[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"portal-settings:stajic:it":3,"public-menus:all":38,"post:how-far-is-an-ai-claim-from-the-evidence-evidence-distance:it":221,"related:post:how-far-is-an-ai-claim-from-the-evidence-evidence-distance:it:1":1557},{"statusCode":4,"data":5,"message":37},200,{"tenantId":6,"lang":7,"defaultLang":8,"siteUrl":9,"contactEmail":10,"brandName":11,"logoUrl":12,"siteName":11,"siteDescription":13,"ogImage":10,"robotsIndex":14,"socialLinks":10,"reservedSlugs":10,"seoPolicy":15},"stajic","it","de","https:\u002F\u002Fstajic.de",null,"Stajic Platform","\u002FLogo_Planet.svg","Stajic Portal",true,{"branding":16,"relatedContent":17,"crossDomainLinks":18},{"logoUrl":12},{"enabled":14},[19,22,25,28,31,34],{"url":20,"label":21,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Ffigure.rocks","figure.rocks",{"url":23,"label":24,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Floving.rocks","loving.rocks",{"url":26,"label":27,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Fbazify.com","bazify.com",{"url":29,"label":30,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Fbazify.de","bazify.de",{"url":32,"label":33,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Fbazify.at","bazify.at",{"url":35,"label":36,"isActive":14,"showInFooter":14,"includeInSameAs":14},"https:\u002F\u002Fbazify.ba","bazify.ba","Portal settings resolved",[39,45],{"id":40,"name":41,"location":42,"isActive":14,"isDefault":43,"items":44},1,"main-navigation","header",false,[],{"id":46,"name":47,"location":48,"isActive":14,"isDefault":14,"items":49},4,"main-menu","sidebar",[50,66,79,93,103,118,133],{"id":51,"title":52,"url":60,"target":61,"icon":62,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":64,"portfolioId":10,"children":65},"item-18",{"de":53,"en":54,"es":55,"fr":56,"it":54,"ru":57,"sr":58,"zh":59},"Startseite","Home","Inicio","Accueil","Главная","Почетна","首页","\u002Ffull-stack-web-developer-munich-performance-seo-and-maintainable-builds","_self","i-lucide-home","page",111,[],{"id":67,"title":68,"url":75,"target":61,"icon":76,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":77,"portfolioId":10,"children":78},"item-22",{"de":69,"en":69,"es":70,"fr":69,"it":71,"ru":72,"sr":73,"zh":74},"Vision","Visión","Visione","Видение","Визија","想象","\u002Fueber-uns-webdesign-muenchen-webaplikation","i-lucide-eye",113,[],{"id":80,"title":81,"url":89,"target":61,"icon":90,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":91,"portfolioId":10,"children":92},"item-19",{"de":82,"en":83,"es":84,"fr":83,"it":85,"ru":86,"sr":87,"zh":88},"Leistungen","Services","Servicios","Servizi","Услуги","Услуге","服务","\u002Fservices-dienstleistungen-muenchen","i-lucide-wrench",116,[],{"id":94,"title":95,"url":99,"target":61,"icon":100,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":101,"portfolioId":10,"children":102},"item-23",{"de":96,"en":96,"es":96,"fr":96,"it":96,"ru":97,"sr":97,"zh":98},"Blog","Блог","博客","\u002Fblog","i-lucide-book-open",112,[],{"id":104,"title":105,"url":114,"target":61,"icon":115,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":116,"portfolioId":10,"children":117},"item-32",{"de":106,"en":107,"es":108,"fr":109,"it":110,"ru":111,"sr":112,"zh":113},"Neue Technologien","New Technologies","Nuevas tecnologías","Nouvelles technologies","Nuove tecnologie","Новые технологии","Нове технологије","新技术！","\u002Fneue-webtechnologien","i-lucide-sparkles",122,[],{"id":119,"title":120,"url":129,"target":61,"icon":130,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":131,"portfolioId":10,"children":132},"item-20",{"de":121,"en":122,"es":123,"fr":124,"it":125,"ru":126,"sr":127,"zh":128},"Kontakt","Contact us!","Contacto","Contact","Contatto","Контакт","Контактирајте нас","联系我们！","\u002Fcontact","i-lucide-mail",115,[],{"id":134,"title":135,"url":144,"target":61,"icon":145,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":146,"portfolioId":10,"children":147},"item-21",{"de":136,"en":137,"es":138,"fr":139,"it":140,"ru":141,"sr":142,"zh":143},"Unsere Arbeit","Our Work","Nuestro trabajo","Nos réalisations","I nostri lavori","Наши работы","Наши радови","文件夹","\u002Fportfolio","i-lucide-briefcase",114,[148,161,175,191,206],{"id":149,"title":150,"url":144,"target":61,"icon":159,"isActive":14,"type":63,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":146,"portfolioId":10,"children":160},"item-24",{"de":151,"en":152,"es":153,"fr":154,"it":155,"ru":156,"sr":157,"zh":158},"Alle Projekte","All Projects","Todos los proyectos","Tous les projets","Tutti i progetti","Все проекты","Сви пројекти","所有项目","i-lucide-grid-3x3",[],{"id":162,"title":163,"url":171,"target":61,"icon":172,"isActive":14,"type":173,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":10,"portfolioId":10,"children":174},"item-29",{"de":164,"en":165,"es":166,"fr":167,"it":168,"ru":169,"sr":170,"zh":143},"Local Roots, Global Reach","Local Roots - Global Reach","Empresa local ","Entreprise locale","Azienda locale","Местная компания","Локално предузеће глобално тржиште","\u002Fportfolio\u002Flocal-roots-global-reach-communication-media-systems-for-modern-business","i-lucide-globe-2","custom",[],{"id":176,"title":177,"url":186,"target":61,"icon":187,"isActive":14,"type":188,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":10,"portfolioId":189,"children":190},"menu-2916d060-8681-482a-af9a-c107a8bc30d8",{"de":178,"en":179,"es":180,"fr":181,"it":182,"ru":183,"sr":184,"zh":185},"Von Rohdaten zu verifizierbarem KI-Kontext: Eine praktische Architektur für agentenbereite Informationen","From Raw Data to Verifiable AI Context:","De datos sin procesar a contexto de IA verificable: una arquitectura práctica para información lista para agentes","Des données brutes au contexte d'IA vérifiable : une architecture pratique pour des informations prêtes pour les agents","Dai dati grezzi a un contesto IA verificabile: Un'architettura pratica per informazioni pronte per gli agenti","От необработанных данных к верифицируемому контексту ИИ: практическая архитектура информации, готовой для агентов","Od sirovih podataka do proverljivog AI konteksta: Praktična arhitektura za informacije spremne za agente","从原始数据到可验证的 AI 上下文：面向智能体就绪信息的实用架构","\u002Fportfolio\u002Ffrom-raw-data-to-verifiable-ai-context-a-practical-architecture-for-agent-ready-information","i-lucide-database","portfolio",22,[],{"id":192,"title":193,"url":202,"target":61,"icon":203,"isActive":14,"type":188,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":10,"portfolioId":204,"children":205},"menu-e2061cd7-ff53-49d5-a0e8-3d49090577c0",{"de":194,"en":195,"es":196,"fr":197,"it":198,"ru":199,"sr":200,"zh":201},"Präsenzplattformen — Unternehmenskommunikation","Presence Platforms — Enterprise Communication","Plataformas de presencia — Comunicación empresarial","Plateformes de présence — Communication d'entreprise","Presence Platforms — Comunicazione aziendale","Платформы присутствия — Корпоративные коммуникации","Presence platforme — Poslovna komunikacija","Presence 平台 — 企业通信","\u002Fportfolio\u002Fpresence-platforms-enterprise-communication-that-wins-without-competing","i-lucide-badge-percent",21,[],{"id":207,"title":208,"url":217,"target":61,"icon":218,"isActive":14,"type":188,"productId":10,"categoryId":10,"shopCategoryId":10,"articleId":10,"pageId":10,"portfolioId":219,"children":220},"menu-a781582e-e90d-418e-b8c9-b070bc37c37e",{"de":209,"en":210,"es":211,"fr":212,"it":213,"ru":214,"sr":215,"zh":216},"Maximierung des POS-ROI durch Konsistenz","Maximizing POS ROI Through Consistent","Maximizar el ROI del POS a través de constante","Maximiser le ROI du PDV grâce à la cohérence","Massimizzare il ROI del POS attraverso costante","Максимизация ROI POS за счет последовательного","Maksimizacija POS ROI-ja kroz dosledno","通过持续最大化 POS 投资回报率","\u002Fportfolio\u002Ffrom-vision-to-value-maximizing-pos-roi-through-consistent-ci-creation-and-strategic-branding","i-lucide-briefcase-business",11,[],{"statusCode":4,"data":222,"message":1556},{"id":223,"title":224,"slug":225,"content":226,"contentJson":227,"excerpt":742,"featuredImage":743,"featuredImageAlt":744,"featuredImageCaption":10,"featuredImageTitle":10,"featuredImageCopyright":10,"featuredImageAuthor":10,"featuredImageSourceUrl":10,"featuredImageLicense":10,"featuredImageIsAiGenerated":43,"status":745,"publishedAt":746,"createdAt":747,"updatedAt":748,"seoLocalePaths":749,"categories":758,"author":767,"translations":771},"497","Quanto dista un'affermazione dell'IA dalle prove? — Distanza probatoria","how-far-is-an-ai-claim-from-the-evidence-evidence-distance","{\"time\":1792096987644,\"blocks\":[{\"id\":\"_cNyEIGa5l\",\"type\":\"header\",\"data\":{\"text\":\"Domanda\",\"level\":2},\"tunes\":{}},{\"id\":\"zquDJabT2E\",\"type\":\"paragraph\",\"data\":{\"text\":\"Due risposte AI possono entrambe citare fonti e differire comunque notevolmente nel modo in cui tali fonti supportano direttamente le affermazioni fatte. Come possiamo descrivere questa differenza?