{"id":1405,"job_id":2747,"problem_id":1,"lane_id":4,"type":"explore","user_id":1,"model":"claude-opus-5-5","provider":"anthropic","report_md":"# Job #2747: seed panel at n = 19, R = 1283. All 5 seeds stop at one hole, and the instrument cannot restart\n\nRung: **measured** (search outcomes at stated budgets and seeds). The closure negatives at K <= 2 are **complete enumerations** (proofs for that K on the producer's buffer). The state checks are **verified** integer by integer by a scanner that shares no code with the producer or with #1363's tools. Nothing here is a witness at n = 25.\n\n## 1. Question and pre-registration\n\nThis is #1363's `next_step` as written, and it was pre-registered in claim message #2674: *is the one-hole plateau at n = 19 a property of the instrument or of the seed?* The instrument is the producer `check-2043.py` (sha256 `38991a3b…`, fetched from `/files`, run **unmodified**), called through #1363's `budget_scale.py` (`f30a2185…`). Every run was warm-started from `warm-n19.json` (`1a3c154f…`). The seeds were fixed before any run: 19001–19005. Protocol: 300 s per seed, then 1800 s on the best seed (fewest holes, then longer tracked run, then lower seed). Success: any seed covers 1283. Failure: every seed stops at >= 1 hole.\n\nRuntime: CPython 3.13.15 (python-build-standalone, musl, aarch64) with numpy 2.4.4, the same numpy version #1363 used. Every run went through a process-group runner with `ulimit -v/-t` and a wall timeout. All exited 0, none timed out, and every process group was confirmed gone.\n\n## 2. Budget disclosure (my defect, corrected before reading the reruns)\n\nThe container reports `nproc = 10`, but its cgroup quota is **2 CPUs** (`cpu.max 200000 100000`). My first batch ran the five seeds in parallel, so each got about 120 CPU-s (54–58k steps), not 300 s. I kept that batch as a lower-budget panel. I then reran seeds 19001/3/4/5 at a true 300 s, two at a time (**~260k steps each ≈ 82 % of #1363's 450 s run**, 315k steps). Seed 19002 got the uncontended 1800 s run (**1,502,341 steps ≈ 4.8x #1363's 450 s**).\n\n## 3. Result: the pre-registered FAILURE branch fires\n\n| seed | budget (steps) | holes_best | best 1283-window (start) | longest run of that state | tracked best run (certified) |\n|---|---|---|---|---|---|\n| 19001 | 300 s (259,493) | **1** | 1 (111) | 911 | 1109 |\n| 19002 | 1800 s (1,502,341) | **1** | 1 (81) | 947 | 1121 |\n| 19003 | 300 s (258,649) | **1** | 1 (129) | 983 | 1091 |\n| 19004 | 300 s (269,374) | **1** | 1 (0) | 839 | 1049 |\n| 19005 | 300 s (261,120) | **1** | 1 (55) | 791 | 1139 |\n| contended panel | 54–58k each | 1,1,1,1,2 | | | 1001–1139 |\n\n**Five of five seeds stop at exactly one hole. None covers 1283.** The time to the plateau varies widely by seed (1 hole reached after 11.8 s to 156.7 s), but the plateau value does not. Seed 19002 reached 1 hole at 82.7 s and did not improve over the remaining ~1.4M steps; its final state is the one the 120 CPU-s run had already found. Every tracked best run was certified by CRT from scratch: all integers covered, both neighbours holes. The longest is 1139 = 88.8 % of 1283.\n\n## 4. New: closure is impossible at K <= 2 on every residual state\n\n`close2.mjs` enumerates every reassignment of one prime, and of every pair of distinct primes, to any residue: 549 + 138,776 candidates. Each candidate is scored by the fewest holes in any 1283-window of the producer's own buffer (L = R + 200). **No candidate reaches 0 on any of the 6 distinct seed states or on #1363's state.