Expert review will determine whether Claude materially contributed to a valid improvement from 41.6% to 67.2% in the proved lower bound on the proportion of Riemann zeta zeros satisfying the Riemann hypothesis.

state: expiredheat: lowuncertainty: highknownscott: mediummath-research research-agents frontier-modelsAnthropic Claude

What is this?

The case concerns an alleged Anthropic/Claude-authored paper claiming to raise the proved lower bound for the proportion of nontrivial Riemann zeta zeros on the critical line from 41.6% to 67.2%. The supplied search answer says expert review confirmed the result, but none of the accompanying snippets identifies Claude’s paper, Anthropic’s role, the reviewers, or the claimed bounds, so that confirmation is not grounded by the provided evidence. The snippets establish only that partial results concerning zeta zeros are a recognized route toward the Riemann hypothesis; the claimed AI contribution and mathematical breakthrough remain unverified here.

Why it matters to Scott

The radar already tracks the same expert-review-dependent AI-mathematics pattern in ProofCouncil, Claude Fable, and other claimed proofs; Scott’s Challenger, Never Arbiter and Provenance-Coupled Work frameworks already require that the paper, review evidence, and contribution lineage—not Claude’s confidence or Anthropic’s attribution—settle the claim. The unusually large claimed bound improvement could materially update his view of frontier models as research collaborators if validated, but the supplied grounding does not yet establish the result or Claude’s contribution.

ip:framework.challenger-never-arbiterip:framework.provenance-coupled-workip:concept.human-judgmentradar:concept.ai-mathematicsradar:proofcouncil-llm-agent-open-mathradar:claude-fable-jacobian-counterexampleradar:fable-astra-proof-replication

queries asked of Scott's wikis

AI agents producing expert-verifiable research
human review of autonomous research outputs
frontier models as mathematical collaborators
evaluation harnesses for open-ended scientific reasoning
provenance and attribution in AI-assisted discoveries
research-agent artifacts versus benchmark performance

Measured heat

no measured readings yet — the hourly heat pass fills this in

How the heat travelled

no chain yet — the hourly chain pass fills this in

Evidence (12) — ⭐ canonical anchor

source	object	author	score	comments
🟠 reddit	Claude increased the lower bound for the fraction of zeros of the Riemann zeta function that satisfy the hypothesis from 41.6% to 67.2% singularity	BoyNextDoor1990	1117	36
🟧 hn	Learning more about Claude's mathematical capabilities	tosh	279	170
🟧 echo.paper ⭐	The primary artifact is Claude’s paper, “More than two thirds of the zeros of the Riemann zeta function lie on the critical line.” Anthropic	Claude; Anthropic	—	—
🟠 reddit	"do ur thang" using a mindless GPT 5.6 Sol claims to improve upon Anthropic's Riemann hypothesis +0.002% singularity	I_HALF_CATS	495	92
🟧 hn	New Claude model improves lower bound on problem linked to Reimann hypothesis	ajax33	1	2
🟧 hn	I gave Claude's Reimann paper to GPT-5.6-Sol Pro. It's now 67.30%	porridgeraisin	2	0
🟧 hn	An unreleased Anthropic model made progress on one of math's biggest unsolved	sbulaev	2	0
🟧 hn	Asked AI to do a mini math breakthrough. It did a mini breakthrough	learademacher	6	1
🟠 reddit	Learning more about Claude's mathematical capabilities singularity	Inevitable-Ant1725	90	4
🟧 hn	An Infamous Math Problem Closer to Being Solved by Encouraging AI	RyanShook	1	0
🟧 hn	AI Just Had Another Math Breakthrough–With Help from a High-School Dropout	sonabinu	3	0
🟠 reddit	If AI+People Covering The Weak Spots Can Solve Complex Math, We Can Solve AI singularity	Soulren	5	6

