{
 "question": "Naming, domain, and visual-glyph identity strategy for a deep-time human-knowledge nonprofit (currently \"Memory Commons\", domains memorycommons.com/.foundation) and its sibling agent-identity system (alters.id / AltersID: DIDs + verifiable credentials for AI agents). The project needs (1) a possibly-new name/domain that avoids the overused \"memory\" prefix in the AI space, and (2) a visual glyph identity \u2014 an ornamental mark that both humans and AI can easily remember and recognize, usable on websites, as agent identity avatars, and etched on physical deep-time artifacts. Research these questions with numbers and citations, flagging vendor claims vs primary sources: (A) TLD LONGEVITY AND PRICE ECONOMICS: which TLDs are safest for a 50+ year institutional hold \u2014 current ICANN price-cap status of .com (Verisign contract, 2024 renewal, allowed increases), .net, .org (the 2019 Ethos Capital / price-cap removal controversy and its outcome, current PIR pricing); documented price-hike cases in new gTLDs (Uniregistry/UNR 2017 price increases up to 3000%, .xyz premium games, Identity Digital/Donuts renewal pricing patterns); registry failure/transition cases (what happens to domains when a gTLD registry fails or a TLD is retired \u2014 EBERO process, .wed case); ccTLD sovereignty risks (.io Chagos Islands sovereignty transfer 2024-2026 status and what it means for .io domains, .af/.sy precedents of politically-removed ccTLDs, .ai Anguilla boom); wholesale price comparison at Cloudflare Registrar for .com/.net/.org/.info/.institute/.foundation/.earth/.world/.archive if it exists. (B) NAMING LINGUISTICS AND LONGEVITY: peer-reviewed or expert evidence on what makes names memorable and durable \u2014 phonestheme research, pronounceability across major language families, name length vs recall, the \"bouba/kiki\" effect for brand sound symbolism; case studies of long-lived institution names and renames (Wikipedia/Wikimedia, Mozilla, Internet Archive, Long Now Foundation's \"Clock of the Long Now\", Software Heritage) \u2014 how they chose names and whether the names aged well; evidence that generic-word names vs coined names age differently. (C) NAMESPACE CROWDING: current (2025-2026) AI products/projects using \"memory\" in name or core branding (mem0, Letta/MemGPT, LangMem, Memary, OpenAI memory features, Claude memory) and \"commons\" naming in knowledge projects (Creative Commons, Digital Public Goods, Wikimedia Commons) \u2014 how crowded is each term. (D) GLYPH / VISUAL IDENTITY SYSTEMS for human+machine memorability: Urbit sigils (@p patp visual sigils \u2014 design system, generator, how they encode identity); GitHub identicons and other hash-to-image avatar systems (robohash, jazzicon, blockies in Ethereum wallets) \u2014 collision/recognition properties; Japanese mon/kamon family crests as a centuries-old compact glyph system (design constraints, count, why they endured); hallmarks/maker's marks and their legal longevity; deep-time semiotics: the Sandia/WIPP long-term nuclear waste warning markers report (Trauth, Hora, Guzowski 1993), \"ray cat\" and atomic priesthood proposals, the radiation trefoil and biohazard symbol design history and recognition rates; ISO 7010 symbol standardization; the C2PA \"cr\" Content Credentials pin as a new machine-verifiable trust mark; research on symbol recognition and recall vs text (icon memorability studies); machine-readable+human-aesthetic dual codes (QR code with logo, ArUco/AprilTag fiducials, Spotify Codes, Apple App Clip codes) \u2014 can one mark be both ornamental and machine-decodable. (E) NAME CANDIDATE SEMANTIC FIELDS: for a project about preserving human knowledge/essence for future intelligence, survey the semantic fields and existing-use collisions of candidate roots: continuity, testament, provenance, ark/arca, keel, anchor, ember, kernel, seed, strata, bedrock, meridian, aeon/eon, epoch, canon, corpus, anima, ethos, legacy, heirloom, palimpsest, keystone, lodestar, polaris, beacon \u2014 for each check: major existing tech/nonprofit brand collisions, trademark-crowding signals, cross-language meaning hazards in major languages. Output: cited findings with confidence levels per section.",