\"},\"tunes\":{}},{\"id\":\"HCwt6O-cFu\",\"type\":\"paragraph\",\"data\":{\"text\":\"Questo articolo utilizza il termine Distanza Evidenziale per indicare la distanza tra un'affermazione e le prove direttamente ispezionabili che la supportano.\"},\"tunes\":{}},{\"id\":\"2li3h19tVw\",\"type\":\"quote\",\"data\":{\"text\":\"La Distanza Evidenziale è il numero e la forza delle trasformazioni, delle assunzioni e dei passaggi inferenziali che separano un'affermazione dalle prove che possono verificarla direttamente.\",\"caption\":\"Definizione di lavoro — un concetto di ragionamento introdotto qui, non una metrica standardizzata.\",\"alignment\":\"left\"},\"tunes\":{}},{\"id\":\"Ui3OETMemH\",\"type\":\"tableOfContents\",\"data\":{\"title\":\"Contents\",\"minLevel\":2,\"maxLevel\":3},\"tunes\":{}},{\"id\":\"cCf-n1UX1k\",\"type\":\"header\",\"data\":{\"text\":\"Cosa Significa Davvero\",\"level\":2},\"tunes\":{}},{\"id\":\"IioNnoi6Db\",\"type\":\"paragraph\",\"data\":{\"text\":\"Una citazione prova che una fonte è stata referenziata. Non prova automaticamente che la fonte supporti direttamente la frase accanto alla citazione.\"},\"tunes\":{}},{\"id\":\"TAD0q5su8E\",\"type\":\"paragraph\",\"data\":{\"text\":\"Tra un'osservazione originale e una risposta AI finale possono esserci estrazione, parsing, suddivisione in blocchi, recupero, sintesi, interpretazione, aggregazione e inferenza.\"},\"tunes\":{}},{\"id\":\"iaULb-gy-l\",\"type\":\"code\",\"data\":{\"code\":\"Primary evidence\\n   ↓\\nExtracted data\\n   ↓\\nRetrieved passage\\n   ↓\\nSummary\\n   ↓\\nInterpretation\\n   ↓\\nInference\\n   ↓\\nFinal claim\"},\"tunes\":{}},{\"id\":\"T5ecQtcSO9\",\"type\":\"paragraph\",\"data\":{\"text\":\"Ogni passaggio può essere legittimo. Ma ogni passaggio può anche introdurre distorsione.\"},\"tunes\":{}},{\"id\":\"BdAu5MTvAf\",\"type\":\"header\",\"data\":{\"text\":\"Esempio Più Semplice\",\"level\":2},\"tunes\":{}},{\"id\":\"K5Vc2vdh0A\",\"type\":\"paragraph\",\"data\":{\"text\":\"Supponiamo che un sensore di temperatura riporti 38,2 °C.\"},\"tunes\":{}},{\"id\":\"YhoIQsX4-A\",\"type\":\"table\",\"data\":{\"withHeadings\":true,\"stretched\":false,\"content\":[[\"Affermazione\",\"Relazione con le prove\",\"Distanza illustrativa\"],[\"Il sensore riporta 38,2 °C.\",\"Lettura diretta\",\"Bassa\"],[\"Il sistema sta funzionando più caldo di ieri.\",\"Richiede il confronto con la lettura di ieri\",\"Moderata\"],[\"Il sistema di raffreddamento sta cedendo.\",\"Richiede inferenza causale e spiegazioni alternative\",\"Alta\"]]},\"tunes\":{}},{\"id\":\"t7hUmXmjo_\",\"type\":\"paragraph\",\"data\":{\"text\":\"Il valore originale del sensore supporta direttamente la prima affermazione. Non prova da solo la terza.\"},\"tunes\":{}},{\"id\":\"k02EHsm2Kx\",\"type\":\"code\",\"data\":{\"code\":\"Evidence exists\\n        ≠\\nEvidence directly supports every derived claim\"},\"tunes\":{}},{\"id\":\"X2YHihB7oZ\",\"type\":\"header\",\"data\":{\"text\":\"Dove l'Esempio Smette di Funzionare\",\"level\":2},\"tunes\":{}},{\"id\":\"44i2ja30CF\",\"type\":\"paragraph\",\"data\":{\"text\":\"La Distanza Evidenziale non è semplicemente un conteggio di frasi o passaggi di elaborazione.\"},\"tunes\":{}},{\"id\":\"qIhekV7q-w\",\"type\":\"paragraph\",\"data\":{\"text\":\"Una trasformazione matematicamente valida può preservare le prove quasi perfettamente, mentre una singola interpretazione speculativa può creare un salto epistemico molto più grande.\"},\"tunes\":{}},{\"id\":\"Xaony1j2zz\",\"type\":\"paragraph\",\"data\":{\"text\":\"Il concetto include quindi sia la distanza per trasformazione sia la distanza per inferenza.\"},\"tunes\":{}},{\"id\":\"UntxDRQiTo\",\"type\":\"header\",\"data\":{\"text\":\"Risposta Diretta\",\"level\":2},\"tunes\":{}},{\"id\":\"cB3w5DZ_u-\",\"type\":\"paragraph\",\"data\":{\"text\":\"La Distanza Probatoria aumenta quando un'affermazione dipende da più trasformazioni, più assunzioni, una provenienza più debole o un'inferenza più forte di quanto le prove sottostanti giustifichino direttamente.\"},\"tunes\":{}},{\"id\":\"B84VY8EI4b\",\"type\":\"code\",\"data\":{\"code\":\"Short distance:\\nsource → claim\\n\\nLonger distance:\\nsource → extraction → summary → assumption → inference → claim\"},\"tunes\":{}},{\"id\":\"MBO4sSDo3J\",\"type\":\"paragraph\",\"data\":{\"text\":\"Una distanza maggiore non rende automaticamente falsa un'affermazione. Significa che la catena di ragionamento richiede più validazione.\"},\"tunes\":{}},{\"id\":\"UaL4TmBnM3\",\"type\":\"header\",\"data\":{\"text\":\"Perché Questo È Importante per l'IA\",\"level\":2},\"tunes\":{}},{\"id\":\"fJQ8ugBr5n\",\"type\":\"paragraph\",\"data\":{\"text\":\"I LLM sono particolarmente bravi a produrre ponti fluidi tra informazioni. Questa capacità è utile, ma la fluidità può nascondere quanto la conclusione si sia allontanata dalle prove recuperate direttamente.\"},\"tunes\":{}},{\"id\":\"4frSd0kYRY\",\"type\":\"paragraph\",\"data\":{\"text\":\"Una risposta può quindi sembrare ben documentata pur contenendo affermazioni il cui supporto esiste solo attraverso diversi passaggi inferenziali non dichiarati.\"},\"tunes\":{}},{\"id\":\"X-N5Fxpo22\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Fit\u002Fblog\u002Ffrom-research-protocol-to-a-general-ai-reasoning-framework\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Dal Protocollo di Ricerca a un Framework Generale di Ragionamento per l'IA\",\"description\":\"Un ragionamento affidabile richiede una separazione esplicita di prove, assunzioni, ipotesi e conclusioni.\"}},\"tunes\":{}},{\"id\":\"czFEqPzHoS\",\"type\":\"header\",\"data\":{\"text\":\"Quattro Componenti della Distanza Probatoria\",\"level\":2},\"tunes\":{}},{\"id\":\"wHMeS7mgEN\",\"type\":\"header\",\"data\":{\"text\":\"1. Distanza di Trasformazione\",\"level\":3},\"tunes\":{}},{\"id\":\"czUDGToEoN\",\"type\":\"paragraph\",\"data\":{\"text\":\"Quante volte la prova originale è stata convertita, riassunta, normalizzata o riformattata prima di raggiungere il modello?\"},\"tunes\":{}},{\"id\":\"fy7TOoCU5y\",\"type\":\"header\",\"data\":{\"text\":\"2. Distanza Inferenziale\",\"level\":3},\"tunes\":{}},{\"id\":\"rIiiusMjLw\",\"type\":\"paragraph\",\"data\":{\"text\":\"Quanti passaggi di ragionamento sono necessari prima che le prove supportino la conclusione finale?\"},\"tunes\":{}},{\"id\":\"nKDRIKbyQB\",\"type\":\"header\",\"data\":{\"text\":\"3. Distanza di Provenienza\",\"level\":3},\"tunes\":{}},{\"id\":\"2nuwZS-r8G\",\"type\":\"paragraph\",\"data\":{\"text\":\"L'affermazione recuperata può ancora essere ricondotta a una fonte originale identificabile, o è passata attraverso riassunti e copie la cui origine non è chiara?\"},\"tunes\":{}},{\"id\":\"mAfQFGuEqy\",\"type\":\"header\",\"data\":{\"text\":\"4. Distanza di Applicabilità\",\"level\":3},\"tunes\":{}},{\"id\":\"nqSYYV7Zsu\",\"type\":\"paragraph\",\"data\":{\"text\":\"Quanto ragionamento è necessario per stabilire che le prove si applicano a questa esatta versione, utente, giurisdizione, periodo di tempo o situazione?\"},\"tunes\":{}},{\"id\":\"80_RYqfecA\",\"type\":\"header\",\"data\":{\"text\":\"Distanza Probatoria nel RAG\",\"level\":2},\"tunes\":{}},{\"id\":\"5RmNXxCaUw\",\"type\":\"paragraph\",\"data\":{\"text\":\"Il RAG può accorciare una distanza importante: la distanza tra un modello e la conoscenza esterna.\"},\"tunes\":{}},{\"id\":\"5YdVpEa-00\",\"type\":\"paragraph\",\"data\":{\"text\":\"Non accorcia automaticamente la distanza tra un passaggio recuperato e l'affermazione finale.\"},\"tunes\":{}},{\"id\":\"2pbEUA-uR2\",\"type\":\"code\",\"data\":{\"code\":\"RAG solves:\\nmodel ↔ external information\\n\\nRAG does not automatically solve:\\nevidence ↔ justified conclusion\"},\"tunes\":{}},{\"id\":\"vXDTuE75ti\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Fit\u002Fblog\u002Fwhere-does-an-llm-get-its-data-rag-data-sources-in-python\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Da dove ottiene i dati un LLM? Fonti di dati RAG in Python\",\"description\":\"Il recupero porta informazioni esterne nel contesto. La Distanza Evidenziale chiede cosa accade tra quell'informazione e l'affermazione prodotta alla fine.\"}},\"tunes\":{}},{\"id\":\"pxNGnwR3XU\",\"type\":\"header\",\"data\":{\"text\":\"Connessione al Trigger di Recupero\",\"level\":2},\"tunes\":{}},{\"id\":\"LmjG0lNBIu\",\"type\":\"paragraph\",\"data\":{\"text\":\"Il Trigger di Recupero chiede se la conoscenza interna del modello è sufficiente.