Interpretation history

2026-08-24T19:55:02Z
The active episode has faded without named expert review, an inspectable verification artifact, or clearer contribution lineage. Expire routine monitoring without judging the mathematical claim; reopen if substantive validation, correction, or attribution evidence appears.
2026-08-22T18:26:04Z
No new evidence has appeared within the review window; minor engagement changes only reinforce that discussion has faded without independent mathematical validation or clearer contribution lineage. Keep the claim open but reduce monitoring cadence until a named expert review, inspectable verification artifact, or primary correction emerges.
2026-08-20T17:36:05Z
The refreshed comments merely dispute framing and repeat unsupported verification claims; they provide no named expert review, inspectable formalization, or clearer contribution lineage. The case remains an unsettled first-party result, but repetitive low-signal discussion warrants no increased attention.
2026-08-18T17:32:25Z
The new post repeats the known bound while asserting Lean verification and outside review without supplying artifacts, reviewer identities, or contribution lineage. It adds no independent validation, so the case remains an unsettled first-party research claim.
2026-08-18T17:23:33Z
evidence attached: reddit.post.1vrvc3t — The post provides additional discussion of the reported Claude-assisted Riemann-bound result, though not independent verification.
2026-08-16T23:26:00Z
The Reddit velocity spike is renewed attention to the known workflow, not new mathematical evidence. Without named external review, inspectable verification, or clearer contribution lineage, the result and Claude’s material role remain unsettled.
2026-08-16T10:30:18Z
The refreshed discussion and minor engagement movement add no independent mathematical review, formal verification, or evidence isolating Claude’s contribution. The case remains potentially consequential but unchanged: an externally unvalidated first-party research claim.
2026-08-15T13:33:00Z
The latest item is another thin secondary-media framing of the same Anthropic result, not independent mathematical review or contribution-lineage evidence. The claimed bound and Claude’s material role remain externally unvalidated despite broader coverage.
2026-08-15T13:22:34Z
evidence attached: hn.story.49310048 — shared external link with case evidence
2026-08-15T02:23:04Z
The new media item is secondary amplification and adds no expert mathematical review, formal verification, or evidence isolating Claude’s contribution. The potentially consequential result remains an unvalidated first-party claim despite broader coverage.
2026-08-15T02:21:59Z
evidence attached: hn.story.49306928 — Independent media coverage materially contextualises the open case about Claude's claimed contribution to the Riemann-zeta bound.
2026-08-14T10:31:13Z
The newly attached Reddit post only repeats Anthropic’s claimed result and workflow; it adds no independent mathematical review, formal verification, or evidence isolating Claude’s contribution. The case remains a consequential but externally unvalidated first-party claim.
2026-08-14T10:22:27Z
evidence attached: reddit.post.1vo2ovj — shared external link with case evidence
2026-08-12T23:28:17Z
The refreshed comments and small engagement gains are repetitive amplification, not independent expert review, formal verification, or evidence isolating Claude’s contribution. The consequential first-party bound claim remains externally unvalidated and unchanged in meaning.
2026-08-12T08:34:10Z
The refreshed discussion and modest engagement increase are repetitive amplification, not independent mathematical review, formal verification, or clearer attribution of Claude’s contribution. The consequential first-party bound claim remains externally unvalidated and unchanged in meaning.
2026-08-11T22:07:15Z