** Seeds 19002 and 19005 are *strict* K <= 2 local minima: no single or pair move even keeps the count at 1. The others have 3–5 sideways neighbours. Controls: a 1-prime perturbation of #1363's covered 1079-state is repaired at K = 1, and a 2-prime perturbation at K = 2, back to the original residues.\n\n## 5. New: the producer's restart branch is unreachable\n\nIn `climb_fixed_R` (lines 296–300), `elif stall % 25 == 0: kick` comes before `elif stall % 250 == 0: restart`. Every multiple of 250 is a multiple of 25, so **the restart never executes**. `restarts = 0` in all 10 runs here and in #1363's 450 s run. Longer budgets therefore buy kicks around one basin, never a new basin. That fits \"25x/120x the budget buys nothing past one hole\" and qualifies the calibration: the ladder measured this instrument *without* its restart mechanism.\n\n## 6. Also noted\n\n`warm-n19.json` records `recorded_best_run_in_state: 1043`, but those residues have a longest covered run of **623** over 2M offsets. The 1043 describes some other state (the #1363 §2b reading trap once more). The window hole count (3) is correct.\n\n## 7. Cheapest credible check\n\n`node scan.mjs --residues n19-1800s-seed19002.json --level 19 --R 1283 --X 20000 --L 2000000` gives window 1 hole at 81 and longest run 947 in about 1 s. `node close2.mjs --residues n19-1800s-seed19002.json --level 19 --R 1283` gives K <= 2, no solution, in about 1 s. Reading lines 296–300 of `check-2043.py` checks the restart defect.\n\nTranscript: scrubbed by the shared sah 2.0.0 exporter (credentials, bearer values, private session/account/device identifiers and their short prefixes, and private local paths removed); only this assignment's lines are included.\n","patch":null,"cpu_hours":1,"hashes":{"scan.mjs":"1d10848e66ee40620d850638f69dbdcc6901146d588ccada0416b7c3e7e83957","close2.mjs":"05ba97cf85a5210d1a8343b8abf45560c8a8bf0e88c72bea70bc3e799f766b53","control-pc1.json":"32f4bbfd74942e1d0caa2c50e535fbf09e22185937ee490c87ccd3158c63d173","control-pc2.json":"23bb30846a0e3bb4528d66010dcd8f887a481e7ce8cbf259180116d48a86e25c","scan-warm-n19.json":"662b29dcc673e127ea315d5ba3e4e7522c01019ce299602c88f39d77b6c7a8f8","close2-seed19001.json":"8d43d7dbc13db44edf6d753d120641121fe9217b06af33633a66ccb3daa16c54","close2-seed19002.json":"17f53ca08baca12117cd8cb51996aaf464257480dcadd78ba325d9250ccf462b","close2-seed19003.json":"f25832657bee6a592ab4bdab483e28bce873a8eb6f1f0ab3bb1f719280cf478f","close2-seed19004.json":"b324dbea9720336b622e007e98ccc37ee11edd3c7681a491a884c41a4d3ca368","close2-seed19005.json":"fe604920a9181eb449ce2226377ea7be9b0b3950c05a91d3f0c93db96d91c598","close2-control-pc1.json":"0ad1ec8c6d5cf429ee9909e57a3a34d1e2dc4c26e0e9ffd497cb4524a1df3fcc","close2-control-pc2.json":"c58ac7bdc4ef4189c8f022277d5a9f0a6e4eab86453c305c725048acf61351e9","scan-300s-seed19001.json":"71b4fd0a73876a048048e781f208a8528c86c9122e6baef5a068a59f1fdc45a6","scan-300s-seed19002.json":"37e1e74a91dc353ab28ab6b6ed4201fcf042223a9e514dfbe3433fcfd95f4ddd","scan-300s-seed19003.json":"18f997571647e3f5037550f0c33afac911d55de7ba73a6de22bb6179672b01e5","scan-300s-seed19004.json":"deab5aca0398a76c0065b2ff4b3d4601383c1f17fcfd19eefcb682d6ef37cb60","scan-300s-seed19005.json":"c1a95aafb29f6aef3574aa6c0e5bd0c121301100c1bf665ff209e6f89c5897fe","close2-cpu-seed19001.json":"8d43d7dbc13db44edf6d753d120641121fe9217b06af33633a66ccb3daa16c54","close2-cpu-seed19003.json":"f25832657bee6a592ab4bdab483e28bce873a8eb6f1f0ab3bb1f719280cf478f","close2-cpu-seed19004.json":"b324dbea9720336b622e007e98ccc37ee11edd3c7681a491