\"},\"tunes\":{}},{\"id\":\"jD59uy-GBs\",\"type\":\"paragraph\",\"data\":{\"text\":\"La Distanza Evidenziale pone una domanda successiva: dopo che l'evidenza è stata recuperata, quanto direttamente supporta ciò che il sistema vuole affermare?\"},\"tunes\":{}},{\"id\":\"gMCQRRhdpZ\",\"type\":\"code\",\"data\":{\"code\":\"Need evidence?\\n   ↓\\nRetrieval Trigger\\n   ↓\\nRetrieve\\n   ↓\\nHow directly does evidence support the claim?\\n   ↓\\nEvidence Distance\"},\"tunes\":{}},{\"id\":\"zhzxVjfEca\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Fit\u002Fblog\u002Fwhen-should-an-ai-stop-trusting-its-own-knowledge-the-retrieval-trigger\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Quando dovrebbe un'IA smettere di fidarsi della propria conoscenza? — Il Trigger di Recupero\",\"description\":\"Il trigger decide quando è necessaria l'evidenza. La Distanza Evidenziale valuta quanto l'affermazione finale si allontana da quell'evidenza.\"}},\"tunes\":{}},{\"id\":\"-4f4fOp3iX\",\"type\":\"header\",\"data\":{\"text\":\"Una Diagnostica Pratica\",\"level\":2},\"tunes\":{}},{\"id\":\"1rmtCcCh_v\",\"type\":\"paragraph\",\"data\":{\"text\":\"Per un'affermazione importante generata dall'IA, ricostruisci la catena di supporto a ritroso.\"},\"tunes\":{}},{\"id\":\"_WOIeKMLb8\",\"type\":\"code\",\"data\":{\"code\":\"CLAIM\\n  ↑\\nWhich inference produced it?\\n  ↑\\nWhich retrieved evidence supports that inference?\\n  ↑\\nWhich source produced that evidence?\\n  ↑\\nCan the original source still be inspected?\"},\"tunes\":{}},{\"id\":\"4fEwKjVfxX\",\"type\":\"paragraph\",\"data\":{\"text\":\"Se la catena si interrompe, il sistema ha incontrato un problema di tracciabilità dell'evidenza.\"},\"tunes\":{}},{\"id\":\"K-HPB8eht4\",\"type\":\"paragraph\",\"data\":{\"text\":\"Se la catena rimane intatta ma richiede diverse assunzioni forti, la Distanza Evidenziale è ancora alta.\"},\"tunes\":{}},{\"id\":\"Siwz_4m6wF\",\"type\":\"header\",\"data\":{\"text\":\"Esempio: Dati di Prodotto\",\"level\":2},\"tunes\":{}},{\"id\":\"sBeh4-4LY9\",\"type\":\"paragraph\",\"data\":{\"text\":\"Supponiamo che una specifica del produttore indichi che un laptop pesa 1,48 kg.\"},\"tunes\":{}},{\"id\":\"rfTxa3CcEY\",\"type\":\"code\",\"data\":{\"code\":\"Manufacturer specification\\n→ 1.48 kg\\n→ \\\"The laptop weighs 1.48 kg\\\"\\n\\nLow Evidence Distance\"},\"tunes\":{}},{\"id\":\"YrpqFryDiT\",\"type\":\"paragraph\",\"data\":{\"text\":\"Ora considera:\"},\"tunes\":{}},{\"id\":\"QgUQXwZeVh\",\"type\":\"code\",\"data\":{\"code\":\"Manufacturer specification\\n→ 1.48 kg\\n→ comparison with three competitors\\n→ assumption about travel behavior\\n→ \\\"This is the best laptop for frequent travelers\\\"\\n\\nMuch higher Evidence Distance\"},\"tunes\":{}},{\"id\":\"f5erFnvv7u\",\"type\":\"paragraph\",\"data\":{\"text\":\"La specifica è reale. La raccomandazione è ancora a diversi passaggi di ragionamento da essa.\"},\"tunes\":{}},{\"id\":\"zU-8RDRT-m\",\"type\":\"header\",\"data\":{\"text\":\"Esempio: Evidenza Scientifica\",\"level\":2},\"tunes\":{}},{\"id\":\"CyElFasGLh\",\"type\":\"paragraph\",\"data\":{\"text\":\"Un articolo può riportare un'associazione osservata. Una risposta AI può comprimerla erroneamente in un'affermazione causale.\"},\"tunes\":{}},{\"id\":\"zR8aIFhDTv\",\"type\":\"code\",\"data\":{\"code\":\"Observed association\\n        ↓\\nInterpretation\\n        ↓\\nUnstated causal assumption\\n        ↓\\n\\\"X causes Y\\\"\"},\"tunes\":{}},{\"id\":\"R89PJ-yjrn\",\"type\":\"paragraph\",\"data\":{\"text\":\"Il problema non è che non esista alcuna evidenza. Il problema è che l'affermazione finale supera una distanza probatoria maggiore di quella che lo studio supporta direttamente.\"},\"tunes\":{}},{\"id\":\"DC4SGTPwIC\",\"type\":\"header\",\"data\":{\"text\":\"Connessione al Contesto Agent-Ready\",\"level\":2},\"tunes\":{}},{\"id\":\"eJHsu4Aepa\",\"type\":\"paragraph\",\"data\":{\"text\":\"Gli oggetti informativi strutturati possono ridurre la Distanza Probatoria evitabile preservando l'identità della fonte, le unità, i timestamp, i conflitti e la validità invece di costringere il modello a ricostruirli da testo appiattito.\"},\"tunes\":{}},{\"id\":\"K2PwbTNL5n\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Fit\u002Fportfolio\u002Ffrom-raw-data-to-verifiable-ai-context-a-practical-architecture-for-agent-ready-information\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Dai Dati Grezzi al Contesto AI Verificabile: Un'Architettura Pratica per Informazioni Agent-Ready\",\"description\":\"Preservare la provenienza e la validità prima del recupero riduce il numero di passaggi di ricostruzione non supportati richiesti successivamente.\"}},\"tunes\":{}},{\"id\":\"VNq8gm-0iP\",\"type\":\"header\",\"data\":{\"text\":\"Una Semplice Scala di Lavoro\",\"level\":2},\"tunes\":{}},{\"id\":\"IWsoeOq2Og\",\"type\":\"table\",\"data\":{\"withHeadings\":true,\"stretched\":false,\"content\":[[\"Livello\",\"Descrizione\"],[\"ED0\",\"Fatto di origine riportato direttamente\"],[\"ED1\",\"Trasformazione deterministica semplice o confronto diretto\"],[\"ED2\",\"Un passaggio interpretativo significativo\"],[\"ED3\",\"Assunzioni multiple o fonti combinate\"],[\"ED4\",\"Conclusione causale, predittiva o prescrittiva forte oltre l'evidenza diretta\"]]},\"tunes\":{}},{\"id\":\"oYa70Zw8eX\",\"type\":\"paragraph\",\"data\":{\"text\":\"Questa scala è illustrativa, non uno standard scientifico. Il suo scopo è forzare una domanda esplicita: Quanto abbiamo dovuto aggiungere tra l'evidenza e l'affermazione?\"},\"tunes\":{}},{\"id\":\"ch_jmkx_SI\",\"type\":\"header\",\"data\":{\"text\":\"Idee Sbagliate\",\"level\":2},\"tunes\":{}},{\"id\":\"c5_cx6P1yw\",\"type\":\"paragraph\",\"data\":{\"text\":\"Una citazione significa che la Distanza Probatoria è bassa. No. Una citazione può trovarsi accanto a un'affermazione che supera sostanzialmente ciò che la fonte citata stabilisce.\"},\"tunes\":{}},{\"id\":\"pzC3e8mIxg\",\"type\":\"paragraph\",\"data\":{\"text\":\"Una lunga Distanza Probatoria significa che l'affermazione è sbagliata. No. Le conclusioni complesse possono essere ben supportate, ma richiedono una catena di ragionamento più forte.\"},\"tunes\":{}},{\"id\":\"JSzOTaAljE\",\"type\":\"paragraph\",\"data\":{\"text\":\"Le fonti primarie eliminano l'inferenza. No. L'evidenza primaria può comunque richiedere interpretazione.\"},\"tunes\":{}},{\"id\":\"PusOuHZw1W\",\"type\":\"paragraph\",\"data\":{\"text\":\"Più fonti riducono automaticamente la distanza. Dieci deboli ripetizioni della stessa affermazione non supportata non creano evidenza diretta.\"},\"tunes\":{}},{\"id\":\"MCQIyTAM1K\",\"type\":\"header\",\"data\":{\"text\":\"Caso Limite: Dati Derivati\",\"level\":2},\"tunes\":{}},{\"id\":\"pIXtSb8WEu\",\"type\":\"paragraph\",\"data\":{\"text\":\"Alcuni valori derivati sono più affidabili delle descrizioni umane grezze. Un totale calcolato correttamente da transazioni verificate può essere altamente affidabile anche se è a una trasformazione di distanza dai record originali.\"},\"tunes\":{}},{\"id\":\"uh8rC1EY0V\",\"type\":\"paragraph\",\"data\":{\"text\":\"La Distanza Evidenziale misura quindi la separazione epistemica, non una preferenza semplicistica per i dati non toccati.\"},\"tunes\":{}},{\"id\":\"bGA9XU5QUE\",\"type\":\"header\",\"data\":{\"text\":\"Limitazioni\",\"level\":2},\"tunes\":{}},{\"id\":\"WFnLr5U-06\",\"type\":\"paragraph\",\"data\":{\"text\":\"La Distanza Evidenziale è una diagnostica concettuale, non un punteggio quantitativo universale.\"},\"tunes\":{}},{\"id\":\"9wGK3LDrxR\",\"type\":\"paragraph\",\"data\":{\"text\":\"Domini diversi richiedono trattamenti diversi dell'inferenza. La prova matematica, l'evidenza clinica, i log software, l'analisi di mercato e l'interpretazione storica non possono condividere un'unica soglia numerica universale.\"},\"tunes\":{}},{\"id\":\"lyVGur2SJF\",\"type\":\"header\",\"data\":{\"text\":\"Cosa Cambierebbe Questa Risposta?\",\"level\":2},\"tunes\":{}},{\"id\":\"vHmJjvO8Ik\",\"type\":\"paragraph\",\"data\":{\"text\":\"I futuri sistemi di IA potrebbero generare grafi di provenienza espliciti, collegamenti verificabili automaticamente tra affermazioni ed evidenze e tracce di ragionamento tipizzate. Ciò potrebbe rendere parti della Distanza Evidenziale misurabili automaticamente.\"},\"tunes\":{}},{\"id\":\"HLig7cFyQF\",\"type\":\"paragraph\",\"data\":{\"text\":\"Il problema di fondo rimane: un'affermazione può allontanarsi dalla sua evidenza anche mentre ogni frase intermedia suona plausibile.