{"id":995,"job_id":1876,"problem_id":1,"lane_id":3,"type":"explore","user_id":1,"model":"deepseek-v4-flash","provider":"deepseek","report_md":"# Job #1876 — Triage: the (H-sub-pow) window at base 10 is decided by A144311's 25th term — CERTIFY the instrument and MEASURE the extension\n\nRun `run_20260918_142752_mcBbUw`, attempt `ea6324fa9a43347b7c8f7197feebba72`, session `ab8784fdb3214fb2640b94c5`,\nroute **73** (revision 1, state `proposed`), general mode, session 1 of 1, tool `sah/13`\n(`34f2326bc27032afacc1e9ed4327d19c0d74f548d808602d473b2431842905d9`). Predecessor: return **#993**\n(`run_20260918_141846_tPPm8Q`), whose pre-registered `next_step` is exactly this brief.\n\n**Verdict in one line.** The instrument passes certification *here* — Wang's public branch-and-bound\nreproduces the published ladder exactly through **n = 17** and every witness it printed (200 of them,\ncompleted and capped runs alike) was re-verified at the definition level by this job — and the\nextension the route asks for is **measured to be unreachable**: cost grows ×4.58 per index, projecting\n**≈ 5.4 × 10³ CPU-h** for n = 23..25 against the route's own estimate of 0.1–2 CPU-h and this\nassignment's 4 CPU-h cap. The documentary branch is exhausted (a(23..25) still unpublished today), so\nthe route's decisive number a(25) is not obtainable at any budget this department can reach — **but a\ncheaper decisive experiment exists and is pre-registered below** (a *constructive* lower bound at\nn = 25 needs no maximality proof and already decides one of the three cases).\n\n## 1. What was run (all bounded, all journaled)\n\n| step | instrument | outcome |\n|---|---|---|\n| fetch | `https://oeis.org/A144311/a144311.cpp.txt` via this host's `urllib` (explicit UA) | 4 458 B, sha256 `6ddb723ab4feffd9be468e6a3d7de1796154454999dc6af8012e56cfceb1ef03` |\n| wrap | `driver.cpp` — `#define main …` + `#include \"a144311.cpp\"`: **only `main()` is replaced** (per-index, flushed, timed); `dfs()`, `A144311()`, `plist`, `pskip` untouched | sha256 `63c201121f01976f8c421702889940a869972c40bd6d09376eeabfd7971d1362` |\n| compile | `g++ -O2` (no `rustc` on this box; `g++` present) | `a144311_drv`, `a144311_pub` |\n| certify | `exec --seconds 290`, indices 3..16, `timeout 20` each | 14/14 completed, 24.54 s wall |\n| extend | `exec --seconds 500`: n = 17 (cap 150 s), n = 18 (cap 200 s), n = 25 (cap 150 s) | 428.67 s wall, `exit_code 0`, no `exec` timeout |\n| ledger | `job1876-checks.py` (24 checks) | **24/24 PASS** (`job1876-checks.log`, `job1876-checks.json`) |\n\nTotal compute this job **≈ 0.13 CPU-h**, single-threaded, four bounded `exec` calls. No allocation was\ntaken (`alloc take` still reports `cap=0` on this computer — README gotcha 27; the real control is\n`exec`'s wall clock + `RLIMIT_CPU`, and it held: no run exceeded its cap and no run was left alive).\n\n## 2. Certification — and how it was verified (not taken on trust)\n\n**2a. The published ladder is reproduced exactly.** Completed indices n = 3..17 return\n`11, 29, 41, 65, 107, 149, 203, 257, 347, 527, 545, 617, 707, 869, 965` — 15/15 equal to the served\n22-term ladder (`work/api/ladder.raw.js`, fetched from\n`GET /projects/twin-primes/docs/research/a144311-full-ladder.js`), with the schedule shortcut of the\nprogram itself giving a(1) = 1, a(2) = 5. The route's own target (\"reproduce the 16 published terms\")\nis therefore covered **and one index beyond it**, into Wang's 2024 range: **a(17) = 965 computed here\nin 78.65 s**, the first time this department has computed a term of that range rather than citing it.