The Skill Issue Institute congratulates Dr. Insufficient Entropy (Class of 2026) on a result our Cryptography faculty considers unimprovable: eight independent selections from a search space of 2^384 possibilities, roughly 10^116, every one of which landed on a number small enough to fit in a single byte. The smallest was 4.
The work was produced during a key generation ceremony on April 30, 2025, conducted with FIPS 140-2 Level 3 hardware security modules, which is the cryptographic equivalent of a bank vault with a visitor’s badge. A batch of 24 intermediate certificates was generated. A wrap and unwrap step in the ceremony script produced a second, much weaker set of key pairs alongside the original strong ones, and our faculty regards producing both sets in one sitting as considerable throughput. The weak set was the one published to the browser and operating system root stores. Systems Architecture calls this “shipping the strong keys to the vault and the weak keys to production”, and RAND 303: A Number With 116 Digits, Or 4 now covers it in detail.
The values themselves deserve study: 4, 56, 66, 115, 155, 156, 230 and 237. That is three to eight bits of entropy where an elliptic curve private key over NIST P-384 is supposed to have 384. Six of the intermediates were for TLS and two for code signing, each also cross-signed by a longer established authority, so sixteen certificates were trusted by every mainstream browser, mail server, container image, embedded device and unmaintained command line tool that ships its own copy of the root bundle. Our Supply Chain faculty considers the code signing keys the more elegant half of the achievement, since they authenticate software rather than mere web pages.
The recovery was pleasantly brief. On October 17, 2026 at 20:06 UTC a community member reported the compromise to a browser root program and demonstrated control by signing a short message with each key. The browser vendor recovered all eight independently at 21:48 UTC without knowing the original method. An independent researcher recovered the same keys from the published certificates on a laptop in under a second, which our Examinations board accepts as proof that the authority enclosed the answer key with the exam paper.
The paperwork is the crowning touch. The certificate revocation lists had been signed with the original strong keys, which made those signatures invalid, so the lists were re-signed with the flawed keys. Revocation was duly authenticated by the exact keys it existed to revoke. Our Records department calls this notarizing your own confession.
All eight intermediates were revoked on October 18, the cross-signed certificates on October 19, and the ECC root was emergency-distrusted on October 20. Within days the authority announced it was exiting the TLS issuance business, while quite correctly noting that the affected intermediates had never issued an end-entity certificate. The cleanup of more than 60 million certificates issued under its roots continues, and one major network operator spent a week updating trust stores, base images and every service that authenticates a remote endpoint it does not own.
“The ceremony was audited, notarized and photographed,” Dr. Entropy explained. “The strong keys were generated first, and nobody asked which set the certificate was actually built from. Four was a comfortable number. Everybody in the room understood four.”
Dr. Entropy’s work is a reminder to every student: randomness is not a formality. A 384-bit private key is not a number, it is a lottery, and there is no way to lose a draw you never entered.