 "summary": "For a 50+ year institutional hold, the verified evidence favors legacy gTLDs: .com wholesale pricing remains contractually constrained (~$10/yr cap, 7% raises allowed in only 4 of the next 6 years, no increase before 1 Sep 2026) under the NTIA-Verisign Cooperative Agreement that auto-renewed 30 Nov 2024; .org's 2019 sale to Ethos Capital was blocked by the ICANN Board in April 2020; and even a failed new gTLD does not go dark \u2014 ICANN's three-provider EBERO regime (proven on .WED in 2017, still running 7+ years later) can sustain a dead registry indefinitely \u2014 while .io carries a real but bounded sovereignty risk (an ISO 3166-1 evaluation at worst, followed by a 5-year, extendable phase-out). On naming, peer-reviewed sound-symbolism research shows a coined name can carry meaning on its own: bouba/kiki shape associations replicate at ~72% across 25 languages (with round/sonorant sounds generalizing more reliably than spiky ones, 22/25 vs 11/25), vowel connotations measurably shift brand preference when matched to the product's semantic field, fluently pronounceable names are evaluated more favorably, and a concrete phoneme recipe (continuants, nasals, front vowels, diphthongs; avoiding plosives/affricates) lifted pseudoword recall from 36-41% to 53% in a 2025 study. For the visual identity, Urbit's sigil system is the strongest verified template: 512 hand-drawn syllable glyphs compose deterministically into unique 2x2 marks for all 2^32 (~4.2B) IDs, demonstrating that an ornamental, human-recognizable glyph system can be fully generative from a pronounceable name \u2014 directly applicable to AltersID agent avatars. No claims survived verification for namespace crowding (section C), candidate-root semantic fields (section E), or most of the deep-time semiotics material (WIPP, trefoil, mon, C2PA, machine-decodable marks), so those parts of the question remain open.",
 "findings": [
  {
   "claim": ".com is the safest priced long-hold TLD in the verified record: the NTIA-Verisign Cooperative Agreement automatically renewed effective 30 Nov 2024 with existing requirements and price constraints unchanged (after NTIA and Verisign failed to agree on pricing changes), capping wholesale .com at approximately $10/domain/year, permitting 7% increases in only four of the following six years, and permitting no increase at all before 1 Sep 2026.",
   "confidence": "high",
   "sources": [
    "https://www.ntia.gov/blog/2024/com-cooperative-agreement-ensuring-internet-stability-and-security"
   ],
   "evidence": "NTIA (primary, the government counterparty): \"Per its terms, the Cooperative Agreement will automatically renew and take effect on November 30, with its current requirements and existing price constraints in place... The Agreement currently caps those prices at approximately $10 per domain per year, and would allow a 7% increase four out of the next six years... The current terms do not permit any increases in wholesale .com prices until September 1, 2026.\" Verifiers confirmed the renewal took effect and that Verisign's first post-renewal 7% increase ($10.26 to $10.97) was announced Apr 2026 effective Nov 2026 \u2014 consistent with the schedule. Note NTIA's 'absent the Agreement, higher prices would be possible' framing is contestable (the ICANN Registry Agreement independently caps prices), but is accurately attributed.",
   "vote": "9-0 (three merged claims, each 3-0)"
  },
  {
   "claim": "The 2019 .org privatization attempt was defeated: PIR announced on 13 Nov 2019 that Ethos Capital would acquire it from the Internet Society, converting PIR from a Pennsylvania nonprofit into a for-profit LLC; on 30 Apr 2020 the ICANN Board rejected (withheld consent for) the change of control, and PIR remains an ISOC subsidiary. Important scope limit: this decision did NOT reverse the separate June 2019 removal of .org price caps from the PIR registry agreement.",
   "confidence": "high",
   "sources": [
    "https://www.icann.org/en/blogs/details/icann-board-withholds-consent-for-a-change-of-control-of-the-public-interest-registry-pir-30-04-2020-en"
   ],