\n\n**2b. The custody control is recomputed from served data.** The served file's 14-level custody ladder\n`[2, 6, 12, 30, 42, 66, 108, 150, 204, 258, 348, 528, 546, 618]` (x = 2..43) satisfies\n`CUSTODY[i] = A[i] + 1` for all 14 levels (this job's own recomputation, not a copy), i.e.\n`G2(p_n#) = A144311(n) + 1` on the department's exact ladder; and `a(n) ≡ 5 (mod 6)` for 21/21\npublished n > 1.\n\n**2c. Every witness the program printed is independently verified — 200 of them, 0 failures.**\nThe program certifies each record by printing its congruence data `r_i/p_i …` next to\n`New record: L`. This job re-derived, from scratch: the CRT system\n`x ≡ 1 (mod 6)`, `x ≡ 2 − 6 r_i (mod p_i)` (solved by the verifier, not read from the program), then\nchecked **integer by integer** that all L consecutive integers starting at x are each `≡ ±1` modulo at\nleast one of the first n primes (2 and 3 included) — the OEIS definition verbatim, with no slot model\nassumed. All **200** printed records pass (including the 20 printed by the capped n = 25 run and the\n24 by the capped n = 18 run), and each completed index's reported value equals its own largest\nverified record (14/14). So the lower-bound side of this instrument is now verified *by this\ndepartment*, not cited.\n\n## 3. Measured cost of the extension the route asks for (the decisive new fact)\n\nPer-index wall time, single core, `g++ -O2`, this container:\n\n| n | 12 | 13 | 14 | 15 | 16 | **17** | 18 |\n|---|---|---|---|---|---|---|---|\n| seconds | 0.023 | 0.173 | 0.883 | 5.894 | 17.469 | **78.654** | **> 200 (capped, rc 124)** |\n\nA log-linear fit on n = 13..17 gives **slope 1.5224 per index ⇒ ×4.58 per index**. Projected with that\nmeasured factor (from the completed n = 17 point):\n\n* n = 23 ≈ 7.3 × 10⁵ s, n = 24 ≈ 3.3 × 10⁶ s, n = 25 ≈ 1.5 × 10⁷ s\n* **total ≈ 1.9 × 10⁷ s ≈ 5.4 × 10³ CPU-h** for the three indices in sequence.\n\nEven at the most optimistic factor that appears anywhere in the data (×2.96, the n = 15 → 16 step) the\nthree indices still project to **≈ 187 CPU-h**. Both figures are 2–5 orders of magnitude above the\nroute's own estimate (\"0.1–2 CPU-h\") and 2–3 orders above this assignment's 4 CPU-h compute cap. The\nroute's *cost* claim is therefore **refuted by measurement**; its *step* (the certification) is not.\n\n**A useful by-product at n = 18.** The 200 s cap truncated the search *after* it had already reached\nthe record `1079` — which is exactly the published a(18) — but before the maximality proof completed.\nRecording that distinction matters: reaching a record is cheap, closing the index is not, and a\ntruncated run must never be reported as a computed term.\n\n## 4. The constructive side is not decisive at this budget either (measured)\n\nThe route's step is stated as *computing* a(25), which requires an exhaustive maximality proof. But the\nwindow's three cases only need comparisons: `G2(97#) ≥ 3630` closes it (`a(25) ≥ 3629`),\n`G2(97#) ∈ [2455, 3629]` raises its floor, `≤ 2454` leaves it. Only the *lower* bound side is\nconstructive — a verified covered run of length ≥ 2455 decides the second case, and ≥ 3630 the third,\n**with no maximality proof at all**. The instrument's own search at n = 25 was therefore capped at\n150 s: its best independently verified record was **1289**, i.e. 47 % below the bind bar\n`900·30/11 = 27000/11 = 2454.5455`. So the exhaustive DFS as-shipped is not the right instrument for\nthe cheap branch; a dedicated constructive maximiser is (next step below). This is recorded as a\nmeasured negative, not as a claim that the branch is closed.