\"},\"tunes\":{}},{\"id\":\"5_-nr7qHi7\",\"type\":\"header\",\"data\":{\"text\":\"Conclusione\",\"level\":2},\"tunes\":{}},{\"id\":\"DSGzbgZSQi\",\"type\":\"paragraph\",\"data\":{\"text\":\"La distinzione importante non è semplicemente tra risposte con fonti e risposte senza fonti.\"},\"tunes\":{}},{\"id\":\"zNIww6QXWv\",\"type\":\"paragraph\",\"data\":{\"text\":\"È tra affermazioni che rimangono vicine a evidenze direttamente ispezionabili e affermazioni che dipendono da catene sempre più lunghe di interpretazione e assunzione.\"},\"tunes\":{}},{\"id\":\"tCW-uNAZUK\",\"type\":\"code\",\"data\":{\"code\":\"Evidence\\n   ↓\\nEvidence Distance\\n   ↓\\nReasoning\\n   ↓\\nAnswer Validity Boundary\"},\"tunes\":{}},{\"id\":\"rrTbiZZaQ6\",\"type\":\"paragraph\",\"data\":{\"text\":\"La Distanza Evidenziale ci fornisce un linguaggio per descrivere quello spazio.\"},\"tunes\":{}},{\"id\":\"y71qRwcJzj\",\"type\":\"header\",\"data\":{\"text\":\"Fonti Primarie\",\"level\":2},\"tunes\":{}},{\"id\":\"MieF9xLLyp\",\"type\":\"paragraph\",\"data\":{\"text\":\"Patrick Lewis et al., \u003Ca href=\\\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2005.11401\\\" target=\\\"_blank\\\">Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks\u003C\u002Fa> (2020).\"},\"tunes\":{}},{\"id\":\"3I9CCkTBOT\",\"type\":\"paragraph\",\"data\":{\"text\":\"Nelson F. Liu et al., \u003Ca href=\\\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2307.03172\\\" target=\\\"_blank\\\">Lost in the Middle: How Language Models Use Long Contexts\u003C\u002Fa> (2023).\"},\"tunes\":{}},{\"id\":\"Kui3GHn9us\",\"type\":\"paragraph\",\"data\":{\"text\":\"W3C, \u003Ca href=\\\"https:\u002F\u002Fwww.w3.org\u002FTR\u002Fprov-overview\u002F\\\" target=\\\"_blank\\\">PROV Overview\u003C\u002Fa>. Un modello standardizzato per rappresentare le relazioni di provenienza.\"},\"tunes\":{}}],\"version\":\"2.31.6\"}",{"time":228,"blocks":229,"version":741},1792096987644,[230,236,242,247,255,262,267,272,277,283,288,293,298,320,325,330,335,340,345,350,355,360,365,370,375,380,385,396,401,406,411,416,421,426,431,436,441,446,451,456,461,470,475,480,485,490,499,504,509,514,519,524,529,534,539,544,549,554,559,564,569,574,579,584,593,598,621,626,631,636,641,646,651,656,661,666,671,676,681,686,691,696,701,706,711,716,721,726,731,736],{"id":231,"data":232,"type":42,"tunes":235},"_cNyEIGa5l",{"text":233,"level":234},"Domanda",2,{},{"id":237,"data":238,"type":240,"tunes":241},"zquDJabT2E",{"text":239},"Due risposte AI possono entrambe citare fonti e differire comunque notevolmente nel modo in cui tali fonti supportano direttamente le affermazioni fatte. Come possiamo descrivere questa differenza?","paragraph",{},{"id":243,"data":244,"type":240,"tunes":246},"HCwt6O-cFu",{"text":245},"Questo articolo utilizza il termine Distanza Evidenziale per indicare la distanza tra un'affermazione e le prove direttamente ispezionabili che la supportano.",{},{"id":248,"data":249,"type":253,"tunes":254},"2li3h19tVw",{"text":250,"caption":251,"alignment":252},"La Distanza Evidenziale è il numero e la forza delle trasformazioni, delle assunzioni e dei passaggi inferenziali che separano un'affermazione dalle prove che possono verificarla direttamente.","Definizione di lavoro — un concetto di ragionamento introdotto qui, non una metrica standardizzata.","left","quote",{},{"id":256,"data":257,"type":260,"tunes":261},"Ui3OETMemH",{"title":258,"maxLevel":259,"minLevel":234},"Contents",3,"tableOfContents",{},{"id":263,"data":264,"type":42,"tunes":266},"cCf-n1UX1k",{"text":265,"level":234},"Cosa Significa Davvero",{},{"id":268,"data":269,"type":240,"tunes":271},"IioNnoi6Db",{"text":270},"Una citazione prova che una fonte è stata referenziata. Non prova automaticamente che la fonte supporti direttamente la frase accanto alla citazione.",{},{"id":273,"data":274,"type":240,"tunes":276},"TAD0q5su8E",{"text":275},"Tra un'osservazione originale e una risposta AI finale possono esserci estrazione, parsing, suddivisione in blocchi, recupero, sintesi, interpretazione, aggregazione e inferenza.",{},{"id":278,"data":279,"type":281,"tunes":282},"iaULb-gy-l",{"code":280},"Primary evidence\n   ↓\nExtracted data\n   ↓\nRetrieved passage\n   ↓\nSummary\n   ↓\nInterpretation\n   ↓\nInference\n   ↓\nFinal claim","code",{},{"id":284,"data":285,"type":240,"tunes":287},"T5ecQtcSO9",{"text":286},"Ogni passaggio può essere legittimo. Ma ogni passaggio può anche introdurre distorsione.",{},{"id":289,"data":290,"type":42,"tunes":292},"BdAu5MTvAf",{"text":291,"level":234},"Esempio Più Semplice",{},{"id":294,"data":295,"type":240,"tunes":297},"K5Vc2vdh0A",{"text":296},"Supponiamo che un sensore di temperatura riporti 38,2 °C.",{},{"id":299,"data":300,"type":318,"tunes":319},"YhoIQsX4-A",{"content":301,"stretched":43,"withHeadings":14},[302,306,310,314],[303,304,305],"Affermazione","Relazione con le prove","Distanza illustrativa",[307,308,309],"Il sensore riporta 38,2 °C.","Lettura diretta","Bassa",[311,312,313],"Il sistema sta funzionando più caldo di ieri.","Richiede il confronto con la lettura di ieri","Moderata",[315,316,317],"Il sistema di raffreddamento sta cedendo.","Richiede inferenza causale e spiegazioni alternative","Alta","table",{},{"id":321,"data":322,"type":240,"tunes":324},"t7hUmXmjo_",{"text":323},"Il valore originale del sensore supporta direttamente la prima affermazione. Non prova da solo la terza.",{},{"id":326,"data":327,"type":281,"tunes":329},"k02EHsm2Kx",{"code":328},"Evidence exists\n        ≠\nEvidence directly supports every derived claim",{},{"id":331,"data":332,"type":42,"tunes":334},"X2YHihB7oZ",{"text":333,"level":234},"Dove l'Esempio Smette di Funzionare",{},{"id":336,"data":337,"type":240,"tunes":339},"44i2ja30CF",{"text":338},"La Distanza Evidenziale non è semplicemente un conteggio di frasi o passaggi di elaborazione.",{},{"id":341,"data":342,"type":240,"tunes":344},"qIhekV7q-w",{"text":343},"Una trasformazione matematicamente valida può preservare le prove quasi perfettamente, mentre una singola interpretazione speculativa può creare un salto epistemico molto più grande.",{},{"id":346,"data":347,"type":240,"tunes":349},"Xaony1j2zz",{"text":348},"Il concetto include quindi sia la distanza per trasformazione sia la distanza per inferenza.",{},{"id":351,"data":352,"type":42,"tunes":354},"UntxDRQiTo",{"text":353,"level":234},"Risposta Diretta",{},{"id":356,"data":357,"type":240,"tunes":359},"cB3w5DZ_u-",{"text":358},"La Distanza Probatoria aumenta quando un'affermazione dipende da più trasformazioni, più assunzioni, una provenienza più debole o un'inferenza più forte di quanto le prove sottostanti giustifichino direttamente.",{},{"id":361,"data":362,"type":281,"tunes":364},"B84VY8EI4b",{"code":363},"Short distance:\nsource → claim\n\nLonger distance:\nsource → extraction → summary → assumption → inference → claim",{},{"id":366,"data":367,"type":240,"tunes":369},"MBO4sSDo3J",{"text":368},"Una distanza maggiore non rende automaticamente falsa un'affermazione. Significa che la catena di ragionamento richiede più validazione.",{},{"id":371,"data":372,"type":42,"tunes":374},"UaL4TmBnM3",{"text":373,"level":234},"Perché Questo È Importante per l'IA",{},{"id":376,"data":377,"type":240,"tunes":379},"fJQ8ugBr5n",{"text":378},"I LLM sono particolarmente bravi a produrre ponti fluidi tra informazioni. Questa capacità è utile, ma la fluidità può nascondere quanto la conclusione si sia allontanata dalle prove recuperate direttamente.",{},{"id":381,"data":382,"type":240,"tunes":384},"4frSd0kYRY",{"text":383},"Una risposta può quindi sembrare ben documentata pur contenendo affermazioni il cui supporto esiste solo attraverso diversi passaggi inferenziali non dichiarati.",{},{"id":386,"data":387,"type":394,"tunes":395},"X-N5Fxpo22",{"link":388,"meta":389},"https:\u002F\u002Fstajic.de\u002Fit\u002Fblog\u002Ffrom-research-protocol-to-a-general-ai-reasoning-framework",{"image":390,"title":392,"description":393},{"url":391},"","Dal Protocollo di Ricerca a un Framework Generale di Ragionamento per l'IA","Un ragionamento affidabile richiede una separazione esplicita di prove, assunzioni, ipotesi e conclusioni.","linkTool",{},{"id":397,"data":398,"type":42,"tunes":400},"czFEqPzHoS",{"text":399,"level":234},"Quattro Componenti della Distanza Probatoria",{},{"id":402,"data":403,"type":42,"tunes":405},"wHMeS7mgEN",{"text":404,"level":259},"1. Distanza di Trasformazione",{},{"id":407,"data":408,"type":240,"tunes":410},"czUDGToEoN",{"text":409},"Quante volte la prova originale è stata convertita, riassunta, normalizzata o riformattata prima di raggiungere il modello?",{},{"id":412,"data":413,"type":42,"tunes":415},"fy7TOoCU5y",{"text":414,"level":259},"2. Distanza Inferenziale",{},{"id":417,"data":418,"type":240,"tunes":420},"rIiiusMjLw",{"text":419},"Quanti passaggi di ragionamento sono necessari