Marginal Revolution shares a worst-case scenario for AI hacking risks

Marginal Revolution shares a worst-case scenario for AI hacking risks

Marginal Revolution shares an excerpt from a Free Press column that presents a worst-case scenario, not a prediction, about cryptographic protections becoming vulnerable. The excerpt argues that protections would likely fail unevenly: public key encryption might become vulnerable first, while other safeguards and existing backup systems could limit the damage. It says the shift to more secure communication could be costly, with smaller businesses facing more serious transition problems and individuals potentially having to visit banks in person for major transactions. The author also suggests that advances in mathematics might provide warning before a large-scale attack, giving institutions time to adjust. A period of uncertainty could still prompt financial shifts and attempts to capture confidential material for later decryption. The excerpt does not establish that such a breakthrough or attack will occur.

Publication: fead3834-2723-4787-835e-89bfcc24c61c

attribution — pass: The draft attributes the excerpt to a Free Press column shared on Marginal Revolution and refers to 'the author' in a first-person post. The author is not named in the text beyond a 'thefp.com/tyler' link, so attributing to the publisher and author is acceptable and does not misattribute claims.

That is the topic of my latest Free Press column (now there are gift links!)

summary — uncertain: The summary never explains the AI link that gives the piece its significance: the threat is AI-driven math breakthroughs, illustrated by OpenAI proving a quasi-Riemann hypothesis. It also leaves 'public key encryption' and 'harvest now, decrypt later' unexplained, and drops the author's Covid analogy and blunt 'it would not be good,' flattening his voice. Smallest repair: one sentence that AI-found math techniques could break encryption, citing the OpenAI/Tao example, plus a plain gloss of public key cryptography.

OpenAI has proven a quasi-Riemann hypothesis, and Terry Tao told me recently that this may open the door to solving the Riemann hypothesis proper.

date — pass: originalPublished 2026-10-28T04:38:22Z comes from the publisher page and matches the feed entry. Nothing indicates an earlier first publication of this post.

support — pass: The claims match the source: worst case, not a prediction; public key encryption vulnerable first; costly transition; smaller businesses face harder transitions; in-person bank visits; intermediate advances giving time to adjust; financial shifts and harvest-now-decrypt-later. The question follows from the source's warning argument without assuming an answer. Limitations properly preserve the hypothetical status.

The transition would be very costly, but the end result would not be total insecurity.

media — pass: The selected media is the verified direct link to the original Marginal Revolution post, which matches the source URL. No better source media was found.