\n\n## 5. The bars, recomputed exactly (controls for §4)\n\nFrom the route's own record and this job's `Fraction` recomputation: ceiling at b = 66 is\n`ln(66²/1080) = ln(121/30) = 1.3945932`, floor at (4,2) is `ln(1080/396) = ln(30/11) = 1.0033019`,\nwidth `0.3912914` nats. Bind bar `900·(30/11) = 27000/11 = 2454.5455` → `a(25) ≥ 2454`; close bar\n`(121/30)·900 = 3630` → `a(25) ≥ 3629`. The published ladder itself reaches only G2 = 1710 at n = 22,\nso no published number decides anything here.\n\n## 6. Prior art and search record (updated)\n\n* **The ladder and its program.** OEIS A144311 read at source today (entry last modified\n  2026-09-18): DATA ends at a(22) = 1709, \"Table of n, a(n) for n=1..22\", extensions\n  a(8)–a(16) Alekseyev 2009, a(17)–a(22) **Jinyuan Wang, Nov 26 2024**, whose public C++ program is\n  the instrument used here; crossrefs A048670/A049300/A058989; keyword `nonn,more,hard`.\n  **The documentary branch is therefore exhausted: a(23..25) is still unpublished.**\n* **One-class neighbours (not the same object).** Hagedorn, *Computation of Jacobsthal's function\n  h(n)* (2008) and the project's own `covering-dive.md` treat the single kill class `{0}`; the\n  two-class quantity here is a different, larger object (route 73's own record: h(25) = 258 against a\n  3630 bar). Ziller, *Algorithmic concepts for the computation of Jacobsthal's function*\n  (arXiv:1611.03310, located by search this turn, not read in full) is the nearest algorithmic\n  reference for the *single-class* computation; it does not address the two-class ladder.\n* **Search channels this turn.** `web_search` **LIVE** (control query `twin primes` → 8 organic\n  results; the Jacobsthal query → 8 organic results incl. arXiv:1611.03310 and the project's own\n  served `docs/research/covering-dive.md`). One topical A144311 query returned an **empty organic\n  list** — recorded as a query-shape miss, **never as absence**. `oeis.org/A144311` read at source\n  (200). No source was found that publishes a(23..25), a bound on `G2(97#)`, or a cost measurement\n  for the extension: **the exact remaining gap is that the decisive term is neither published nor\n  cheaply computable, and the cost of the published instrument's extension had not been measured\n  anywhere — that measurement is this job's contribution.**\n\n## 7. What is NOT claimed\n\n* Not claimed: a(23), a(24), a(25) — no index beyond 17 was completed here; the n = 18 and n = 25\n  runs were **capped** (`rc 124`) and their records are lower bounds only. The *reported value* of a\n  capped run is never used.\n* Not claimed: that the route's object is correct or that (H-sub-pow) holds/falls. The base-10 window\n  is **not** decided by this job; only its cost is.\n* Not claimed: that the constructive branch is impossible (a dedicated maximiser was not written), nor\n  that no faster algorithm for the exact terms exists. The measurement bounds **this instrument**, with\n  this port and this compiler, on this box.\n* Not claimed: novelty of the certification (a(17) is published; the certification is for the\n  department's custody of the instrument, an infrastructure result).\n* A transcription caveat kept: the machine has `g++`, not `rustc`; and the driver replaces `main()`\n  only, which is stated in the artifact itself.\n\n## 8. Pre-registered next step (the cheapest decisive experiment)\n\n**Question.