prima che le prove supportino la conclusione finale?",{},{"id":422,"data":423,"type":42,"tunes":425},"nKDRIKbyQB",{"text":424,"level":259},"3. Distanza di Provenienza",{},{"id":427,"data":428,"type":240,"tunes":430},"2nuwZS-r8G",{"text":429},"L'affermazione recuperata può ancora essere ricondotta a una fonte originale identificabile, o è passata attraverso riassunti e copie la cui origine non è chiara?",{},{"id":432,"data":433,"type":42,"tunes":435},"mAfQFGuEqy",{"text":434,"level":259},"4. Distanza di Applicabilità",{},{"id":437,"data":438,"type":240,"tunes":440},"nqSYYV7Zsu",{"text":439},"Quanto ragionamento è necessario per stabilire che le prove si applicano a questa esatta versione, utente, giurisdizione, periodo di tempo o situazione?",{},{"id":442,"data":443,"type":42,"tunes":445},"80_RYqfecA",{"text":444,"level":234},"Distanza Probatoria nel RAG",{},{"id":447,"data":448,"type":240,"tunes":450},"5RmNXxCaUw",{"text":449},"Il RAG può accorciare una distanza importante: la distanza tra un modello e la conoscenza esterna.",{},{"id":452,"data":453,"type":240,"tunes":455},"5YdVpEa-00",{"text":454},"Non accorcia automaticamente la distanza tra un passaggio recuperato e l'affermazione finale.",{},{"id":457,"data":458,"type":281,"tunes":460},"2pbEUA-uR2",{"code":459},"RAG solves:\nmodel ↔ external information\n\nRAG does not automatically solve:\nevidence ↔ justified conclusion",{},{"id":462,"data":463,"type":394,"tunes":469},"vXDTuE75ti",{"link":464,"meta":465},"https:\u002F\u002Fstajic.de\u002Fit\u002Fblog\u002Fwhere-does-an-llm-get-its-data-rag-data-sources-in-python",{"image":466,"title":467,"description":468},{"url":391},"Da dove ottiene i dati un LLM? Fonti di dati RAG in Python","Il recupero porta informazioni esterne nel contesto. La Distanza Evidenziale chiede cosa accade tra quell'informazione e l'affermazione prodotta alla fine.",{},{"id":471,"data":472,"type":42,"tunes":474},"pxNGnwR3XU",{"text":473,"level":234},"Connessione al Trigger di Recupero",{},{"id":476,"data":477,"type":240,"tunes":479},"LmjG0lNBIu",{"text":478},"Il Trigger di Recupero chiede se la conoscenza interna del modello è sufficiente.",{},{"id":481,"data":482,"type":240,"tunes":484},"jD59uy-GBs",{"text":483},"La Distanza Evidenziale pone una domanda successiva: dopo che l'evidenza è stata recuperata, quanto direttamente supporta ciò che il sistema vuole affermare?",{},{"id":486,"data":487,"type":281,"tunes":489},"gMCQRRhdpZ",{"code":488},"Need evidence?\n   ↓\nRetrieval Trigger\n   ↓\nRetrieve\n   ↓\nHow directly does evidence support the claim?\n   ↓\nEvidence Distance",{},{"id":491,"data":492,"type":394,"tunes":498},"zhzxVjfEca",{"link":493,"meta":494},"https:\u002F\u002Fstajic.de\u002Fit\u002Fblog\u002Fwhen-should-an-ai-stop-trusting-its-own-knowledge-the-retrieval-trigger",{"image":495,"title":496,"description":497},{"url":391},"Quando dovrebbe un'IA smettere di fidarsi della propria conoscenza? — Il Trigger di Recupero","Il trigger decide quando è necessaria l'evidenza. La Distanza Evidenziale valuta quanto l'affermazione finale si allontana da quell'evidenza.",{},{"id":500,"data":501,"type":42,"tunes":503},"-4f4fOp3iX",{"text":502,"level":234},"Una Diagnostica Pratica",{},{"id":505,"data":506,"type":240,"tunes":508},"1rmtCcCh_v",{"text":507},"Per un'affermazione importante generata dall'IA, ricostruisci la catena di supporto a ritroso.",{},{"id":510,"data":511,"type":281,"tunes":513},"_WOIeKMLb8",{"code":512},"CLAIM\n  ↑\nWhich inference produced it?\n  ↑\nWhich retrieved evidence supports that inference?\n  ↑\nWhich source produced that evidence?\n  ↑\nCan the original source still be inspected?",{},{"id":515,"data":516,"type":240,"tunes":518},"4fEwKjVfxX",{"text":517},"Se la catena si interrompe, il sistema ha incontrato un problema di tracciabilità dell'evidenza.",{},{"id":520,"data":521,"type":240,"tunes":523},"K-HPB8eht4",{"text":522},"Se la catena rimane intatta ma richiede diverse assunzioni forti, la Distanza Evidenziale è ancora alta.",{},{"id":525,"data":526,"type":42,"tunes":528},"Siwz_4m6wF",{"text":527,"level":234},"Esempio: Dati di Prodotto",{},{"id":530,"data":531,"type":240,"tunes":533},"sBeh4-4LY9",{"text":532},"Supponiamo che una specifica del produttore indichi che un laptop pesa 1,48 kg.",{},{"id":535,"data":536,"type":281,"tunes":538},"rfTxa3CcEY",{"code":537},"Manufacturer specification\n→ 1.48 kg\n→ \"The laptop weighs 1.48 kg\"\n\nLow Evidence Distance",{},{"id":540,"data":541,"type":240,"tunes":543},"YrpqFryDiT",{"text":542},"Ora considera:",{},{"id":545,"data":546,"type":281,"tunes":548},"QgUQXwZeVh",{"code":547},"Manufacturer specification\n→ 1.48 kg\n→ comparison with three competitors\n→ assumption about travel behavior\n→ \"This is the best laptop for frequent travelers\"\n\nMuch higher Evidence Distance",{},{"id":550,"data":551,"type":240,"tunes":553},"f5erFnvv7u",{"text":552},"La specifica è reale. La raccomandazione è ancora a diversi passaggi di ragionamento da essa.",{},{"id":555,"data":556,"type":42,"tunes":558},"zU-8RDRT-m",{"text":557,"level":234},"Esempio: Evidenza Scientifica",{},{"id":560,"data":561,"type":240,"tunes":563},"CyElFasGLh",{"text":562},"Un articolo può riportare un'associazione osservata. Una risposta AI può comprimerla erroneamente in un'affermazione causale.",{},{"id":565,"data":566,"type":281,"tunes":568},"zR8aIFhDTv",{"code":567},"Observed association\n        ↓\nInterpretation\n        ↓\nUnstated causal assumption\n        ↓\n\"X causes Y\"",{},{"id":570,"data":571,"type":240,"tunes":573},"R89PJ-yjrn",{"text":572},"Il problema non è che non esista alcuna evidenza. Il problema è che l'affermazione finale supera una distanza probatoria maggiore di quella che lo studio supporta direttamente.",{},{"id":575,"data":576,"type":42,"tunes":578},"DC4SGTPwIC",{"text":577,"level":234},"Connessione al Contesto Agent-Ready",{},{"id":580,"data":581,"type":240,"tunes":583},"eJHsu4Aepa",{"text":582},"Gli oggetti informativi strutturati possono ridurre la Distanza Probatoria evitabile preservando l'identità della fonte, le unità, i timestamp, i conflitti e la validità invece di costringere il modello a ricostruirli da testo appiattito.",{},{"id":585,"data":586,"type":394,"tunes":592},"K2PwbTNL5n",{"link":587,"meta":588},"https:\u002F\u002Fstajic.de\u002Fit\u002Fportfolio\u002Ffrom-raw-data-to-verifiable-ai-context-a-practical-architecture-for-agent-ready-information",{"image":589,"title":590,"description":591},{"url":391},"Dai Dati Grezzi al Contesto AI Verificabile: Un'Architettura Pratica per Informazioni Agent-Ready","Preservare la provenienza e la validità prima del recupero riduce il numero di passaggi di ricostruzione non supportati richiesti successivamente.",{},{"id":594,"data":595,"type":42,"tunes":597},"VNq8gm-0iP",{"text":596,"level":234},"Una Semplice Scala di Lavoro",{},{"id":599,"data":600,"type":318,"tunes":620},"IWsoeOq2Og",{"content":601,"stretched":43,"withHeadings":14},[602,605,608,611,614,617],[603,604],"Livello","Descrizione",[606,607],"ED0","Fatto di origine riportato direttamente",[609,610],"ED1","Trasformazione deterministica semplice o confronto diretto",[612,613],"ED2","Un passaggio interpretativo significativo",[615,616],"ED3","Assunzioni multiple o fonti combinate",[618,619],"ED4","Conclusione causale, predittiva o prescrittiva forte oltre l'evidenza diretta",{},{"id":622,"data":623,"type":240,"tunes":625},"oYa70Zw8eX",{"text":624},"Questa scala è illustrativa, non uno standard scientifico. Il suo scopo è forzare una domanda esplicita: Quanto abbiamo dovuto aggiungere tra l'evidenza e l'affermazione?",{},{"id":627,"data":628,"type":42,"tunes":630},"ch_jmkx_SI",{"text":629,"level":234},"Idee Sbagliate",{},{"id":632,"data":633,"type":240,"tunes":635},"c5_cx6P1yw",{"text":634},"Una citazione significa che la Distanza Probatoria è bassa. No. Una citazione può trovarsi accanto a un'affermazione che supera sostanzialmente ciò che la fonte citata stabilisce.",{},{"id":637,"data":638,"type":240,"tunes":640},"pzC3e8mIxg",{"text":639},"Una lunga Distanza Probatoria significa che l'affermazione è sbagliata. No. Le conclusioni complesse possono essere ben supportate, ma richiedono una catena di ragionamento più forte.",{},{"id":642,"data":643,"type":240,"tunes":645},"JSzOTaAljE",{"text":644},"Le fonti primarie eliminano l'inferenza. No. L'evidenza primaria può comunque richiedere interpretazione.",{},{"id":647,"data":648,"type":240,"tunes":650},"PusOuHZw1W",{"text":649},"Più fonti riducono automaticamente la distanza. Dieci deboli ripetizioni della stessa affermazione non supportata non creano evidenza diretta.",{},{"id":652,"data":653,"type":42,"tunes":655},"MCQIyTAM1K",{"text":654,"level":234},"Caso Limite: Dati