Canopus public run / sidon-sets
pending reviewBounded work, independently verified, not accepted.
This is a validated, sanitized projection of one retained Canopus run. gpt-5.6-sol produced the bounded artifact; a separate frozen verifier checked it; neither actor could accept scientific state.
Produced the exact reconstructed 7,194-point Sidon witness for {0,1}^24.
Verifier success is not scientific acceptance. The bounded result does not establish maximality or settle the broader scientific problem. Canopus did not sign or perform a human decision.
Evidence chain
Each handoff is root-bound. Defer is the successful policy outcome.
- 01
Mission
sidon:a24-improve
sha256:3b529df4304…
- 02
gpt-5.6-sol worker
success
sha256:09977a08357…
- 03
Artifact
bounded output
sha256:878b05e01db…
- 04
Frozen verifier
pass
sha256:89130864b3c…
- 05
Vela Receipt
recorded
sha256:91b9f0c72e2…
- 06
Defer
human decision retained
One new point, exactly verified
The mission started from a mechanically verified but unaccepted 7,193-point seed. GPT-5.6 proposed an exchange; the frozen verifier reconstructed a 7,194-point subset of {0,1}^24 and checked every unordered componentwise pair sum for uniqueness.
- 25,880,415 pair sums checked
- clean-clone replay matched
- policy route: defer
- accepted-state delta: zero
Verify the pending result from a clean clone
git clone https://github.com/vela-science/sidon-frontier.git
cd sidon-frontier
git checkout 825657d7e87618c0aa6fc9af7e3182e05f324750
vela reproduce artifacts/sidon-a24-gpt56-7194.witness.json
node verification/verify-sidon-a24-7194.mjs artifacts/sidon-a24-gpt56-7194.witness.jsonThe first verifier is Vela's frozen Sidon verifier. The second is an independent base-3 implementation that also rejects a bound collision injection. Both verify the pending artifact; neither accepts it.
Open independent audit evidence