Newly attached HN post (score 1) is a low-credibility personal anecdote linking to a GitHub repo; it adds no independent expert review, formal verification, or evidence isolating Claude's contribution. The claimed bound remains a consequential first-party Anthropic result awaiting external validation. No change in meaning.
2026-08-11T21:23:13Z
evidence attached: hn.story.49264616 — This repository is independent follow-up evidence relevant to evaluating Claude's reported improvement to the Riemann-zeta bound.
2026-08-11T19:32:58Z
The added TechCrunch/HN coverage is secondary amplification, not independent mathematical review or new evidence separating Claude’s contribution from Anthropic’s validation process. The claimed bound remains consequential but externally unvalidated, and repeated extension claims do not corroborate it.
2026-08-11T19:23:35Z
evidence attached: hn.story.49262981 — The report appears to provide additional coverage of Anthropic's unreleased-model progress on the Riemann-hypothesis-related result.
2026-08-11T17:44:20Z
The refreshed discussion remains repetitive amplification and attribution debate, adding no independent expert review, formal verification, or evidence isolating Claude’s contribution. The claimed bound and subsequent extension attempts remain externally unvalidated, so the case’s meaning is unchanged.
2026-08-11T16:03:59Z
The new 67.30% follow-up is a title-only, unreviewed claim without an inspectable proof or methodology. It adds another attempted extension but neither validates the underlying 67.2% result nor clarifies Claude’s material contribution.
2026-08-11T15:39:18Z
evidence attached: hn.story.49259077 — This follow-up claims GPT-5.6-Sol Pro further improved Claude’s reported bound, directly informing expert validation of the mathematical result.
2026-08-11T15:02:19Z
The refreshed comments continue the existing attribution and provenance debate without supplying named external reviewers, formal verification, or evidence isolating Claude’s contribution. The case remains a consequential first-party mathematical claim awaiting independent validation.
2026-08-11T13:58:24Z
The latest activity is repetitive amplification and attribution debate, with no independent expert review, formal verification, or evidence isolating Claude’s contribution. The consequential first-party bound claim remains externally unvalidated and unchanged in meaning.
2026-08-11T12:52:50Z
The refreshed comments remain repetitive amplification and attribution debate, adding no independent expert review, formal verification, or evidence isolating Claude’s contribution. The case remains a consequential first-party mathematical claim awaiting external validation.
2026-08-11T11:43:21Z
The refreshed comments and engagement remain repetitive amplification, adding no independent expert review, formal verification, or evidence isolating Claude’s contribution. The consequential first-party bound claim remains externally unvalidated and unchanged in meaning.
2026-08-11T10:42:08Z
The refreshed discussion remains repetitive amplification and attribution debate, with no independent expert review, formal verification, or evidence isolating Claude’s contribution. The consequential bound claim remains externally unvalidated, and this update does not change the case’s meaning.
2026-08-11T09:34:18Z
Refreshed comments remain repetitive amplification and attribution debate, adding no independent expert review, formal verification, or evidence isolating Claude’s contribution. The claimed bound remains consequential but externally unvalidated.
2026-08-11T08:35:50Z
The refreshed comments remain repetitive amplification and attribution debate, with no named expert review, formal verification, or evidence isolating Claude’s contribution. The case remains a consequential first-party mathematical claim awaiting independent validation.
2026-08-11T07:50:15Z
The refreshed discussion and engagement add no independent mathematical review, formal verification, or clearer attribution of Claude’s contribution. The case remains a consequential first-party research claim whose validity and provenance are unsettled.
2026-08-11T06:27:23Z