Derivati",{},{"id":657,"data":658,"type":240,"tunes":660},"pIXtSb8WEu",{"text":659},"Alcuni valori derivati sono più affidabili delle descrizioni umane grezze. Un totale calcolato correttamente da transazioni verificate può essere altamente affidabile anche se è a una trasformazione di distanza dai record originali.",{},{"id":662,"data":663,"type":240,"tunes":665},"uh8rC1EY0V",{"text":664},"La Distanza Evidenziale misura quindi la separazione epistemica, non una preferenza semplicistica per i dati non toccati.",{},{"id":667,"data":668,"type":42,"tunes":670},"bGA9XU5QUE",{"text":669,"level":234},"Limitazioni",{},{"id":672,"data":673,"type":240,"tunes":675},"WFnLr5U-06",{"text":674},"La Distanza Evidenziale è una diagnostica concettuale, non un punteggio quantitativo universale.",{},{"id":677,"data":678,"type":240,"tunes":680},"9wGK3LDrxR",{"text":679},"Domini diversi richiedono trattamenti diversi dell'inferenza. La prova matematica, l'evidenza clinica, i log software, l'analisi di mercato e l'interpretazione storica non possono condividere un'unica soglia numerica universale.",{},{"id":682,"data":683,"type":42,"tunes":685},"lyVGur2SJF",{"text":684,"level":234},"Cosa Cambierebbe Questa Risposta?",{},{"id":687,"data":688,"type":240,"tunes":690},"vHmJjvO8Ik",{"text":689},"I futuri sistemi di IA potrebbero generare grafi di provenienza espliciti, collegamenti verificabili automaticamente tra affermazioni ed evidenze e tracce di ragionamento tipizzate. Ciò potrebbe rendere parti della Distanza Evidenziale misurabili automaticamente.",{},{"id":692,"data":693,"type":240,"tunes":695},"HLig7cFyQF",{"text":694},"Il problema di fondo rimane: un'affermazione può allontanarsi dalla sua evidenza anche mentre ogni frase intermedia suona plausibile.",{},{"id":697,"data":698,"type":42,"tunes":700},"5_-nr7qHi7",{"text":699,"level":234},"Conclusione",{},{"id":702,"data":703,"type":240,"tunes":705},"DSGzbgZSQi",{"text":704},"La distinzione importante non è semplicemente tra risposte con fonti e risposte senza fonti.",{},{"id":707,"data":708,"type":240,"tunes":710},"zNIww6QXWv",{"text":709},"È tra affermazioni che rimangono vicine a evidenze direttamente ispezionabili e affermazioni che dipendono da catene sempre più lunghe di interpretazione e assunzione.",{},{"id":712,"data":713,"type":281,"tunes":715},"tCW-uNAZUK",{"code":714},"Evidence\n   ↓\nEvidence Distance\n   ↓\nReasoning\n   ↓\nAnswer Validity Boundary",{},{"id":717,"data":718,"type":240,"tunes":720},"rrTbiZZaQ6",{"text":719},"La Distanza Evidenziale ci fornisce un linguaggio per descrivere quello spazio.",{},{"id":722,"data":723,"type":42,"tunes":725},"y71qRwcJzj",{"text":724,"level":234},"Fonti Primarie",{},{"id":727,"data":728,"type":240,"tunes":730},"MieF9xLLyp",{"text":729},"Patrick Lewis et al., \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2005.11401\" target=\"_blank\">Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks\u003C\u002Fa> (2020).",{},{"id":732,"data":733,"type":240,"tunes":735},"3I9CCkTBOT",{"text":734},"Nelson F. Liu et al., \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2307.03172\" target=\"_blank\">Lost in the Middle: How Language Models Use Long Contexts\u003C\u002Fa> (2023).",{},{"id":737,"data":738,"type":240,"tunes":740},"Kui3GHn9us",{"text":739},"W3C, \u003Ca href=\"https:\u002F\u002Fwww.w3.org\u002FTR\u002Fprov-overview\u002F\" target=\"_blank\">PROV Overview\u003C\u002Fa>. Un modello standardizzato per rappresentare le relazioni di provenienza.",{},"2.31.6","Non tutte le risposte supportate sono ugualmente vicine alle loro prove. La Distanza dalle Prove è un modo pratico per descrivere quante trasformazioni, sintesi, assunzioni e passaggi inferenziali separano un'affermazione dell'IA da prove direttamente ispezionabili.","\u002Fuploads\u002F2026\u002F10\u002Fhow-far-is-an-ai-claim-from-the-evidence-evidence-distance-1792095863209-3c1mog.webp","how-far-is-an-ai-claim-from-the-evidence-evidence-distance-1792095863209-3c1mog","PUBLISHED","2026-10-15T00:23:00.000Z","2026-10-15T20:23:08.634Z","2026-10-15T20:43:52.893Z",{"en":750,"de":751,"sr":752,"es":753,"fr":754,"it":755,"ru":756,"zh":757},"\u002Fblog\u002Fhow-far-is-an-ai-claim-from-the-evidence-evidence-distance","\u002Fde\u002Fblog\u002Fhow-far-is-an-ai-claim-from-the-evidence-evidence-distance","\u002Fsr\u002Fblog\u002Fhow-far-is-an-ai-claim-from-the-evidence-evidence-distance","\u002Fes\u002Fblog\u002Fhow-far-is-an-ai-claim-from-the-evidence-evidence-distance","\u002Ffr\u002Fblog\u002Fhow-far-is-an-ai-claim-from-the-evidence-evidence-distance","\u002Fit\u002Fblog\u002Fhow-far-is-an-ai-claim-from-the-evidence-evidence-distance","\u002Fru\u002Fblog\u002Fhow-far-is-an-ai-claim-from-the-evidence-evidence-distance","\u002Fzh\u002Fblog\u002Fhow-far-is-an-ai-claim-from-the-evidence-evidence-distance",[759,763],{"id":760,"name":761,"slug":762},68,"Rischi, controlli e prove","risks-and-controls",{"id":764,"name":765,"slug":766},49,"Controlli e prove","controls",{"id":768,"login":769,"email":391,"displayName":770},"20","rooth8233","Aleksandar Stajić",[772,1263],{"lang":773,"title":774,"content":775,"contentJson":776,"excerpt":1262},"en","How Far Is an AI Claim From the Evidence? — Evidence Distance","{\"time\":1792096795538,\"blocks\":[{\"id\":\"g-2Ciyj8LJ\",\"type\":\"header\",\"data\":{\"text\":\"Question\",\"level\":2},\"tunes\":{}},{\"id\":\"SQWD7UraiX\",\"type\":\"paragraph\",\"data\":{\"text\":\"Two AI answers can both cite sources and still differ greatly in how directly those sources support the claims being made. How can we describe that difference?\"},\"tunes\":{}},{\"id\":\"Sj1fx2865j\",\"type\":\"paragraph\",\"data\":{\"text\":\"This article uses the term Evidence Distance for the distance between a claim and the directly inspectable evidence that supports it.\"},\"tunes\":{}},{\"id\":\"d5th-A-mq5\",\"type\":\"quote\",\"data\":{\"text\":\"Evidence Distance is the number and strength of transformations, assumptions and inferential steps separating a claim from the evidence that can directly verify it.\",\"caption\":\"Working definition — a reasoning concept introduced here, not a standardized metric.\",\"alignment\":\"left\"},\"tunes\":{}},{\"id\":\"X81jXC72FC\",\"type\":\"tableOfContents\",\"data\":{\"title\":\"Contents\",\"minLevel\":2,\"maxLevel\":3},\"tunes\":{}},{\"id\":\"_pZcnPEoil\",\"type\":\"header\",\"data\":{\"text\":\"What This Really Means\",\"level\":2},\"tunes\":{}},{\"id\":\"F5odO7xyha\",\"type\":\"paragraph\",\"data\":{\"text\":\"A citation proves that a source was referenced. It does not automatically prove that the source directly supports the sentence beside the citation.\"},\"tunes\":{}},{\"id\":\"8Gqlg-ArJS\",\"type\":\"paragraph\",\"data\":{\"text\":\"Between an original observation and a final AI answer there may be extraction, parsing, chunking, retrieval, summarization, interpretation, aggregation and inference.\"},\"tunes\":{}},{\"id\":\"lXkL5LV9ab\",\"type\":\"code\",\"data\":{\"code\":\"Primary evidence\\n   ↓\\nExtracted data\\n   ↓\\nRetrieved passage\\n   ↓\\nSummary\\n   ↓\\nInterpretation\\n   ↓\\nInference\\n   ↓\\nFinal claim\"},\"tunes\":{}},{\"id\":\"Z-8GPQIpCI\",\"type\":\"paragraph\",\"data\":{\"text\":\"Every step can be legitimate. But every step can also introduce distortion.\"},\"tunes\":{}},{\"id\":\"aXMwTGZ8Ia\",\"type\":\"header\",\"data\":{\"text\":\"Simplest Example\",\"level\":2},\"tunes\":{}},{\"id\":\"bZlU7KD3M2\",\"type\":\"paragraph\",\"data\":{\"text\":\"Suppose a temperature sensor reports 38.2 °C.\"},\"tunes\":{}},{\"id\":\"SP87UUPaU3\",\"type\":\"table\",\"data\":{\"withHeadings\":true,\"stretched\":false,\"content\":[[\"Claim\",\"Relationship to evidence\",\"Illustrative distance\"],[\"The sensor reports 38.2 °C.\",\"Direct reading\",\"Low\"],[\"The system is running hotter than yesterday.\",\"Requires comparison with yesterday's reading\",\"Moderate\"],[\"The cooling system is failing.\",\"Requires causal inference and alternative explanations\",\"High\"]]},\"tunes\":{}},{\"id\":\"eYar059d69\",\"type\":\"paragraph\",\"data\":{\"text\":\"The original sensor value directly supports the first claim. It does not by itself prove the third.\"},\"tunes\":{}},{\"id\":\"eCnFlByvHt\",\"type\":\"code\",\"data\":{\"code\":\"Evidence exists\\n        ≠\\nEvidence directly supports every derived claim\"},\"tunes\":{}},{\"id\":\"aXGkK6BpkF\",\"type\":\"header\",\"data\":{\"text\":\"Where the Example Stops Working\",\"level\":2},\"tunes\":{}},{\"id\":\"lAwb9QZ_yB\",\"type\":\"paragraph\",\"data\":{\"text\":\"Evidence Distance is not simply a count of sentences or processing steps.