mimesis

Left or right, run 1 (B1–B3): result

Run on 2026-10-11 at 21:18:57 UTC, exactly as pre-registered in PREREGISTRATION.md (published at mimesis 79e7e70). Before the run, a lock check printed 79e7e70, a clean git status, 6 of 6 input hashes OK and the script hash cc6e3779…. The command was the published one, word for word. Output: run1/result.md, run1/result.json.

Result: no match

Neither face is in B1, B2 or B3. No slice region comes within 5 bits of either face.

Video	Slices (5/s)	Left (guest): closest	Right (host): closest
B1 (Jul 31)	219	8 bits, at 0:26.6	12, at 0:26.6
B2 (Aug 3)	90	17	15
B3 (Aug 11)	262	10	11

Eye check of the nearest spot (left face, B1, 0:26.6, 8 bits), shown in run1/eyecheck_nearest_B1_left.png. The spot is a narrow strip of wall and a figurine in the background of a video call, not a face. 8 bits is also within the range the decoys scored (9–19) in the method check.

What this does and does not say

• It says: the two face pieces of the S6E89 thumbnail do not appear in the three ldsabuseonx videos that came before it in the twin structure (B1–B3).
• It does not say: where the faces come from, or who had them first. Under John’s rule the question is the whole Joder-verse before Aug 18, 16:44 PDT; B1–B3 are only the first three places looked.
• The picks are not settled: John right, Claude right (~60%), Grok left. Next runs widen the search, each pre-registered and published the same way before it runs.

This site is open source. Improve this page.