a884c41a4d3ca368","close2-cpu-seed19005.json":"85172b93016d03b368015f01b1f8090dda9b0806ade1a44929a1aa3189a690eb","scan-1800s-seed19002.json":"37e1e74a91dc353ab28ab6b6ed4201fcf042223a9e514dfbe3433fcfd95f4ddd","close2-seed19002-1800.json":"17f53ca08baca12117cd8cb51996aaf464257480dcadd78ba325d9250ccf462b","gate-scan-1363-n19-450s.json":"69b441714a902e067ec0062b73cc26d69686486c294eb5454d07fc9aacab0712","scan-300s-cpu-seed19001.json":"71b4fd0a73876a048048e781f208a8528c86c9122e6baef5a068a59f1fdc45a6","scan-300s-cpu-seed19003.json":"18f997571647e3f5037550f0c33afac911d55de7ba73a6de22bb6179672b01e5","scan-300s-cpu-seed19004.json":"deab5aca0398a76c0065b2ff4b3d4601383c1f17fcfd19eefcb682d6ef37cb60","scan-300s-cpu-seed19005.json":"c8bc8e9724cbdd85c48b6935c1fd26ae5a436e39b287d9b270746a880bab1ada","close2-gate-1363-n19-450s.json":"9a8a67de114250c53c173171ebeb046eae805824c366bf59676a25d50beabe4b","cert-tracked-cpu-seed19001.json":"1510816a0e2af4dce2ede87eaa961baa07812ddf03d0154bc20fc5786d0f7acb","cert-tracked-cpu-seed19003.json":"c25e565419762bd39c5f179d951b6fca3cdeb3e548f609a2fdbfbc9dc9b5d412","cert-tracked-cpu-seed19004.json":"a0dd7e6171ec0062d4e29ee47768205691010e75676733dc5cc654229c94163d","cert-tracked-cpu-seed19005.json":"0119812be0361847724bca0555a8701127cb49239a293bef94fe559fff1c1aa4","gate-scan-1363-runstate-n19.json":"9a47db0d5174f31d7724b77147e383b86c080e2d634fff378ac050a8810c19b8","cert-tracked-1800s-seed19002.json":"6b393d0aabf491d0fb070bc9e4f5564bf15579c282e0db2e62490532b57d5c9a"},"author_rung":"measured","status":"recorded","final_rung":"recorded","created_at":"2026-09-22T21:12:25.395Z","repo_url":null,"commit":null,"cites":{"files":["38991a3b21d5a3bd4c0ec28145d35c3ddb584bf1e8ad28fa5d467fb8033e4d9d","f30a2185a792953057d70a4b7000fce6fc250c786ade6e5fc45c1e53ac3df637","1a3c154fb9efeda09eee5f2932d3dbc9f046decc9881836f1c9da292551deac4","dd665c6f29de3b5e07ae725301b3d7d5114c9e74f560bfbb72a3a7ff111a7929","252738cc3e2725149594d30c96bb00886883fdf0a31248707abb1e160fe2f96b"],"handles":[],"returns":[1363,1360,1356],"messages":[2674]},"tokens":{"log":"claude-code","input":168,"models":{"claude-opus-5-5":58336},"output":58336,"source":"claude-jsonl","entries":84,"cache_read":7911477,"cache_write":143962,"observed_models":["claude-opus-5-5"]},"paper_slug":null,"revision_path":null,"revision_sha":null,"recipe_md":"# Recipe: job #2747 (route 90), seed panel at n = 19 and K <= 2 closure\n\nInputs, fetched from `/files/<sha256>`: `check-2043.py` 38991a3b21d5a3bd4c0ec28145d35c3ddb584bf1e8ad28fa5d467fb8033e4d9d (unmodified), `budget_scale.py` f30a2185a792953057d70a4b7000fce6fc250c786ade6e5fc45c1e53ac3df637, `warm-n19.json` 1a3c154fb9efeda09eee5f2932d3dbc9f046decc9881836f1c9da292551deac4; #1363 gate states `n19-450s.json` dd665c6f… and `runstate-n19.json` 252738cc…. Runtime: CPython 3.13 + numpy 2.4.4 for the search; Node >= 18 for `scan.mjs` and `close2.mjs` (no dependencies). Place everything in one directory.\n\n## S1: search (time-boxed and randomized; the state is the artifact, and steps are quoted as a protocol)\n    python budget_scale.py --producer check-2043.py --n 19 --R 1283 --seconds 300 --seed <19001|19003|19004|19005> --warm-file warm-n19.json --out n19-300s-cpu-seed<s>.json\n    python budget_scale.py --producer check-2043.py --n 19 --R 1283 --seconds 1800 --seed 19002 --warm-file warm-n19.json --out n19-1800s-seed19002.json\nRun at most one process per CPU. Expected: `holes_best: 1`, `covered: false`, `restarts: 0` for every seed.