behind them is not given. The post is a short excerpt of a tweet relayed by a blog. The definition of Minicrypt is the blog author's own look-up, not part of Green's words.
Risks and caveats
Green frames himself as the "goofball" raising worst-case scenarios, so the numbers are best read as a deliberate stress test rather than a forecast. No timescale is given for the 15% chance, and no specific AI system, attack or research result is named as the trigger. The quote is excerpted, with an omission marked in the original, so context may be missing.
“You only recover from a surprise like this if you do the preparation in advance.”
— Matthew Green, on Twitter