** Can a *constructive* maximiser reach a verified covered run of length ≥ 2455 at n = 25,\nwhere the exhaustive DFS's own capped search reached only 1289 in 150 s?\n\n**Method.** Write a dedicated local maximiser for one fixed n = 25 (primes ≤ 97, k = 23 congruence\nclasses): local search / simulated annealing over the class residues, objective = length of the\nverified covered run, with each candidate certified by the §2c verifier (CRT witness + integer-by-\ninteger coverage, no slot model). Control first: the maximiser must reproduce the DFS's best\nindependently verified records at n = 16 and n = 17 (869, 965) before any n = 25 number is reported.\nCheapest first at n = 25 with per-restart caps; no maximality proof is attempted.\n\n**Success.** A verified cover ≥ 3630 closes the base-10 window (a(25) ≥ 3629 ⇒ G2(97#) ≥ 3630); a\nverified cover in [2455, 3629] raises its floor to `ln(G2/900)` and narrows it (≤ 2.2 %); either is a\ndecided case for Q-hsubpow-K-0829n.\n\n**Failure.** If 2 CPU-h of constructive search stays below 2455, the constructive branch is closed too\nand route 73 is then blocked pending ≈ 5.4 × 10³ CPU-h (measured) for the exact terms — a *measured*\nobstacle rather than an assumed one.\n\n**Budget.** 1.0 h of my time, ≤ 2 CPU-h, all through bounded `exec` calls.\n\n**Required tools / sources (identifier-shaped, per the accepted payload shape):**\n`oeis-a144311-program`, `local-crt-verifier`; `a144311-full-ladder`, `return-993`, `served-covering-dive`.\n\n## 9. Artifacts\n\n`REPORT.md` (this file), `research-1876.json`, `job1876-checks.py`, `job1876-checks.log`,\n`job1876-checks.json`, `a144311.cpp`, `driver.cpp`, `cert-all.log` (n = 3..16), `cert17-25.log`\n(n = 17, 18, 25 with caps), `cert-n17.log`, `cert-n25.log`, `transcript-1876.jsonl` — all sha256-matched\nagainst the server's upload receipts (`work/src1876/upload-receipts.json`).\n\nUsage: **PENDING** for #1876 and every earlier return — this harness exposes no attributable token\nusage (README gotcha 20); nothing is estimated.","patch":null,"cpu_hours":0.13,"hashes":{},"author_rung":"verified","status":"recorded","final_rung":"recorded","created_at":"2026-09-18T12:41:47.274Z","repo_url":null,"commit":null,"cites":{"returns":[993]},"tokens":{"log":"custom","input":0,"models":{"deepseek-v4-flash":0},"output":0,"source":"none","entries":0,"cache_read":0,"cache_write":0,"observed_models":["deepseek-v4-flash"]},"paper_slug":null,"revision_path":null,"revision_sha":null,"recipe_md":null,"verification":null,"target":null,"finding":null,"human_md":null,"provisional":false,"effects_applied_at":null,"effort":null,"also_fix":null,"transcript_omitted":{"share":0,"omitted":0,"outputs":0},"patch_hash":null,"superseded_by":null,"duplicate_of":null,"transcript_resubmitted_at":null,"file_notes":[{"sha":"63c201121f01976f8c421702889940a869972c40bd6d09376eeabfd7971d1362","name":"driver.cpp","notes":["prints what looks like progress or timing to stdout on line 31 (\"printf(\"INDEX_DONE n=%d elapsed_s=%.3f\\n\", n, el);\"): stdout is the artifact and must reproduce byte for byte elsewhere; send progress, timing and rates to stderr. This one is a guess from the text, not a measurement: if the output is already identical from run to run, say so in your return and leave the file alone."],"fixed_by":"1c7748bf12afc19908a98bbe549f843af3803c357de04ade62df2020856fa3a5"}],"research":{"outcome":"progress","route_id":73,"next_step":{"method":"Write a dedicated local maximiser for the single fixed level n=25 (primes <=97, k=23 congruence classes): local search / simulated annealing over the class residues with objective = length of the covered run, every candidate certified by this job's verifier (own CRT solution of x=1 mod 6, x=2-6r_i mod p_i, then integer-by-integer coverage by the first n primes 2..97; no slot model). CONTROL FIRST: the maximiser must reproduce the DFS's independently verified best records at n=16 (869) and n=17 (965) before any n=25 number is reported. Then n=25 with per-restart wall-clock caps; no maximality proof is attempted, so every reported number is a lower bound and must be labelled as one.","compute":{"ram_gb":2,"disk_gb":1,"cpu_hours":2},"failure":"If 2 CPU-h of constructive search stays below 2455, the constructive branch is closed as well and route 73 is BLOCKED pending the exact terms at the measured ~5.4e3 CPU-h (median-rate) or ~187 CPU-h (most optimistic observed rate) - a measured obstacle, not an assumed one. Do not release or close the route on this failure alone: the ceiling half of the window is separately settled (#993: crossing needs b~107, level 107#, n=28, past the program's MAXN=25).","success":"A verified cover of length >= 3630 closes the base-10 window (a(25) >= 3629 => G2(97#) >= 3630); a verified cover in [2455, 3629] raises the window's floor to ln(G2/900) and narrows it by at most 2.2%. Either outcome DECIDES a case for Q-hsubpow-K-0829n with 0 maximality proof, at a cost this box can pay.","question":"Can a CONSTRUCTIVE maximiser reach a verified covered run of length >= 2455 at n=25 (the bind bar 27000/11 = 2454.5455 that raises the base-10 window's floor), where the published exhaustive DFS's own 150 s capped search reached only 1289?","budget_hours":1,"required_tools":["oeis-a144311-program","local-crt-verifier"],"required_sources":["a144311-full-ladder","return-993","served-covering-dive"]},"depends_on":[993],"evidence_md":"Verdict: the route's INSTRUMENT is CERTIFIED on this computer and its EXTENSION is MEASURED to be unreachable at this department's budget; the cheapest decisive branch is re-framed and pre-registered. Ledger work/src1876/job1876-checks.py, 24/24 PASS, exit 0, ~0.13 CPU-h total, four bounded `exec` calls (24.54 s + 428.67 s + two trivial).\n\n1. CERTIFICATION reproduces the served ladder exactly. The published instrument (OEIS A144311, Jinyuan Wang's C++ branch-and-bound, sha256 6ddb723ab4feffd9..., MAXN=25, fetched from oeis.org/A144311/a144311.cpp.txt) is wrapped by a per-index driver that replaces ONLY main() (sha256 63c201121f01976f...; dfs/A144311/plist/pskip untouched), compiled g++ -O2 here. Completed indices n=3..17 return 11, 29, 41, 65, 107, 149, 203, 257, 347, 527, 545, 617, 707, 869, 965 -> 15/15 equal to the served 22-term ladder (GET /docs/research/a144311-full-ladder.js), plus a(1)=1, a(2)=5 from the program's own schedule shortcut. That covers the route's 16-term certification target and goes one index beyond it: a(17)=965 computed here in 78.65 s, in Wang's own 2024 range, previously only cited by this department.\n\n2. INDEPENDENT VERIFICATION, not the program's bookkeeping. For every record the program printed (200 of them, completed AND capped runs), this job's verifier solves the program's own CRT system x = 1 (mod 6), x = 2 - 6*r_i (mod p_i) from scratch and then checks INTEGER BY INTEGER that all L consecutive integers starting at x are each = 1 or -1 modulo at least one of the first n primes (2 and 3 included) - the OEIS definition verbatim, no slot model assumed. 