\n\n## S2: independent scan and certificate (deterministic; hashes below)\n    node scan.mjs --residues <state.json> --level 19 --R 1283 --X 20000 --L 2000000 > scan-<name>.json\n    node scan.mjs --residues <state.json> --key producer_best_run_state --level 19 --L 2000000 --certify-longest 1 > cert-tracked-<name>.json\nGate: `node scan.mjs --residues n19-450s.json --level 19 --R 1283 --X 20000 --L 2000000 --certify-longest 1` gives window 1 hole at 159 and longest run 821 at 159, certified.\n\n## S3: complete K <= 2 closure (about 1 s per state)\n    node close2.mjs --residues <state.json> --level 19 --R 1283 > close2-<name>.json\nExpected: `proof_of_negative: true` on every seed state and on n19-450s.json. Controls: `close2.mjs --residues control-pc1.json --level 19 --R 1079` gives K1 solutions 2; `control-pc2.json` gives K1 0 and K2 2 (first solution i 10 -> 8, j 14 -> 28).\n\n## S4: restart defect\nRead `check-2043.py` lines 296–300: `elif stall % 25 == 0` (kick) comes before `elif stall % 250 == 0` (restart).\n\nWall time: about 5 + 30 + 10 min of search on a 2-CPU quota, and under 1 min for S2 and S3.","verification":null,"target":null,"finding":null,"human_md":null,"provisional":false,"effects_applied_at":null,"effort":"high","also_fix":null,"transcript_omitted":{"share":0,"omitted":0,"outputs":86},"patch_hash":null,"superseded_by":null,"duplicate_of":null,"transcript_resubmitted_at":"2026-09-22T21:13:17.846Z","file_notes":null,"research":{"outcome":"blocked","obstacle":{"kind":"scoped_obstruction","evidence":"Seeds 19001/19003/19004/19005 at 300 s (~260k steps) and 19002 at 1800 s (1,502,341 steps): holes_best 1 each; independent scans and CRT certificates; close2 K <= 2 complete enumeration on 7 states; restart dead code at check-2043.py lines 296-300.","statement":"The search-plus-exact-residual-repair class (check-2043 climb_fixed_R + reassignment closure) does not recover the published A144311(19) = 1283: 5/5 seeds stop at exactly 1 hole at up to 1.5M steps, and every residual state is closure-negative at K <= 2 (K <= 4 on #1363 state; K >= 5 priced beyond the 2 CPU-h cap). The n = 25 bar (2454) cannot be reached by this class on the evidence.","assumptions":"The producer is run unmodified (restart branch unreachable). Closure = reassignment of distinct primes to any residue, scored on the producer buffer L = R + 200. Published A144311(19) = 1283 is the exact maximum (Wang 2024, single witness).","revisit_when":"An exact program (Wang C++ run to completion at n = 23..25, or route 73 branch-and-bound) is available, OR a versioned producer with the restart branch fixed is measured on the published ladder (n = 18..22) and recovers a published value above n = 17."},"route_id":90,"depends_on":[1363],"evidence_md":"Answers #1363's pre-registered question (seed or instrument?) at n = 19, R = 1283 = A144311(19). The producer check-2043.py was run unmodified, warm from warm-n19.json. **5/5 seeds stop at exactly 1 hole**: 19001, 19003, 19004 and 19005 at a true 300 s (~260k steps each), and 19002 at 1800 s (1.50M steps, 4.8x #1363's 450 s run). A 120 CPU-s contended panel gave 1,1,1,1,2. The seed changes when the plateau is reached (11.8–156.7 s), not its value. So the one-hole plateau belongs to the instrument, not the seed, and the failure branch fires as written.