   "evidence": "ICANN Board Chair (primary): \"the ICANN Board made the decision to reject the proposed change of control and entity conversion request that Public Interest Registry (PIR) submitted... withholding consent of the transfer of PIR from the Internet Society (ISOC) to Ethos Capital is reasonable, and the right thing to do.\" Corroborated by EFF, CircleID coverage of 30 Apr 2020. Verifiers explicitly flagged that .org price caps removed in June 2019 stayed removed \u2014 so .org's governance risk resolved well, but its price-cap protection is weaker than .com's, and current PIR pricing was not verified.",
   "vote": "6-0 (two merged claims, each 3-0)"
  },
  {
   "claim": ".io sovereignty risk is real but bounded and slow-moving: as of ICANN's 14 Nov 2024 statement, the UK-Mauritius Chagos sovereignty transfer had been announced but not ratified; the 'IO' code persists in ISO 3166-1 with no change, and a completed transfer would likely only trigger an evaluation by the ISO 3166 maintenance committee, not automatic removal. Even if 'IO' were removed, ICANN's ccTLD retirement policy provides a five-year phase-out window, extendable (per the ccNSO policy, by up to five more years).",
   "confidence": "high",
   "sources": [
    "https://www.icann.org/en/blogs/details/the-chagos-archipelago-and-the-io-domain-14-11-2024-en"
   ],
   "evidence": "Kim Davies, VP IANA Services / President PTI (the office that runs ccTLD delegation \u2014 primary): \"'IO' persists in the ISO 3166-1 standard and there has been no change... likely to trigger an evaluation of the circumstances by the committee that maintains the ISO 3166-1 standard... a five-year time window will commence during which time usage of the domain will need to be phased out. That time window might be extended.\" Verifiers updated the timeline: treaty signed 22 May 2025; UK CRaG period completed Jul 2025 but implementing legislation and Mauritius approval incomplete, with the process reportedly paused/shelved April 2026; 'IO' remained in ISO 3166-1 and the IANA root as of mid-2026. Practical read for alters.id-style holdings: worst realistic case is 5-10 years of forced migration notice, not sudden loss \u2014 but .io (and any ccTLD) is governed by geopolitics, not ICANN gTLD contracts.",
   "vote": "9-0 (three merged claims, each 3-0)"
  },
  {
   "claim": "There is a tested, functioning backstop when a gTLD registry fails: on 8 Dec 2017 ICANN placed .WED into the EBERO program after its registry operator Atgron, Inc. suffered a Registration Data Directory Services (WHOIS) failure, and designated Nominet as emergency interim registry operator, which restored service. This remains the only EBERO activation to date and is the concrete precedent for 'what happens to domains in a failed new gTLD'.",
   "confidence": "high",
   "sources": [
    "https://www.icann.org/en/announcements/details/wed-placed-in-emergency-back-end-registry-operator-ebero-program-8-12-2017-en"
   ],
   "evidence": "ICANN announcement (primary): \"Registry operator, Atgron, Inc., which operates gTLD .WED, experienced a Registration Data Directory Services failure, and ICANN designated EBERO provider Nominet as emergency interim registry operator. Nominet has now stepped in and is restoring service for the TLD.\" Corroborated by Domain Incite and ICANN's May 2021 follow-up. One verifier nuance: the WHOIS outage was likely a deliberate shutdown after Atgron's back-end contract expired \u2014 a registry-abandonment scenario, which is arguably the more relevant precedent for a 50-year hold on a niche gTLD like .foundation.",
   "vote": "6-0 (two merged claims, each 3-0)"
  },
  {
   "claim": "EBERO mechanics quantify new-gTLD survival risk: activation is triggered by emergency downtime thresholds on any of five critical registry functions (DNS resolution, shared registration system, RDDS/WHOIS, data escrow, DNSSEC-signed zone); ICANN keeps three contracted providers (CIRA Canada, CNNIC China, Nominet UK) on five-year contracts; and while the goal is resolution within 12 months, ICANN and the providers can sustain an event as long as necessary \u2014 .WED has been under emergency operation for 7+ years. Domains in a failed gTLD do not simply go dark, but EBERO only maintains critical functions (a continuity backstop, not a guarantee of a healthy TLD, and a TLD with no successor can eventually be retired).",
   "confidence": "high",
   "sources": [
    "https://www.icann.org/resources/pages/faqs-2013-04-02-en",
