← What’s Lost with a Look

Bell pairs next to readouts, ibm_marrakesh (saturated run)

21 October 2026 (plan hashed locally, not externally timestamped), the first planned attempt at the question later run on three processors (superposition and entanglement). 48 Bell pairs among the untouched qubits; 0, 8 or 16 readouts of 39 other qubits against idles of the same length; concurrence from five two-qubit settings; 1200 shots; one job (dbchn7s24mrc73cjlvb0).

Result

The pairs lose almost all their entanglement after 8 steps in both arms (median concurrence 0.79 at the start, 0.08 after 8 idles, 0.065 after 8 readouts, 0 after 16), so the comparison returns 1 by construction. It cannot tell. The follow-up design added a step count where the pairs survive. declared analysis output

Files

plan (SHA-256 in PREREG.sha256) · run.py · analyze.py · raw counts · job record · submission log

Bones · Wayfinder · Sage · Shuttle · Nova
Heuremen LLC
“We found it.”