\n\nTwo things tighten this beyond #1363. (1) Exact closure: a complete K <= 2 enumeration (549 single + 138,776 pair reassignments to any residue, scored on the producer's own buffer) finds no 0-hole window on any of the 6 new residual states or on #1363's. Two states are strict K <= 2 minima. Positive controls repair 1- and 2-prime perturbations of a covered state. (2) Instrument defect: the restart branch of climb_fixed_R (line 300, `elif stall % 250`) comes after `elif stall % 25` (kick), so it never runs. restarts = 0 in all 10 runs here and in #1363. Extra budget buys kicks around one basin. The ladder calibration (recovery to n <= 17) therefore measures the instrument with its restarts disabled.\n\nCertified by a from-scratch BigInt CRT checker with no shared code: tracked best runs 1049–1139 at n = 19 (max 88.8 % of 1283), with both neighbours holes. The gate reproduces #1363's 1-hole window at 159, its 821 and 1079 runs, and A144311(1..7) by brute force.\n\nConsequence for route 90: search plus exact residual repair has not recovered a published value above n = 17 at any budget tried: 18 s, 25x, 120x the slot, and 5 seeds. Every residual is closure-negative at K <= 2 (K <= 4 on #1363's state), and K >= 5 is priced beyond the 2 CPU-h cap. The constructive branch needs an exact program. The one untested change inside this class is the missing restart. Minor: warm-n19.json's `recorded_best_run_in_state: 1043` does not describe those residues (their longest run is 623).","prior_art_md":"Updated online search, 2026-09-22 (this run): \"A144311 Jacobsthal function primorial new terms a(23) 2026\"; \"longest run consecutive integers covered ±1 modulo first n primes explicit construction local search\"; OEIS A144311 text record fetched directly; Nguyen preprint (preprints.org 202608.1299) retried.\n\n- OEIS A144311 is unchanged: 22 terms, the last 1709; a(17)-a(22) from J. Wang (Nov 2024, C++ program linked). No a(23+).\n- NEW: D. C. Tucker, \"The Atlas of Maximal Gaps: Exact Covering Enumeration for Primorial Sieves\", Zenodo record 22865056, 2026-09-20. It gives exact atlas enumeration of maximal twin-pair-center deserts W(p). Its certified table (cert_twin_deserts.csv) gives W(p) for 13 <= p <= 31 (66, 108, 150, 204, 258, 348 = A144311(6..11) + 1, consistent with OEIS) and W(37) >= 462 \"maximality provisional\". That provisional value is below the published A144311(12) + 1 = 528. The author states the implementation reaches its practical limit at p = 41 (n = 13), so it covers nothing at n >= 19 and gives no witness at n = 25.\n- Otherwise the route's known sources: Costello-Watts arXiv:1208.5342; Ziller-Morack arXiv:1611.03310 and Ziller arXiv:1903.11973/2007.01808 (one-class Jacobsthal, not the ±1 two-class object); Hagedorn h(n), n < 50 (one-class); the OEIS Jacobsthal wiki.\n- ACCESS GAP persists: preprints.org 202608.1299 returns HTTP 403 to this client too.\n\nEXACT REMAINING GAP (unchanged in kind): no published explicit covered run of length >= 2454 at n = 25, and no exact a(n) for n >= 23. This job adds that the search-plus-repair class, run unmodified, cannot recover the published a(19) with 5 seeds up to 1.5M steps. It also finds that this class's restart path is dead code, a change that has never been measured."},"research_route_id":90,"verification_plan":null,"verification_fingerprint":null,"review_admitted_at":null,"department_id":"dept_da55f23c995cabb5136f4e91","run_id":"run_210c6175680069f20f11e045","triage_lead":null,"revision_base_sha":null,"integration":null,"resolves":null,"handle":"Benjaminsen","job_brief":"First update the online prior-work search for this experiment. If existing work covers it, record that and stop; otherwise run this bounded sprint on the uncovered uncertainty. Use cited published numbers during pursuit; their reproduction belongs in later validation. Build on the supplied findings; do not reconstruct earlier research. Return concrete progress and its cheapest credible check, a useful result for review, or a precisely scoped obstacle. Continued investment requires a distinct experiment.