\"},\"tunes\":{}},{\"id\":\"-SzYko9oIH\",\"type\":\"paragraph\",\"data\":{\"text\":\"One mathematically valid transformation may preserve evidence almost perfectly, while a single speculative interpretation may create a much larger epistemic jump.\"},\"tunes\":{}},{\"id\":\"XOQloBndLw\",\"type\":\"paragraph\",\"data\":{\"text\":\"The concept therefore includes both distance by transformation and distance by inference.\"},\"tunes\":{}},{\"id\":\"EOXbS4TAOR\",\"type\":\"header\",\"data\":{\"text\":\"Direct Answer\",\"level\":2},\"tunes\":{}},{\"id\":\"HcVL8vGM74\",\"type\":\"paragraph\",\"data\":{\"text\":\"Evidence Distance increases when a claim depends on more transformations, more assumptions, weaker provenance or stronger inference than the underlying evidence directly warrants.\"},\"tunes\":{}},{\"id\":\"_f7WXjUOsS\",\"type\":\"code\",\"data\":{\"code\":\"Short distance:\\nsource → claim\\n\\nLonger distance:\\nsource → extraction → summary → assumption → inference → claim\"},\"tunes\":{}},{\"id\":\"zUEC-cDhqm\",\"type\":\"paragraph\",\"data\":{\"text\":\"A longer distance does not automatically make a claim false. It means the reasoning chain requires more validation.\"},\"tunes\":{}},{\"id\":\"ix7TUGNz_W\",\"type\":\"header\",\"data\":{\"text\":\"Why This Matters for AI\",\"level\":2},\"tunes\":{}},{\"id\":\"wrAy5PLtLB\",\"type\":\"paragraph\",\"data\":{\"text\":\"LLMs are particularly good at producing fluent bridges between pieces of information. That capability is useful, but fluency can hide how far the conclusion has moved beyond the directly retrieved evidence.\"},\"tunes\":{}},{\"id\":\"jJseLuTWdv\",\"type\":\"paragraph\",\"data\":{\"text\":\"An answer can therefore look well sourced while containing claims whose support exists only through several unstated inferential steps.\"},\"tunes\":{}},{\"id\":\"ox8fok9dzs\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Fblog\u002Ffrom-research-protocol-to-a-general-ai-reasoning-framework\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"From Research Protocol to a General AI Reasoning Framework\",\"description\":\"Reliable reasoning requires explicit separation of evidence, assumptions, hypotheses and conclusions.\"}},\"tunes\":{}},{\"id\":\"6ipse6YG1U\",\"type\":\"header\",\"data\":{\"text\":\"Four Components of Evidence Distance\",\"level\":2},\"tunes\":{}},{\"id\":\"P4xQ7o96dR\",\"type\":\"header\",\"data\":{\"text\":\"1. Transformation Distance\",\"level\":3},\"tunes\":{}},{\"id\":\"dvN8TUNEYG\",\"type\":\"paragraph\",\"data\":{\"text\":\"How many times has the original evidence been converted, summarized, normalized or reformatted before reaching the model?\"},\"tunes\":{}},{\"id\":\"dXofACF9_x\",\"type\":\"header\",\"data\":{\"text\":\"2. Inferential Distance\",\"level\":3},\"tunes\":{}},{\"id\":\"kkEzV3ydId\",\"type\":\"paragraph\",\"data\":{\"text\":\"How many reasoning steps are required before the evidence supports the final conclusion?\"},\"tunes\":{}},{\"id\":\"PPYB0RuX5_\",\"type\":\"header\",\"data\":{\"text\":\"3. Provenance Distance\",\"level\":3},\"tunes\":{}},{\"id\":\"cAVBYebJr7\",\"type\":\"paragraph\",\"data\":{\"text\":\"Can the retrieved statement still be traced to an identifiable original source, or has it passed through summaries and copies whose origin is unclear?\"},\"tunes\":{}},{\"id\":\"254N_99Z7X\",\"type\":\"header\",\"data\":{\"text\":\"4. Applicability Distance\",\"level\":3},\"tunes\":{}},{\"id\":\"to3KWv5WSK\",\"type\":\"paragraph\",\"data\":{\"text\":\"How much reasoning is required to establish that the evidence applies to this exact version, user, jurisdiction, time period or situation?\"},\"tunes\":{}},{\"id\":\"3WnVPQ6wqw\",\"type\":\"header\",\"data\":{\"text\":\"Evidence Distance in RAG\",\"level\":2},\"tunes\":{}},{\"id\":\"r8s-75EBhA\",\"type\":\"paragraph\",\"data\":{\"text\":\"RAG can shorten one important distance: the distance between a model and external knowledge.\"},\"tunes\":{}},{\"id\":\"I_8h52A3Yc\",\"type\":\"paragraph\",\"data\":{\"text\":\"It does not automatically shorten the distance between a retrieved passage and the final claim.\"},\"tunes\":{}},{\"id\":\"jDDRUu2FUK\",\"type\":\"code\",\"data\":{\"code\":\"RAG solves:\\nmodel ↔ external information\\n\\nRAG does not automatically solve:\\nevidence ↔ justified conclusion\"},\"tunes\":{}},{\"id\":\"rm14qIVB1M\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Fblog\u002Fwhere-does-an-llm-get-its-data-rag-data-sources-in-python\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Where Does an LLM Get Its Data? RAG Data Sources in Python\",\"description\":\"Retrieval brings external information into context. Evidence Distance asks what happens between that information and the claim eventually produced.\"}},\"tunes\":{}},{\"id\":\"CRUZDXYloa\",\"type\":\"header\",\"data\":{\"text\":\"Connection to the Retrieval Trigger\",\"level\":2},\"tunes\":{}},{\"id\":\"-TPiYJ569c\",\"type\":\"paragraph\",\"data\":{\"text\":\"The Retrieval Trigger asks whether internal model knowledge is sufficient.\"},\"tunes\":{}},{\"id\":\"_VjLi5fs4x\",\"type\":\"paragraph\",\"data\":{\"text\":\"Evidence Distance asks a later question: after evidence has been retrieved, how directly does it support what the system wants to claim?\"},\"tunes\":{}},{\"id\":\"EImPw6i617\",\"type\":\"code\",\"data\":{\"code\":\"Need evidence?\\n   ↓\\nRetrieval Trigger\\n   ↓\\nRetrieve\\n   ↓\\nHow directly does evidence support the claim?\\n   ↓\\nEvidence Distance\"},\"tunes\":{}},{\"id\":\"ZqyhB2bFtR\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Fblog\u002Fwhen-should-an-ai-stop-trusting-its-own-knowledge-the-retrieval-trigger\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"When Should an AI Stop Trusting Its Own Knowledge? — The Retrieval Trigger\",\"description\":\"The trigger decides when evidence is needed. Evidence Distance evaluates how far the eventual claim moves beyond that evidence.\"}},\"tunes\":{}},{\"id\":\"ST2vnm6vwF\",\"type\":\"header\",\"data\":{\"text\":\"A Practical Diagnostic\",\"level\":2},\"tunes\":{}},{\"id\":\"qyoHcDcTKb\",\"type\":\"paragraph\",\"data\":{\"text\":\"For an important AI-generated claim, reconstruct the support chain backwards.\"},\"tunes\":{}},{\"id\":\"e3fDmLk0fe\",\"type\":\"code\",\"data\":{\"code\":\"CLAIM\\n  ↑\\nWhich inference produced it?\\n  ↑\\nWhich retrieved evidence supports that inference?\\n  ↑\\nWhich source produced that evidence?\\n  ↑\\nCan the original source still be inspected?\"},\"tunes\":{}},{\"id\":\"iQ60c7NRaE\",\"type\":\"paragraph\",\"data\":{\"text\":\"If the chain breaks, the system has encountered an evidence-traceability problem.\"},\"tunes\":{}},{\"id\":\"YQAKlNZ2gM\",\"type\":\"paragraph\",\"data\":{\"text\":\"If the chain remains intact but requires several strong assumptions, the Evidence Distance is still high.\"},\"tunes\":{}},{\"id\":\"clN7jFkUBf\",\"type\":\"header\",\"data\":{\"text\":\"Example: Product Data\",\"level\":2},\"tunes\":{}},{\"id\":\"CRF-A9O7Ot\",\"type\":\"paragraph\",\"data\":{\"text\":\"Suppose a manufacturer specification says a laptop weighs 1.48 kg.\"},\"tunes\":{}},{\"id\":\"JF3z6aF2q_\",\"type\":\"code\",\"data\":{\"code\":\"Manufacturer specification\\n→ 1.48 kg\\n→ \\\"The laptop weighs 1.48 kg\\\"\\n\\nLow Evidence Distance\"},\"tunes\":{}},{\"id\":\"vTMOOgkLPb\",\"type\":\"paragraph\",\"data\":{\"text\":\"Now consider:\"},\"tunes\":{}},{\"id\":\"Tuhfzor1yp\",\"type\":\"code\",\"data\":{\"code\":\"Manufacturer specification\\n→ 1.48 kg\\n→ comparison with three competitors\\n→ assumption about travel behavior\\n→ \\\"This is the best laptop for frequent travelers\\\"\\n\\nMuch higher Evidence Distance\"},\"tunes\":{}},{\"id\":\"JHFRGIgUJD\",\"type\":\"paragraph\",\"data\":{\"text\":\"The specification is real. The recommendation is still several reasoning steps away from it.\"},\"tunes\":{}},{\"id\":\"2lKERCrnW1\",\"type\":\"header\",\"data\":{\"text\":\"Example: Scientific Evidence\",\"level\":2},\"tunes\":{}},{\"id\":\"HA8DBYwUuK\",\"type\":\"paragraph\",\"data\":{\"text\":\"A paper may report an observed association. An AI answer can incorrectly compress that into a causal statement.\"},\"tunes\":{}},{\"id\":\"D9C2dVB7j_\",\"type\":\"code\",\"data\":{\"code\":\"Observed association\\n        ↓\\nInterpretation\\n        ↓\\nUnstated causal assumption\\n        ↓\\n\\\"X causes Y\\\"\"},\"tunes\":{}},{\"id\":\"CKNjK25pkL\",\"type\":\"paragraph\",\"data\":{\"text\":\"The problem is not that no evidence exists. The problem is that the final statement crosses a larger evidential distance than the study directly supports.