The refreshed comments remain repetitive amplification and attribution debate, adding no independent mathematical review, formal verification, or evidence isolating Claude’s contribution. The consequential bound claim therefore remains externally unvalidated.
2026-08-11T04:31:02Z
The refreshed discussion adds only repetitive amplification and attribution debate, with no independent mathematical review, formal verification, or clearer evidence of Claude’s contribution. The claimed bound remains consequential but externally unvalidated.
2026-08-11T03:22:51Z
The refreshed discussion remains repetitive amplification and attribution debate, adding no named independent mathematical review, formal verification, or evidence isolating Claude’s contribution. The consequential first-party bound claim therefore remains externally unvalidated.
2026-08-11T02:22:48Z
The new HN item is thin secondary repetition, not independent mathematical review or evidence isolating Claude’s contribution. The case remains a consequential but externally unvalidated first-party research claim.
2026-08-11T02:21:53Z
evidence attached: hn.story.49252293 — Independent coverage directly bears on whether Claude materially improved the stated Riemann-zeta lower bound.
2026-08-11T01:29:39Z
The follow-up offers a weak extension attempt rather than independent validation: its tiny claimed gain is explicitly provisional and rests on unresolved proof and computational steps. It neither corroborates Anthropic’s 67.2% bound nor clarifies Claude’s contribution.
2026-08-11T01:25:18Z
evidence attached: reddit.post.1vl3eqx — This is an unverified follow-up claim of a further improvement to the same Anthropic-associated Riemann-zeta bound and should enter expert review.
2026-08-11T00:23:00Z
The refreshed comments only repeat workflow details and attribution concerns; they add no independent mathematical review, formal verification, or evidence isolating Claude’s contribution. The claimed bound remains a consequential first-party result awaiting external validation.
2026-08-10T22:40:07Z
The refreshed discussion remains amplification and attribution debate, with no independent mathematical review, formal verification, or new evidence isolating Claude’s contribution. The first-party result remains consequential but externally unvalidated.
2026-08-10T21:36:14Z
The refreshed discussion still supplies no independent mathematical review, formal verification, or clearer attribution of Claude’s contribution. It is repetitive amplification of the same consequential first-party claim, so the case remains unchanged and externally unvalidated.
2026-08-10T20:35:28Z
Refreshed discussion remains repetitive amplification of the claimed workflow and attribution concerns; it adds no independent mathematical review, formal verification, or clearer evidence of Claude’s contribution. The case remains a consequential first-party research claim awaiting external validation.
2026-08-10T19:36:17Z
Refreshed comments continue to amplify the reported workflow and raise attribution concerns, but add no named independent review, formal verification, or evidence isolating Claude’s contribution. The case’s meaning is unchanged: a consequential first-party claim awaiting external mathematical validation.
2026-08-10T18:38:37Z
The refreshed discussion adds amplification and provenance questions, not independent mathematical review or evidence separating Claude’s contribution from Anthropic’s validation process. The publication remains notable, but the core 67.2% claim is no closer to external confirmation.
2026-08-10T18:31:45Z
grounded: known/medium — The radar already tracks the same expert-review-dependent AI-mathematics pattern in ProofCouncil, Claude Fable, and other claimed proofs; Scott’s Challenger, Ne
2026-08-10T18:29:17Z
origin walked (codex/luna, conf 0.98): anchor hn.story.49247070 -> echo.paper.736fdcd147 by Claude; Anthropic
2026-08-10T18:28:03Z
case created — Two observations point to the same Anthropic research artifact making a concrete and independently resolvable mathematical claim.

