OpenAI's 700-Paper Math Dump Rattles Cryptographers, But Shows No Way to Break Encryption

OpenAI's document drop of more than 700 mathematical papers has cryptography leaders rattled, The Information reported. Nothing in the coverage shows a way to break the encryption you use today.

The release came less than a month after OpenAI's September 8 claim that 10,000 agents solved the Navier-Stokes equations in under 90 hours. This time, according to The Independent, a single experimental model worked from one prompt per problem. The Economist called it "judgment day for mathematics." Scott Aaronson called it "surely one of the biggest days in mathematical history."

The alarm came fast. Johns Hopkins cryptographer Matthew Green said, "I think we might lose public key cryptography." Ethereum researcher Justin Drake urged crypto holders to prepare for "bunker mode" by moving funds to secret new addresses.

Now the footnotes. Fortune says the documents reveal solutions to hundreds of math problems, not a method for hacking computers, and will not accelerate "Q-day." The Independent says there is no evidence the results have come close to defeating the cryptography protecting the internet. Valerio Capraro of the University of Milan says "nobody knows whether the proofs are correct." Markets barely moved, and Dragonfly's Haseeb Qureshi puts the odds that "everything is totally fine" at 95%.

My take: a pile of unverified proofs is a claim, not an exploit. Panic is premature, and so is applause. Show me the eval, not the vibe.