    "https://www.icann.org/en/announcements/details/wed-placed-in-emergency-back-end-registry-operator-ebero-program-8-12-2017-en"
   ],
   "evidence": "ICANN FAQ (primary, verified live July 2026): \"They have entered into five-year contracts with ICANN to operate the five critical registry functions in the event of a gTLD registry operator failure. The currently contracted EBERO service providers are: Canadian Internet Registration Authority (CIRA), China Internet Network Information Center (CNNIC), Nominet\" and \"ICANN's goal is to have the emergency event resolved as soon as possible, preferably within 12 months. However, ICANN and the contracted EBERO providers are able to sustain an event as long as necessary.\" Verifiers confirmed via the IANA root database that .wed remained under emergency back-end operations as of Jan 2025, and that EBERO contracts were re-amended in Nov 2024 / Jan 2025 (program active). Caveat: EBERO does not cover some legacy gTLDs, and registrations are effectively frozen during an event.",
   "vote": "9-0 (three merged claims, each 3-0)"
  },
  {
   "claim": "Sound-shape symbolism is robust across cultures but asymmetric, which matters for coining a globally durable name: across 25 languages (917 participants, 9 language families, 10 writing systems), 'bouba' matched round and 'kiki' matched spiky shapes at an estimated 72% (95% CrI 56-82%); bouba was reliably matched in 22/25 languages while kiki was reliable in only 11/25. Round, sonorant-heavy sounds generalize across cultures more reliably than sharp/plosive ones \u2014 but Romanian, Mandarin, and Turkish fell below 50%, so the effect is strong, not universal.",
   "confidence": "high",
   "sources": [
    "https://royalsocietypublishing.org/rstb/article/377/1841/20200390/108730/The-bouba-kiki-effect-is-robust-across-cultures"
   ],
   "evidence": "Cwiek et al. 2022, Phil Trans R Soc B (peer-reviewed primary, largest cross-cultural bouba/kiki study as of 2026): \"The estimated posterior mean proportion of bouba/kiki matches across languages was 72%, with a 95% credible interval ranging from 56% to 82%... bouba was reliably matched with the round shape in 22 out of 25 languages, whereas kiki was reliably matched to the spiky shape in only 11 out of 25 languages.\" Verifier note: 'round/sonorant vs sharp/plosive' is a slightly loose gloss (bouba's /b/ is a voiced plosive), and application to branding is a standard but interpretive extension.",
   "vote": "6-0 (two merged claims, each 3-0)"
  },
  {
   "claim": "Phonetic symbolism measurably shifts brand-name preference, and the effect is context-dependent \u2014 a name's sound must match its intended semantic field. Lowrey & Shrum (2007, J. Consumer Research, two experiments): fictitious names whose vowel sounds connoted category-positive attributes (e.g., 'small, sharp' for a convertible or knife) were preferred, and the SAME names were preferred less when those connotations were category-negative (SUV, hammer). Klink (2000, Marketing Letters, two studies): a brand name's sound by itself communicates product attributes (size, speed, strength, weight), with or without supporting marketing communications. Implication for the project: a coined name can be engineered to sound like 'deep time / permanence / continuity' and will carry that meaning unaided.",
   "confidence": "high",
   "sources": [
    "https://academic.oup.com/jcr/article-abstract/34/3/406/1798924",
    "https://link.springer.com/article/10.1023/A:1008184423824"
   ],
   "evidence": "Lowrey & Shrum abstract (verbatim): \"Participants preferred brand names more when the attributes connoted by the vowel sounds (e.g., small, sharp) were positive for a product category (e.g., convertible, knife), but they preferred the same names less when the attributes connoted were negative.\" Klink abstract: \"Brand name sounds can convey product-related information either in the presence or absence of supporting marketing communications.\" Verifiers ran adversarial replication checks: the vowel sound-symbolism paradigm was one of only 2/10 sensory-marketing studies that replicated in the 2022 Frontiers replication project (though at roughly half the original effect size), Shrum et al. 2012 replicated cross-language (English/French/Spanish), while a related study (Yorkston & Menon 2004 fricatives) had mixed replication and Doyle & Bottomley 2011 flagged letter-shape confounds in written stimuli. Effects are real but small and lab-measured.",