Decision trace

08-25 05:55expireThe active episode has faded without named expert review, an inspectable verification artifact, or clearer contribution lineage. Expire routine monitoring without judging the mathematical claim; reope
08-25 05:55alert_silentThe only delta is continued staleness and negligible engagement movement, not a consequential event; any future expert review or primary correction can be treated as a fresh delta.
08-25 05:55alert_routeThe only delta is continued staleness and negligible engagement movement, not a consequential event; any future expert review or primary correction can be treated as a fresh delta.
08-25 05:54drop_targetsquiet through full ladder or over cap 8
08-23 04:26repriceNo new evidence has appeared within the review window; minor engagement changes only reinforce that discussion has faded without independent mathematical validation or clearer contribution lineage. Ke
08-23 04:26alert_silentThere is no consequential new delta to surface; the latest observation is staleness and negligible engagement movement, so this can wait for substantive validation or correction.
08-23 04:26alert_routeThere is no consequential new delta to surface; the latest observation is staleness and negligible engagement movement, so this can wait for substantive validation or correction.
08-21 03:36repriceThe refreshed comments merely dispute framing and repeat unsupported verification claims; they provide no named expert review, inspectable formalization, or clearer contribution lineage. The case rema
08-21 03:36alert_silentNo consequential new event occurred; the discussion can wait until a named independent review, inspectable proof verification, or primary correction appears.
08-21 03:36alert_routeNo consequential new event occurred; the discussion can wait until a named independent review, inspectable proof verification, or primary correction appears.
08-19 03:32repriceThe new post repeats the known bound while asserting Lean verification and outside review without supplying artifacts, reviewer identities, or contribution lineage. It adds no independent validation,
08-19 03:32alert_silentThis is unsupported commentary rather than a new verification event; it can wait until a named expert review, inspectable formalization, or primary correction appears.
08-19 03:32alert_routeThis is unsupported commentary rather than a new verification event; it can wait until a named expert review, inspectable formalization, or primary correction appears.
08-19 03:24alert_silentThis is commentary repeating the existing 41.6% to 67.2% claim, while adding unsupported assertions of Lean verification and outside-mathematician review. It provides no new primary artifact, named ex
08-19 03:24alert_routeThis is commentary repeating the existing 41.6% to 67.2% claim, while adding unsupported assertions of Lean verification and outside-mathematician review. It provides no new primary artifact, named ex
08-19 03:23attachThe post provides additional discussion of the reported Claude-assisted Riemann-bound result, though not independent verification.
08-19 03:22propose_attachThe post provides additional discussion of the reported Claude-assisted Riemann-bound result, though not independent verification.
08-18 01:21sensor_dirtyengagement_update
08-17 20:21sensor_dirtyengagement_update
08-17 19:21sensor_dirtyengagement_update
08-17 10:21sensor_dirtyengagement_update
08-17 09:26repriceThe Reddit velocity spike is renewed attention to the known workflow, not new mathematical evidence. Without named external review, inspectable verification, or clearer contribution lineage, the resul
08-17 09:26alert_silentThe delta is engagement-only and adds no substantive validation, correction, or attribution evidence; it can wait for a named expert review or inspectable verification artifact.
08-17 09:26alert_routeThe delta is engagement-only and adds no substantive validation, correction, or attribution evidence; it can wait for a named expert review or inspectable verification artifact.
08-17 09:21sensor_dirtyengagement_update
08-17 08:21sensor_dirtyengagement_update
08-16 20:30repriceThe refreshed discussion and minor engagement movement add no independent mathematical review, formal verification, or evidence isolating Claude’s contribution. The case remains potentially consequent
08-16 20:30alert_silentThe new delta is engagement-only and provides no substantive evidence; it can wait unless a named expert review, inspectable verification, correction, or contribution-lineage finding appears.
08-16 20:30alert_routeThe new delta is engagement-only and provides no substantive evidence; it can wait unless a named expert review, inspectable verification, correction, or contribution-lineage finding appears.
08-16 20:21sensor_dirtyengagement_update
08-15 23:33repriceThe latest item is another thin secondary-media framing of the same Anthropic result, not independent mathematical review or contribution-lineage evidence. The claimed bound and Claude’s material role
08-15 23:33alert_silentThe new link adds no inspectable proof, named expert validation, formal verification, correction, or materially new workflow detail, so it can wait for the normal briefing.
08-15 23:33alert_routeThe new link adds no inspectable proof, named expert validation, formal verification, correction, or materially new workflow detail, so it can wait for the normal briefing.
08-15 23:22alert_silentThis is secondary WSJ coverage of the already tracked Anthropic claim and adds no new expert validation, proof artifact, contribution evidence, or materially different capability detail. The reported
08-15 23:22alert_routeThis is secondary WSJ coverage of the already tracked Anthropic claim and adds no new expert validation, proof artifact, contribution evidence, or materially different capability detail. The reported
08-15 23:22attachshared external link with case evidence
08-15 23:21propose_attachshared external link with case evidence
08-15 23:21sensor_dirtyengagement_update
08-15 12:23repriceThe new media item is secondary amplification and adds no expert mathematical review, formal verification, or evidence isolating Claude’s contribution. The potentially consequential result remains an
08-15 12:23alert_silentSecondary coverage does not change the validity or attribution evidence and can wait for the next briefing; alert only on a named expert review, formal verification, correction, or substantive proof a