Anthropic researchers report that Claude Mythos Preview helped discover improved cryptanalytic attacks, including a claimed structural weakness in the post-quantum candidate HAWK and flaws in a weakened version of a widely used encryption standard. Their CryptanalysisBench preprint presents 191 tasks and claims models can independently find previously unknown bugs, design flaws, and end-to-end attacks. The supplied sources characterize the work as preliminary and not currently affecting production systems; independent expert validation is still needed to establish whether the results represent practical novel cryptanalysis rather than benchmark-bounded performance.
2026-08-09T18:36:52Z
After eleven days, no inspectable independent verdict, replication, or implementation response has emerged, and the only trigger is staleness. The validation question remains unresolved, but this episode has faded beyond its useful monitoring horizon and can be reopened if substantive expert evidence appears.
2026-08-07T17:48:28Z
Refreshed comments and small engagement changes add no inspectable expert verdict, replication, or implementation response. The case remains an unresolved technical-validation question, but repetitive discussion no longer warrants frequent review.
2026-07-30T14:25:39Z
The new trigger exposes no inspectable expert verdict, replication, or implementation response beyond the existing independent-analysis lead. The case remains a substantive but dormant validation question, and engagement-only churn does not justify promotion or frequent review.
2026-07-30T09:24:24Z
The independent-analysis lead (hn.story.49099804) remains the only potentially substantive outside assessment, but its technical verdict is still not inspectable from available evidence. All other new activity is redistribution of Anthropic's first-party results with near-zero engagement. The concrete HAWK-256 attack keeps the hypothesis open, but ~48 hours without inspectable independent replication, expert verdict, or implementation response means the case has fully plateaued. No basis for promotion or urgency.
2026-07-30T08:21:19Z
The refreshed discussion and minor engagement growth add no inspectable independent verdict, replication, or implementation response. The concrete first-party attacks keep the question open, but repetitive amplification does not change its meaning or justify frequent review.
2026-07-30T07:24:11Z
The independent-analysis lead (hn.story.49099804) remains the only potentially substantive outside assessment, but its technical verdict is still not inspectable from the available evidence. All other new activity is redistribution of Anthropic's first-party results with near-zero engagement. The concrete HAWK-256 attack keeps the hypothesis open, but ~40 hours without inspectable independent replication, expert verdict, or implementation response means the case has fully plateaued. No basis for promotion or urgency.
2026-07-30T06:23:49Z
The trigger adds no inspectable technical verdict, replication, or implementation response beyond the existing independent-analysis lead. Practical novelty remains unresolved, and repetitive attention should not disturb the case’s extended passive-watch status.
2026-07-30T04:22:08Z
The latest trigger exposes no new inspectable expert verdict, replication, or implementation response beyond the existing independent-analysis lead. Practical novelty remains unresolved, and engagement-only churn should not pull the case out of extended passive monitoring.
2026-07-30T03:22:04Z
The latest attachment does not expose a new independent technical verdict, replication, or implementation response, so it adds no substance beyond the existing expert-analysis lead. Practical novelty remains unresolved and the case belongs on extended passive watch rather than repeated engagement-driven review.
2026-07-30T02:21:47Z
The latest trigger adds no inspectable expert verdict, replication, or implementation response; it is further attention around the same independent-analysis lead rather than substantive corroboration. Practical novelty remains unresolved, so the case should stay on extended passive watch.
2026-07-30T01:22:36Z
The latest trigger adds no inspectable independent verdict, replication, or implementation response; increased engagement around the cryptography-focused analysis remains attention rather than corroboration. The concrete first-party attack keeps the case open, but practical novelty is still unresolved.
2026-07-29T23:25:15Z
The latest trigger still adds no inspectable independent verdict, replication, or implementation response, so attention around the cryptography-focused analysis remains amplification rather than corroboration. The concrete first-party attack keeps the case open, but its practical novelty remains unresolved.
2026-07-29T22:27:57Z
The latest attachment still exposes no inspectable expert verdict, replication, or implementation response; rising discussion around the independent-analysis link is attention rather than corroboration. The case remains substantive but unresolved and should stay on passive watch for actual technical assessment.
2026-07-29T21:25:35Z
The new attachment still provides no inspectable independent verdict, replication, or implementation response; additional attention around the cryptography-focused analysis is not itself corroboration. The case remains substantive but unresolved, awaiting outside technical assessment rather than further amplification.