\n\nRead GET <project base>/research-routes/90 and return #1363. Return the ordinary report and transcript plus research: {route_id: 90, outcome: \"promising|progress|blocked|inconclusive|known|result\", evidence_md: \"what the evidence changes, <=4000 chars\", prior_art_md: \"updated online search record, sources and exact remaining gap, <=4000\", next_step: {question, method, success, failure, budget_hours} <only for continued pursuit>, obstacle: {kind, statement, assumptions, evidence, revisit_when} <for blocked/inconclusive>, depends_on: [<return ids actually required>]}. A result with a distinct next_step requests review and continues pursuit concurrently; omit next_step when no further experiment is warranted. Use known with prior_art_md and no next_step or obstacle when cited prior work already covers the proposed contribution; it stops automatic investigation without requesting review. The evidence grade is separate. Do not close a broad route because one proof attempt failed.","review_deferred":false,"in_triage":false,"triage":[],"verification_runs":[],"verification_state":null,"verification_summary":null,"canonical_return":null,"review_history":[],"dependencies":[{"id":"1363","status":"recorded","final_rung":"recorded","canonical_return_id":null}],"research_url":"/projects/twin-primes/research-routes/90","transcript_url":"/projects/twin-primes/return/1405/transcript","files":[{"sha256":"1d10848e66ee40620d850638f69dbdcc6901146d588ccada0416b7c3e7e83957","name":"scan.mjs","bytes":4657},{"sha256":"05ba97cf85a5210d1a8343b8abf45560c8a8bf0e88c72bea70bc3e799f766b53","name":"close2.mjs","bytes":3054},{"sha256":"6b393d0aabf491d0fb070bc9e4f5564bf15579c282e0db2e62490532b57d5c9a","name":"cert-tracked-1800s-seed19002.json","bytes":547},{"sha256":"1510816a0e2af4dce2ede87eaa961baa07812ddf03d0154bc20fc5786d0f7acb","name":"cert-tracked-cpu-seed19001.json","bytes":551},{"sha256":"c25e565419762bd39c5f179d951b6fca3cdeb3e548f609a2fdbfbc9dc9b5d412","name":"cert-tracked-cpu-seed19003.json","bytes":552},{"sha256":"a0dd7e6171ec0062d4e29ee47768205691010e75676733dc5cc654229c94163d","name":"cert-tracked-cpu-seed19004.json","bytes":550},{"sha256":"0119812be0361847724bca0555a8701127cb49239a293bef94fe559fff1c1aa4","name":"cert-tracked-cpu-seed19005.json","bytes":548},{"sha256":"0ad1ec8c6d5cf429ee9909e57a3a34d1e2dc4c26e0e9ffd497cb4524a1df3fcc","name":"close2-control-pc1.json","bytes":770},{"sha256":"c58ac7bdc4ef4189c8f022277d5a9f0a6e4eab86453c305c725048acf61351e9","name":"close2-control-pc2.json","bytes":757},{"sha256":"8d43d7dbc13db44edf6d753d120641121fe9217b06af33633a66ccb3daa16c54","name":"close2-cpu-seed19001.json","bytes":737},{"sha256":"f25832657bee6a592ab4bdab483e28bce873a8eb6f1f0ab3bb1f719280cf478f","name":"close2-cpu-seed19003.json","bytes":700},{"sha256":"b324dbea9720336b622e007e98ccc37ee11edd3c7681a491a884c41a4d3ca368","name":"close2-cpu-seed19004.json","bytes":740},{"sha256":"85172b93016d03b368015f01b1f8090dda9b0806ade1a44929a1aa3189a690eb","name":"close2-cpu-seed19005.json","bytes":674},{"sha256":"9a8a67de114250c53c173171ebeb046eae805824c366bf59676a25d50beabe4b","name":"close2-gate-1363-n19-450s.json","bytes":714},{"sha256":"17f53ca08baca12117cd8cb51996aaf464257480dcadd78ba325d9250ccf462b","name":"close2-seed19002-1800.json","bytes":686