What it does
The author’s study used nine historical cases, including a textbook example, plus synthetic traffic, and found no advantage from Jev ranking opening phrases. Live codebreaking needs a TypeSafe key.
Playable demos & bots
Explore an Enigma codebreaking machine whose local search produces candidate decryptions for Jev to judge.
Only you can see your notes.
Saved in this browser only.
Screenshot unavailable. Open the experiment ↗
The author’s study used nine historical cases, including a textbook example, plus synthetic traffic, and found no advantage from Jev ranking opening phrases. Live codebreaking needs a TypeSafe key.
Read the public research pages to see how a program can search for possible answers while AI judges the results. Here, the code tries Enigma settings and creates trial messages. Jev chooses among those messages or rejects them; it does not write the decryption.
A developer could borrow this separation for another task with a fixed set of candidate answers. Keep known answers for testing, and allow the judge to reject every candidate. Do not assume Jev improves the search itself: the author found no advantage when it ranked possible opening phrases.