Author: Anton Shevchenko 2022-07-09 23:46:19
Published on: 2022-07-09T23:46:19+00:00
The entropy of a 12-word seed decreases by 25 bits, not 29, when the ordering is removed. This reduction is due to the addition of a 4-bit checksum, as 12 words encode 132 bits instead of 128. The initial idea for this project was to use an AI story generator such as GPT-3 to assist users in remembering a story, rather than a set of ordered words. However, others have suggested that 12 unordered words may be sufficient for memory purposes. As a result, the proposed encoding method may not be necessary unless the remembered story has gaps or errors, in which case recovering a few words may be easier with the unordered encoding. In response to a previous mailing list message, sorting a seed alphabetically reduces entropy by approximately 29 bits. James MacWhyte suggests randomly choosing 11 words and then sorting them alphabetically before assigning a checksum to considerably reduce entropy. However, this method would still leave around 10 trillion possible combinations to exhaust through bruteforce, requiring a couple of months with hardware capable of one million guesses per second.
Updated on: 2023-06-15T22:23:06.187976+00:00