   "vote": "12-0 (four merged claims, each 3-0)"
  },
  {
   "claim": "Processing fluency (ease of pronunciation) affects how favorably humans evaluate a named entity: in one laboratory study and two analyses of real stock-market data, fluently named stocks robustly outperformed disfluently named stocks in the short term (Alter & Oppenheimer 2006, PNAS). For naming, this argues for an easily pronounceable name over a clever but disfluent one.",
   "confidence": "medium",
   "sources": [
    "https://www.pnas.org/doi/10.1073/pnas.0601071103"
   ],
   "evidence": "PNAS 103(24):9369-9372 abstract (verbatim): \"In both a laboratory study and two analyses of naturalistic real-world stock market data, fluently named stocks robustly outperformed stocks with disfluent names in the short term.\" Verifier found no failed replications and consistent later work (Green & Jame 2013, J. Financial Economics: name fluency linked to investor recognition and firm value), but with boundary conditions (effect reverses in high-sentiment periods per Xing et al. 2016; weakens with more information per a 2023 study). Rated medium because the finding rests on a single (peer-reviewed, corroborated) primary source in the verified set.",
   "vote": "3-0"
  },
  {
   "claim": "A concrete phoneme recipe for a memorable coined name exists but is thinly evidenced: pseudowords built from 'appealing' phonemes per Crystal's phonaesthetic framework \u2014 continuants (/l/, /s/), nasals (/m/, /n/), front vowels (/i/, /e/), and diphthongs, while avoiding plosives, postalveolar fricatives, and affricates \u2014 were recalled at 53.25% vs 41.5% (neutral) and 36.0% (unappealing) among 100 native English speakers (p<0.001). Note the recipe predicts recall, not rated appeal (participants did not rate the recipe words as most appealing).",
   "confidence": "medium",
   "sources": [
    "https://pmc.ncbi.nlm.nih.gov/articles/PMC12674511/"
   ],
   "evidence": "Matzinger & Ko\u0161i\u0107, PLoS One, Dec 2025 (peer-reviewed primary): recall 53.25%/41.5%/36.0%, chi-sq=29.91, df=2, p<0.001; stimuli explicitly built to \"contain continuants (e.g., /l/, /s/), nasals (e.g., /m/, /n/), front vowels (e.g., /i/, /e/), and diphthongs\" and fewer \"plosives..., postalveolar fricatives..., and affricates.\" Rated medium despite 3-0 votes: single study, no replication yet, small marginal effect (R2m=0.03), written-modality stimuli (possible graphemic confound), English speakers only, 12 trisyllabic pseudowords. Interesting tension with bouba/kiki asymmetry: both point away from harsh plosive-heavy names for this project's warm/permanence semantic field.",
   "vote": "6-0 (two merged claims, each 3-0)"
  },
  {
   "claim": "Urbit's sigil system is a verified working template for the AltersID glyph requirement \u2014 a fully generative, deterministic name-to-glyph identity system: every one of the 2^32 (~4.2 billion) Urbit IDs gets a unique visual sigil; pronounceable @p names are built from two fixed sets of 256 three-letter syllables (512 total), each syllable was hand-drawn once as a distinct glyph, and planet-rank sigils compose four syllable glyphs in a 2x2 grid. Design cost was 512 manual drawings, not 4.2B designs \u2014 showing an ornamental glyph system can scale to arbitrary agent populations while staying tied to a pronounceable name.",
   "confidence": "high",
   "sources": [
    "https://urbit.org/blog/creating-sigils"
   ],
   "evidence": "Urbit design blog (primary, corroborated by the open-source urbit/sigil-js implementation, '~4.2 billion default profile pics'): \"There are 2^32 or 4.2 billion unique Urbit IDs... Urbit IDs are constructed from two sets of 256 three-letter syllables... Now all we had to do was manually draw 512 unique but consistent individual 'phonemes'.\" Verifiers confirmed the mapping is deterministic and injective (galaxies=1 glyph, stars=2, planets=2x2 grid of 4) and that the system is unchanged since ~2019. The third sub-claim (512 hand-drawn, 2x2 grid) passed 2-1 but the dissent was not substantive \u2014 the majority verified it against both the blog and the sigil-js repo. Scope note: 2^32 covers galaxies/stars/planets; moons/comets extend the address space but do not get sigils in practice.",