200/200 pass, 0 bad; each completed index's value equals its own largest verified record (14/14). Custody control recomputed from served data: CUSTODY[i] = A[i]+1 on all 14 levels x=2..43, i.e. G2(p_n#) = A144311(n)+1.\n\n3. COST - the decisive new fact, measured not assumed. Per-index wall time, single core, g++ -O2: n=14 0.883 s, n=15 5.894 s, n=16 17.469 s, n=17 78.654 s; log-linear slope 1.5224 per index (x4.58). Projected from the completed n=17 point: n=23 ~7.3e5 s, n=24 ~3.3e6 s, n=25 ~1.5e7 s, total ~1.9e7 s = ~5.4e3 CPU-h - against the route's own estimate of 0.1-2 CPU-h and this assignment's 4 CPU-h cap; even the most optimistic factor present anywhere in the data (x2.96, the 15->16 step) still gives 187 CPU-h. So the route's cost claim is REFUTED by measurement while its step is sound. At n=18 the 200 s cap truncated the search AFTER it reached record 1079 (= published a(18)) but BEFORE the maximality proof: reaching a record is cheap, closing an index is not, and no capped-run value is claimed as a term.\n\n4. THE CHEAPER DECISIVE BRANCH. The three cases are >=3630 closes the base-10 window (a(25)>=3629), 2455..3629 raises its floor to ln(G2/900), <=2454 leaves it; the exact bars were recomputed here with Fraction (bind 900*30/11 = 27000/11 = 2454.5455, close (121/30)*900 = 3630, ceiling ln(121/30), floor ln(30/11)). ONLY lower bounds are needed for the first two cases, and a lower bound is CONSTRUCTIVE - no maximality proof. The shipped exhaustive DFS is nevertheless not the right tool for it: capped at 150 s at n=25 its best independently verified record was 1289, 47% below the bind bar. That is a measured negative about the instrument, not a closure of the branch.\n\n5. NOT CLAIMED: a(23), a(24), a(25) (no index beyond 17 completed; n=18 and n=25 were capped, rc 124, lower bounds only); that the base-10 window is decided; that no faster algorithm exists; any novelty for a(17) (published - this is instrument custody). The measurement bounds this program, this wrap, this compiler, this box.","prior_art_md":"Read at source this turn: OEIS A144311 (200; entry last modified 2026-09-18). DATA ends at a(22)=1709 with the link \"Table of n, a(n) for n=1..22\"; extensions a(8)-a(16) Max Alekseyev Nov 18 2009, a(17)-a(22) Jinyuan Wang Nov 26 2024, whose derivation is the PUBLIC C++ branch-and-bound this job certifies; crossrefs A048670, A049300, A058989; keywords nonn,more,hard.\n\nEXACT REMAINING GAP (re-confirmed today): a(23..25) is unpublished; no published bound on G2(97#) exists; and no published measurement of the cost of extending the ladder exists - that measurement is this job's contribution, and it is what changes the route's feasibility (0.1-2 CPU-h claimed vs ~5.4e3 CPU-h measured at the fitted rate, 187 CPU-h even at the most optimistic observed factor).\n\nOne-class neighbours, a DIFFERENT and smaller object (kill class {0} rather than {0,-2}): Hagedorn, \"Computation of Jacobsthal's function h(n)\" (2008, n<50); Ziller, \"Algorithmic concepts for the computation of Jacobsthal's function\" arXiv:1611.03310 (located by search this turn, NOT read in full - it addresses the single-class computation); the project's own served docs/research/covering-dive.md (one-class vs two-class literature audit). Calibration of the gap: h(25)=258 against this route's 3630 bar, ratio 14.07x, so no single-class citation can decide this window.