2026-07-29T20:25:57Z
The new trigger still provides no inspectable technical verdict, replication, or implementation response from the independent-analysis lead. Engagement around that analysis does not itself corroborate practical novelty, so the case remains on passive watch rather than advancing.
2026-07-29T19:28:00Z
The latest attachment still exposes no inspectable expert verdict, replication, or implementation response, so it does not convert the independent-analysis lead into corroboration. The case remains substantive but unresolved and should stay on passive watch for technical assessment rather than further amplification.
2026-07-29T18:24:42Z
The independent-analysis link remains potentially important, but the latest update supplies neither its technical verdict nor replication, and modest engagement adds no substantive corroboration. The case stays open at passive cadence pending inspectable expert assessment.
2026-07-29T17:27:18Z
The first apparently independent cryptography-focused assessment breaks the case’s derivative-amplification plateau and makes outside evaluation active rather than merely awaited. Its substance and verdict are not present here, however, so it does not yet establish practical novelty, replication, or enough independent support for promotion.
2026-07-29T17:21:36Z
evidence attached: hn.story.49099804 — Independent cryptography-focused analysis directly contextualizes whether Anthropic’s cryptanalysis results reflect practical capability rather than benchmark-limited pattern matching.
2026-07-29T17:21:36Z
evidence attached: hn.story.49099548 — shared external link with case evidence
2026-07-29T14:29:38Z
No independent expert review, replication, or implementation response has appeared in the ~36 hours since the case opened. All new activity is redistribution of Anthropic's first-party results with near-zero engagement. The concrete HAWK-256 attack keeps the hypothesis substantive, but the case is fully dormant — nothing expected within hours or days. Maintaining passive monitoring at extended cadence.
2026-07-29T13:33:39Z
No independent expert review, replication, or implementation response has appeared in the ~29 hours since the case opened. All new activity is redistribution of Anthropic's first-party HAWK-256 and weakened-AES results with near-zero engagement. The concrete attack keeps the hypothesis substantive, but the case is fully dormant pending outside expert validation — nothing expected within hours or days. Dropping to extended passive monitoring.
2026-07-29T12:36:02Z
New evidence attached since last reprice appears to be another redistribution of Anthropic's first-party results, with no independent review, replication, or implementation response. The case remains substantively unchanged, still awaiting outside expert validation. Continuing passive monitoring at extended cadence.
2026-07-29T10:29:54Z
No independent cryptographic review, replication, or implementation response has emerged; the latest trigger adds no substance beyond the known first-party attacks. The case remains open but plateaued, warranting only passive monitoring for outside technical assessment.
2026-07-29T09:35:15Z
No independent expert review, replication, or implementation response has appeared since the case opened ~15 hours ago. All new evidence is repetitive redistribution of Anthropic's first-party HAWK-256 and weakened-AES results with near-zero engagement. The concrete attack keeps the hypothesis substantive, but the case has fully plateaued into passive monitoring territory.
2026-07-29T08:30:14Z
The latest HN link (hn.story.49094023) is another redistribution of the already-known weakened-AES result with zero engagement. No independent expert review, replication, or implementation response has appeared since the case opened. The concrete HAWK-256 attack keeps the hypothesis substantive, but 12+ hours of repetitive amplification without any outside assessment means the case has cooled to a passive monitoring cadence.
2026-07-29T07:28:50Z
The new HN link merely redistributes the already-known weakened-AES result and adds no independent expert review, replication, or implementation response. The concrete HAWK-256 attack keeps the case open, but repeated amplification no longer warrants hourly attention.
2026-07-29T07:21:02Z
evidence attached: hn.story.49094023 — shared external link with case evidence
2026-07-29T06:27:31Z
The latest activity adds no independent expert review, replication, or implementation response beyond Anthropic’s existing HAWK-256 demonstration. Repetitive amplification leaves the practical-discovery hypothesis substantive but unresolved, with no basis for promotion or urgency.
2026-07-29T05:24:39Z
The latest activity remains repetitive distribution of Anthropic’s first-party HAWK-256 result, with no independent expert review, replication, or implementation response. The concrete attack keeps the hypothesis open, but there is no change in meaning or basis for promotion.
2026-07-29T04:25:11Z
The latest attachment is another distribution of Anthropic’s existing claim, with no independent expert review, replication, or implementation response. The concrete HAWK-256 attack keeps the case open, but repetitive amplification does not change its meaning.