   "vote": "8-1 (two claims 3-0, one 2-1)"
  }
 ],
 "caveats": "Coverage gaps are the biggest caveat: of the five requested sections, only parts of A (TLD economics), B (naming linguistics), and D (glyph systems) produced claims that survived 3-vote verification. Nothing survived on: section C namespace crowding (mem0/Letta/LangMem/'commons' saturation \u2014 so the premise that 'memory' is overused remains plausible but unverified here); section E candidate-root semantic fields and trademark collisions; new-gTLD price-hike cases (Uniregistry ~3000%, .xyz premiums, Identity Digital/Donuts renewal patterns \u2014 directly relevant to holding memorycommons.foundation, since .foundation is an Identity Digital TLD with no price caps); the Cloudflare wholesale price table; .af/.sy/.ai ccTLD precedents; and most deep-time semiotics (WIPP/Sandia 1993 report, trefoil/biohazard history, ISO 7010, mon/kamon, hallmarks, C2PA, QR/ArUco/Spotify-code dual-decodability). Time-sensitivity: the .io claims are a 14 Nov 2024 snapshot \u2014 verifiers note the treaty was signed 22 May 2025, UK ratification steps completed Jul 2025 but Mauritius approval and implementing legislation were incomplete and the process was reportedly paused in April 2026; 'IO' remained in ISO 3166-1 and the IANA root as of mid-2026, but this needs re-checking before any decision. The .org Ethos rejection did not restore the price caps removed in June 2019 \u2014 current PIR pricing is unverified, so .org's long-term price protection is weaker than the governance victory suggests. Verisign's first post-renewal .com increase (7%, $10.26 to $10.97, effective Nov 2026) was already announced per verifier evidence. On linguistics, all effects are small, lab-measured, and the phonaesthetics recall study is single and unreplicated; sound-symbolism findings mostly but not universally transfer across languages (Romanian, Mandarin, Turkish under 50% on bouba/kiki).",
 "openQuestions": [
  "How crowded are 'memory' and 'commons' in 2025-2026 AI/knowledge-project branding in quantified terms (product counts, trademark filings), and does that crowding actually justify renaming away from Memory Commons?",
  "What is PIR's actual .org wholesale price trajectory since the 2019 cap removal, and what are current renewal prices for .foundation and other Identity Digital gTLDs \u2014 i.e., is memorycommons.foundation exposed to uncapped renewal pricing?",
  "Can a single mark be both ornamental and machine-decodable (C2PA 'cr' pin, ArUco/AprilTag fiducials, Spotify Codes, QR-with-logo error-correction budgets), or should the project pair a human glyph with a separate machine channel as Urbit does (glyph + @p name)?",
  "What is the final status of the Chagos treaty ratification and the ISO 3166-1 'IO' entry as of decision time, given the April 2026 pause \u2014 and does the alters.id ccTLD (.id, Indonesia) carry analogous sovereignty/policy risk that was never assessed?"
 ],
 "refuted": [],
 "sources": [
  {
   "url": "https://www.icann.org/en/blogs/details/icann-board-withholds-consent-for-a-change-of-control-of-the-public-interest-registry-pir-30-04-2020-en",
   "quality": "primary",
   "angle": "Domain policy and TLD longevity risk (primary/regulatory)",
   "claimCount": 5
  },
  {
   "url": "https://www.ntia.gov/blog/2024/com-cooperative-agreement-ensuring-internet-stability-and-security",
   "quality": "primary",
   "angle": "Domain policy and TLD longevity risk (primary/regulatory)",
   "claimCount": 5
  },
  {
   "url": "https://www.icann.org/en/blogs/details/the-chagos-archipelago-and-the-io-domain-14-11-2024-en",
   "quality": "primary",
   "angle": "Domain policy and TLD longevity risk (primary/regulatory)",
   "claimCount": 5
  },
  {
   "url": "https://www.icann.org/en/announcements/details/wed-placed-in-emergency-back-end-registry-operator-ebero-program-8-12-2017-en",
   "quality": "primary",
   "angle": "Domain policy and TLD longevity risk (primary/regulatory)",
   "claimCount": 4
  },
  {
   "url": "https://www.icann.org/resources/pages/faqs-2013-04-02-en",
   "quality": "primary",
   "angle": "Domain policy and TLD longevity risk (primary/regulatory)",
   "claimCount": 5
  },
  {