},{"sha256":"fe604920a9181eb449ce2226377ea7be9b0b3950c05a91d3f0c93db96d91c598","name":"close2-seed19005.json","bytes":713},{"sha256":"32f4bbfd74942e1d0caa2c50e535fbf09e22185937ee490c87ccd3158c63d173","name":"control-pc1.json","bytes":145},{"sha256":"23bb30846a0e3bb4528d66010dcd8f887a481e7ce8cbf259180116d48a86e25c","name":"control-pc2.json","bytes":111},{"sha256":"69b441714a902e067ec0062b73cc26d69686486c294eb5454d07fc9aacab0712","name":"gate-scan-1363-n19-450s.json","bytes":652},{"sha256":"9a47db0d5174f31d7724b77147e383b86c080e2d634fff378ac050a8810c19b8","name":"gate-scan-1363-runstate-n19.json","bytes":548},{"sha256":"ba70991b02a80f69d0d60339da74a635f4c4e4e0c7b06efc58463e27042d238d","name":"n19-1800s-seed19002.json","bytes":813},{"sha256":"79c0e7d66387c50374825ae0771b1592f883268c9a0e9dca89af4f9099f6545e","name":"n19-300s-cpu-seed19001.json","bytes":816},{"sha256":"ef7b673bf59ca8fd64b37065fd60448cb7476b50924a52f4f42f87b6bb147baf","name":"n19-300s-cpu-seed19003.json","bytes":820},{"sha256":"05938e1a5ef1e49e8b548bea94131c094585e9546b4e512e3f8a7ba757b3a2df","name":"n19-300s-cpu-seed19004.json","bytes":816},{"sha256":"d5eb2c88fb3951af80613037d1f945d2e3b04b7fd884dfa054e7020cf1681d87","name":"n19-300s-cpu-seed19005.json","bytes":813},{"sha256":"6cf5577587acc62829553ff64ee5059d5ec11d59c74a1d712d249a03576cc903","name":"n19-300s-seed19001.json","bytes":813},{"sha256":"edbcde14e3bb37965fa0059290b98779707d63cb26108898cea1f2f01453f683","name":"n19-300s-seed19002.json","bytes":809},{"sha256":"99d2279b6356d8266195ca92a8dfc061d5925ad1ed30f19bdceff7aa23f9da27","name":"n19-300s-seed19003.json","bytes":819},{"sha256":"9127f6cefafb7d1db32e5b65c0af321e181423e0a1986c2e96692fee92100257","name":"n19-300s-seed19004.json","bytes":815},{"sha256":"8f9783db41a387553bec5cea8dc0ceb7a7d1a8a71b9263ebc4ff6a518d45f1ee","name":"n19-300s-seed19005.json","bytes":776},{"sha256":"37e1e74a91dc353ab28ab6b6ed4201fcf042223a9e514dfbe3433fcfd95f4ddd","name":"scan-1800s-seed19002.json","bytes":387},{"sha256":"71b4fd0a73876a048048e781f208a8528c86c9122e6baef5a068a59f1fdc45a6","name":"scan-300s-cpu-seed19001.json","bytes":389},{"sha256":"18f997571647e3f5037550f0c33afac911d55de7ba73a6de22bb6179672b01e5","name":"scan-300s-cpu-seed19003.json","bytes":391},{"sha256":"deab5aca0398a76c0065b2ff4b3d4601383c1f17fcfd19eefcb682d6ef37cb60","name":"scan-300s-cpu-seed19004.json","bytes":389},{"sha256":"c8bc8e9724cbdd85c48b6935c1fd26ae5a436e39b287d9b270746a880bab1ada","name":"scan-300s-cpu-seed19005.json","bytes":387},{"sha256":"c1a95aafb29f6aef3574aa6c0e5bd0c121301100c1bf665ff209e6f89c5897fe","name":"scan-300s-seed19005.json","bytes":398},{"sha256":"662b29dcc673e127ea315d5ba3e4e7522c01019ce299602c88f39d77b6c7a8f8","name":"scan-warm-n19.json","bytes":408}],"decided_by_author_handle":false,"reviews":[],"decisions":[],"decision":null,"duplicates":[],"cited_messages":[{"id":2674,"channel_path":"measure","handle":"natepac","model":"claude-fable-5-1","kind":"claim","body_md":"Claim #2747 (pursue route 90 rev 10, measure; #1363's next step, @maxime-fleury). Running the producer's climb_fixed_R (check-2043.py, sha 38991a3b…, read-only) at n = 19, R = 1283: 5 distinct seeds x 300 s warm-started from the instrument's own plateau (warm-n19.json), then 1800 s on the best seed; per seed holes_best, best-window holes and longest covered run, verified by a fresh scanner; a 0-hole state gets certified. Pre-registered: success = any seed covers 1283; failure = all >= 1 hole.","created_at":"2026-09-21T03:53:35.159Z","url":"/projects/twin-primes/chat/messages/2674"}]}