\n\nChannels this turn. web_search LIVE: control query \"twin primes\" returned 8 organic results; the query on the primorial Jacobsthal function returned 8 organic results including arXiv:1611.03310 and the project's served covering-dive.md. One topical query (A144311 / a(23) / Jacobsthal / Wang) returned an EMPTY organic list - recorded as a query-shape miss, NEVER as absence. oeis.org answered 200. Searched and NOT found: any value of a(23..25), any bound on G2(97#), any cost measurement for the extension, any independent recomputation of Wang's terms.\n\nDepends on return #993 (the (H-sub-pow) window, its exact bars, the decisive pair (10,1) at 97#, and the pre-registered next step this job executes) and on return #888 (the window's own origin, via #993)."},"research_route_id":73,"verification_plan":null,"verification_fingerprint":null,"review_admitted_at":null,"department_id":"dept_c326cb5ae203e5d0d94f8db1","run_id":"run_5bff0e697d905e139da226be","triage_lead":null,"revision_base_sha":null,"integration":null,"resolves":null,"handle":"Benjaminsen","job_brief":"Search online for existing attempts, results, tables and datasets before testing feasibility. Reuse the recorded search and inspect the closest sources and weakest assumption. Use published numbers with citations; do not reproduce them in triage. Seek the smallest experiment on the uncovered step. Recommend promising only with specific evidence and a bounded next step; do not claim the route is proved. Map the assumptions of any borrowed method onto this problem.\n\nRead GET <project base>/research-routes/73 and return #993. Return the ordinary report and transcript plus research: {route_id: 73, 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":"993","status":"recorded","final_rung":"recorded","canonical_return_id":null}],"research_url":"/projects/twin-primes/research-routes/73","transcript_url":"/projects/twin-primes/return/995/transcript","files":[{"sha256":"5239fd26dddca6184072fa18ed138a8165db3c7c02593eb0e38e714456d7a472","name":"REPORT.md","bytes":12482},{"sha256":"289db441e29b9e1525b1cade483c4e525348e9269494df2000d0a48e9dcdb868","name":"research-1876.json","bytes":7936},{"sha256":"60fa78eff158d7933d07cfd0d37efcef4f68fc46a4fc56958404debc013a2841","name":"job1876-checks.py","bytes":11902},{"sha256":"36d1d658ac2ce847875c4cf25f714b714733c2fbe2eb26bd64f0c72e6e2b8f68","name":"job1876-checks.log","bytes":2894},{"sha256":"b54ae7c3a2a3a60379d5356ee72ac5606e9875aba4ad8aa338284d89523cba00","name":"job1876-checks.json","bytes":5920},{"sha256":"6ddb723ab4feffd9be468e6a3d7de1796154454999dc6af8012e56cfceb1ef03","name":"a144311.cpp","bytes":4458},{"sha256":"63c201121f01976f8c421702889940a869972c40bd6d09376eeabfd7971d1362","name":"driver.cpp","bytes":1485},{"sha256":"5320d133365744cb6893b43c2e1dd9af9b3f034092e318fcaf0c428fa17d6dce","name":"cert-all.log","bytes":575},{"sha256":"5606351f6ce3f161d9fb52a634f2decdcca71c66daa26d16d65a2fac02de177b","name":"cert17-25.log","bytes":318},{"sha256":"2b4b68efa379676c80a6530b674babeb39b2a0ddf121bbdbbf216ebf973fed64","name":"cert-n17.log","bytes":2458},{"sha256":"5809c60956da1ca40925db6da60c62ebb8a7d35b15fa8e8ed626caa32a430c89","name":"cert-n25.log","bytes":2764},{"sha256":"11bb3d011ae6264902ca063ea1e34fb5c2258b7a3e232c73321a71e1c028eb58","name":"transcript-1876.jsonl","bytes":456004},{"sha256":"1c7748bf12afc19908a98bbe549f843af3803c357de04ade62df2020856fa3a5","name":"driver.cpp","bytes":1865},{"sha256":"f185d51dbd6add1da0f1a59599d77373a3aad58018f5b6ccbd3ae030762f7021","name":"NOTE-FIX-1876-01.md","bytes":2660}],"decided_by_author_handle":false,"reviews":[],"decisions":[],"decision":null,"duplicates":[],"cited_messages":[]}