   "url": "https://royalsocietypublishing.org/rstb/article/377/1841/20200390/108730/The-bouba-kiki-effect-is-robust-across-cultures",
   "quality": "primary",
   "angle": "Naming science (academic/peer-reviewed)",
   "claimCount": 5
  },
  {
   "url": "https://academic.oup.com/jcr/article-abstract/34/3/406/1798924",
   "quality": "primary",
   "angle": "Naming science (academic/peer-reviewed)",
   "claimCount": 4
  },
  {
   "url": "https://link.springer.com/article/10.1023/A:1008184423824",
   "quality": "primary",
   "angle": "Naming science (academic/peer-reviewed)",
   "claimCount": 5
  },
  {
   "url": "https://www.pnas.org/doi/10.1073/pnas.0601071103",
   "quality": "primary",
   "angle": "Naming science (academic/peer-reviewed)",
   "claimCount": 5
  },
  {
   "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC12674511/",
   "quality": "primary",
   "angle": "Naming science (academic/peer-reviewed)",
   "claimCount": 5
  },
  {
   "url": "https://atlan.com/know/best-ai-agent-memory-frameworks-2026/",
   "quality": "blog",
   "angle": "Namespace crowding scan (recent market/practitioner)",
   "claimCount": 5
  },
  {
   "url": "https://medium.com/@bumurzaqov2/top-10-ai-memory-products-2026-09d7900b5ab1",
   "quality": "blog",
   "angle": "Namespace crowding scan (recent market/practitioner)",
   "claimCount": 5
  },
  {
   "url": "https://github.com/TeleAI-UAGI/Awesome-Agent-Memory",
   "quality": "secondary",
   "angle": "Namespace crowding scan (recent market/practitioner)",
   "claimCount": 5
  },
  {
   "url": "https://tryxlr8.ai/blogs/best-open-source-ai-memory-frameworks-2026",
   "quality": "blog",
   "angle": "Namespace crowding scan (recent market/practitioner)",
   "claimCount": 5
  },
  {
   "url": "https://guptadeepak.com/the-ai-memory-wars-why-one-system-crushed-the-competition-and-its-not-openai/",
   "quality": "blog",
   "angle": "Namespace crowding scan (recent market/practitioner)",
   "claimCount": 5
  },
  {
   "url": "https://blog.devgenius.io/ai-agent-memory-systems-in-2026-mem0-zep-hindsight-memvid-and-everything-in-between-compared-96e35b818da8",
   "quality": "blog",
   "angle": "Namespace crowding scan (recent market/practitioner)",
   "claimCount": 5
  },
  {
   "url": "https://urbit.org/blog/creating-sigils",
   "quality": "primary",
   "angle": "Hash-to-glyph and machine-readable identity marks (technical/design)",
   "claimCount": 5
  },
  {
   "url": "https://github.com/urbit/sigil-js",
   "quality": "primary",
   "angle": "Hash-to-glyph and machine-readable identity marks (technical/design)",
   "claimCount": 5
  },
  {
   "url": "https://ieeexplore.ieee.org/document/9559993/",
   "quality": "primary",
   "angle": "Hash-to-glyph and machine-readable identity marks (technical/design)",
   "claimCount": 5
  },
  {
   "url": "https://barro.github.io/2018/02/avatars-identicons-and-hash-visualization/",
   "quality": "blog",
   "angle": "Hash-to-glyph and machine-readable identity marks (technical/design)",
   "claimCount": 5
  },
  {
   "url": "https://community.spotify.com/t5/Spotify-for-Developers/Spotify-Codes-How-do-they-work/td-p/4967330",
   "quality": "forum",
   "angle": "Hash-to-glyph and machine-readable identity marks (technical/design)",
   "claimCount": 5
  },
  {
   "url": "https://arxiv.org/pdf/2407.11748",
   "quality": "primary",
   "angle": "Hash-to-glyph and machine-readable identity marks (technical/design)",
   "claimCount": 5
  },
  {
   "url": "https://wipp.info/921382.pdf",
   "quality": "primary",
   "angle": "Deep-time semiotics and symbol durability (authoritative references)",
   "claimCount": 5
  },
  {
   "url": "https://en.wikipedia.org/wiki/Long-term_nuclear_waste_warning_messages",
   "quality": "secondary",
   "angle": "Deep-time semiotics and symbol durability (authoritative references)",
   "claimCount": 5
  },
  {
   "url": "https://www.sciencehistory.org/collections/blog/the-biohazard-warning-symbol/",
   "quality": "secondary",
   "angle": "Deep-time semiotics and symbol durability (authoritative references)",
   "claimCount": 4
  }
 ],
 "stats": {
  "angles": 5,
  "sourcesFetched": 25,
  "claimsExtracted": 122,
  "claimsVerified": 25,
  "confirmed": 25,
  "killed": 0,
  "afterSynthesis": 10,
  "urlDupes": 0,
  "budgetDropped": 5,
  "agentCalls": 107
 }
}