2026-07-29T04:20:47Z
evidence attached: reddit.post.1v9l877 — shared external link with case evidence
2026-07-29T03:25:56Z
The newly attached material adds no independent cryptographic review, replication, or implementation response beyond Anthropic’s existing HAWK-256 demonstration. Repetitive amplification leaves the practical-discovery hypothesis unresolved and does not justify promotion or urgency.
2026-07-29T02:30:06Z
The latest activity still provides no independent cryptographic review, replication, or implementation response beyond Anthropic’s concrete HAWK-256 demonstration. Repetitive amplification does not change the case’s meaning, so it remains substantive but unresolved pending outside expert assessment.
2026-07-29T01:22:40Z
The apparent update adds no identifiable independent review, replication, or implementation response beyond the already-known first-party HAWK-256 demonstration. The case remains substantive but unresolved, with repetitive amplification providing no basis for promotion or urgency.
2026-07-29T00:22:09Z
No independent review, replication, or implementation response has appeared; the latest activity is repetitive amplification of the already-known first-party HAWK-256 demonstration. The concrete attack keeps the case open, but there is no new basis for promotion or urgent attention.
2026-07-28T23:24:14Z
The new attachment is another low-engagement distribution of Anthropic’s first-party report, not independent cryptographic review or replication. The concrete HAWK-256 attack keeps the case substantive, but its meaning remains unresolved pending outside expert assessment.
2026-07-28T23:21:11Z
evidence attached: reddit.post.1v9dz4x — shared external link with case evidence
2026-07-28T22:27:22Z
grounded: novel/none — No intersection found: the supplied Scott wiki and radar searches returned no hits, so there is no grounded basis to connect these cryptanalysis claims or their
2026-07-28T22:26:47Z
The published HAWK-256 key-recovery demonstration materially strengthens the case beyond benchmark performance by supplying a concrete attack, but it remains first-party and does not satisfy the hypothesis’s independent-review threshold. The case is now worth near-term expert follow-up rather than passive monitoring.
2026-07-28T22:21:29Z
evidence attached: hn.story.49090083 — Anthropic's practical HAWK-256 key-recovery demonstration is direct first-party evidence for the open cryptanalysis-capability case.
2026-07-28T21:28:22Z
No independent cryptographic review, replication, or downstream implementation response has emerged; the apparent additions remain derivative amplification of the authors’ claim. The case’s meaning is unchanged and should stay cool while awaiting expert assessment.
2026-07-28T20:28:32Z
The new posts add derivative reporting and stronger practical-impact framing, but no independent cryptographic analysis, replication, or implementation response. The core claim remains unresolved and the sparse engagement is repetitive amplification rather than corroboration.
2026-07-28T20:21:28Z
evidence attached: reddit.post.1v995bc — This independently framed report reinforces the claim that Claude found a potentially consequential cryptanalytic shortcut, though expert validation remains necessary.
2026-07-28T20:21:28Z
evidence attached: reddit.post.1v99ovm — This provides additional reporting on Claude finding a serious weakness in a post-quantum cryptography candidate, directly bearing on whether the result is practically meaningful.
2026-07-28T20:21:28Z
evidence attached: reddit.post.1v99cuk — shared external link with case evidence
2026-07-28T19:28:56Z
The newly attached material remains first-party amplification rather than independent cryptographic review or replication. The practical-discovery claim is therefore unchanged and should stay cool pending outside technical assessment.
2026-07-28T18:22:30Z
Anthropic’s first-party report adds detail and confirms the lab is standing behind the result, but it is not independent corroboration and does not establish practical discovery of unknown weaknesses. With discussion now flat and no expert replication or review, the case remains open but cools pending outside assessment.
2026-07-28T18:21:52Z
evidence attached: hn.story.49087091 — Anthropic's first-party cryptanalysis report is directly relevant evidence for the open case, though not independent corroboration.
2026-07-28T17:23:59Z
grounded: novel/none — No intersection found: there are no Scott wiki hits connecting this claim to his positions or projects, and no radar hits showing prior coverage. The supplied e
2026-07-28T17:23:16Z
origin walked (codex/luna, conf 0.98): anchor hn.story.49087057 -> echo.paper.f735e1fa0a by Lukas Fluri, Avital Shafran, Nicholas Carlini, Matthew Jagielski, Milad Nasr, Orr Dunkelman, Eyal Ronen, and Florian Tramèr
2026-07-28T17:22:26Z
case created — The technical paper and contemporaneous reporting appear to track one consequential claim about frontier-model cryptanalysis capability and warrant consolidation.