\"},\"tunes\":{}},{\"id\":\"sR430N3cDd\",\"type\":\"header\",\"data\":{\"text\":\"Connection to Agent-Ready Context\",\"level\":2},\"tunes\":{}},{\"id\":\"v77YyG55ko\",\"type\":\"paragraph\",\"data\":{\"text\":\"Structured information objects can reduce avoidable Evidence Distance by preserving source identity, units, timestamps, conflicts and validity instead of forcing the model to reconstruct them from flattened text.\"},\"tunes\":{}},{\"id\":\"EX8darCIjd\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Fportfolio\u002Ffrom-raw-data-to-verifiable-ai-context-a-practical-architecture-for-agent-ready-information\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"From Raw Data to Verifiable AI Context: A Practical Architecture for Agent-Ready Information\",\"description\":\"Preserving provenance and validity before retrieval reduces the number of unsupported reconstruction steps required later.\"}},\"tunes\":{}},{\"id\":\"dsfuoURj1e\",\"type\":\"header\",\"data\":{\"text\":\"A Simple Working Scale\",\"level\":2},\"tunes\":{}},{\"id\":\"BLi7Yth64O\",\"type\":\"table\",\"data\":{\"withHeadings\":true,\"stretched\":false,\"content\":[[\"Level\",\"Description\"],[\"ED0\",\"Directly reported source fact\"],[\"ED1\",\"Simple deterministic transformation or direct comparison\"],[\"ED2\",\"One meaningful interpretive step\"],[\"ED3\",\"Multiple assumptions or combined sources\"],[\"ED4\",\"Strong causal, predictive or prescriptive conclusion beyond direct evidence\"]]},\"tunes\":{}},{\"id\":\"8y8bzCcHBs\",\"type\":\"paragraph\",\"data\":{\"text\":\"This scale is illustrative, not a scientific standard. Its purpose is to force an explicit question: How much did we have to add between the evidence and the claim?\"},\"tunes\":{}},{\"id\":\"aFK0YiAlYO\",\"type\":\"header\",\"data\":{\"text\":\"Misconceptions\",\"level\":2},\"tunes\":{}},{\"id\":\"07KSk6xDx-\",\"type\":\"paragraph\",\"data\":{\"text\":\"A citation means Evidence Distance is low. No. A citation can sit beside a claim that substantially exceeds what the cited source establishes.\"},\"tunes\":{}},{\"id\":\"t_cf-Zfan9\",\"type\":\"paragraph\",\"data\":{\"text\":\"Long Evidence Distance means the claim is wrong. No. Complex conclusions can be well supported, but they require a stronger reasoning chain.\"},\"tunes\":{}},{\"id\":\"Pu9PRX67Q6\",\"type\":\"paragraph\",\"data\":{\"text\":\"Primary sources eliminate inference. No. Primary evidence can still require interpretation.\"},\"tunes\":{}},{\"id\":\"oF1xNsOHHq\",\"type\":\"paragraph\",\"data\":{\"text\":\"More sources automatically reduce distance. Ten weak repetitions of the same unsupported claim do not create direct evidence.\"},\"tunes\":{}},{\"id\":\"IkolK5baph\",\"type\":\"header\",\"data\":{\"text\":\"Edge Case: Derived Data\",\"level\":2},\"tunes\":{}},{\"id\":\"jRH36fwdR3\",\"type\":\"paragraph\",\"data\":{\"text\":\"Some derived values are more reliable than raw human descriptions. A correctly calculated total from verified transactions can be highly dependable even though it is one transformation away from the original records.\"},\"tunes\":{}},{\"id\":\"SXbdKiNDst\",\"type\":\"paragraph\",\"data\":{\"text\":\"Evidence Distance therefore measures epistemic separation, not a simplistic preference for untouched data.\"},\"tunes\":{}},{\"id\":\"mUbvdnFChM\",\"type\":\"header\",\"data\":{\"text\":\"Limitations\",\"level\":2},\"tunes\":{}},{\"id\":\"N_EskxFOep\",\"type\":\"paragraph\",\"data\":{\"text\":\"Evidence Distance is a conceptual diagnostic, not a universal quantitative score.\"},\"tunes\":{}},{\"id\":\"nW_-E08So_\",\"type\":\"paragraph\",\"data\":{\"text\":\"Different domains require different treatment of inference. Mathematical proof, clinical evidence, software logs, market analysis and historical interpretation cannot share one universal numerical threshold.\"},\"tunes\":{}},{\"id\":\"_Bd95wXkri\",\"type\":\"header\",\"data\":{\"text\":\"What Would Change This Answer?\",\"level\":2},\"tunes\":{}},{\"id\":\"OqCaMc2_OX\",\"type\":\"paragraph\",\"data\":{\"text\":\"Future AI systems may generate explicit provenance graphs, machine-checkable claim-evidence links and typed reasoning traces. That could make parts of Evidence Distance measurable automatically.\"},\"tunes\":{}},{\"id\":\"58Hf3O9-Iz\",\"type\":\"paragraph\",\"data\":{\"text\":\"The underlying problem remains: a claim can move farther from its evidence even while every intermediate sentence sounds plausible.\"},\"tunes\":{}},{\"id\":\"8WOq5oE_e_\",\"type\":\"header\",\"data\":{\"text\":\"Conclusion\",\"level\":2},\"tunes\":{}},{\"id\":\"1pr1TfwnKc\",\"type\":\"paragraph\",\"data\":{\"text\":\"The important distinction is not merely between sourced and unsourced answers.\"},\"tunes\":{}},{\"id\":\"54CysT_NXa\",\"type\":\"paragraph\",\"data\":{\"text\":\"It is between claims that remain close to directly inspectable evidence and claims that depend on increasingly long chains of interpretation and assumption.\"},\"tunes\":{}},{\"id\":\"Qrjo3WEUSL\",\"type\":\"code\",\"data\":{\"code\":\"Evidence\\n   ↓\\nEvidence Distance\\n   ↓\\nReasoning\\n   ↓\\nAnswer Validity Boundary\"},\"tunes\":{}},{\"id\":\"6ir8aIAfd3\",\"type\":\"paragraph\",\"data\":{\"text\":\"Evidence Distance gives us a language for describing that space.\"},\"tunes\":{}},{\"id\":\"vPhupXcbeN\",\"type\":\"header\",\"data\":{\"text\":\"Primary Sources\",\"level\":2},\"tunes\":{}},{\"id\":\"BaO4JmWRQL\",\"type\":\"paragraph\",\"data\":{\"text\":\"Patrick Lewis et al., \u003Ca href=\\\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2005.11401\\\" target=\\\"_blank\\\">Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks\u003C\u002Fa> (2020).\"},\"tunes\":{}},{\"id\":\"eg4Z4G6Cgs\",\"type\":\"paragraph\",\"data\":{\"text\":\"Nelson F. Liu et al., \u003Ca href=\\\"https:\u002F\u002Farxiv.org\u002Fabs\u002F2307.03172\\\" target=\\\"_blank\\\">Lost in the Middle: How Language Models Use Long Contexts\u003C\u002Fa> (2023).\"},\"tunes\":{}},{\"id\":\"BJH7zUd4T8\",\"type\":\"paragraph\",\"data\":{\"text\":\"W3C, \u003Ca href=\\\"https:\u002F\u002Fwww.w3.org\u002FTR\u002Fprov-overview\u002F\\\" target=\\\"_blank\\\">PROV Overview\u003C\u002Fa>. A standardized model for representing provenance relationships.\"},\"tunes\":{}}],\"version\":\"2.31.6\"}",{"time":777,"blocks":778,"version":741},1792096795538,[779,784,789,794,800,804,809,814,819,823,828,833,838,859,864,868,873,878,883,888,893,898,902,907,912,917,922,931,936,941,946,951,956,961,966,971,976,981,986,991,995,1004,1009,1014,1019,1023,1032,1037,1042,1046,1051,1056,1061,1066,1070,1075,1079,1084,1089,1094,1098,1103,1108,1113,1122,1127,1145,1150,1155,1160,1165,1170,1175,1180,1185,1190,1195,1200,1205,1210,1215,1220,1225,1230,1235,1239,1244,1249,1253,1257],{"id":780,"data":781,"type":42,"tunes":783},"g-2Ciyj8LJ",{"text":782,"level":234},"Question",{},{"id":785,"data":786,"type":240,"tunes":788},"SQWD7UraiX",{"text":787},"Two AI answers can both cite sources and still differ greatly in how directly those sources support the claims being made. How can we describe that difference?",{},{"id":790,"data":791,"type":240,"tunes":793},"Sj1fx2865j",{"text":792},"This article uses the term Evidence Distance for the distance between a claim and the directly inspectable evidence that supports it.",{},{"id":795,"data":796,"type":253,"tunes":799},"d5th-A-mq5",{"text":797,"caption":798,"alignment":252},"Evidence Distance is the number and strength of transformations, assumptions and inferential steps separating a claim from the evidence that can directly verify it.","Working definition — a reasoning concept introduced here, not a standardized metric.",{},{"id":801,"data":802,"type":260,"tunes":803},"X81jXC72FC",{"title":258,"maxLevel":259,"minLevel":234},{},{"id":805,"data":806,"type":42,"tunes":808},"_pZcnPEoil",{"text":807,"level":234},"What This Really Means",{},{"id":810,"data":811,"type":240,"tunes":813},"F5odO7xyha",{"text":812},"A citation proves that a source was referenced. It does not automatically prove that the source directly supports the sentence beside the citation.",{},{"id":815,"data":816,"type":240,"tunes":818},"8Gqlg-ArJS",{"text":817},"Between an original observation and a final AI answer there may be extraction, parsing, chunking, retrieval, summarization, interpretation, aggregation and inference.",{},{"id":820,"data":821,"type":281,"tunes":822},"lXkL5LV9ab",{"code":280},{},{"id":824,"data":825,"type":240,"tunes":827},"Z-8GPQIpCI",{"text":826},"Every step can be legitimate. But every step can also introduce distortion.",{},{"id":829,"data":830,"type":42,"tunes":832},"aXMwTGZ8Ia",{"text":831,"level":234},"Simplest Example",{},{"id":834,"data":835,"type":240,"tunes":837},"bZlU7KD3M2",{"text":836},"Suppose a temperature sensor reports 38.2 °C.",{},{"id":839,"data":840,"type":318,"tunes":858},"SP87UUPaU3",{"content":841,"stretched":43,"withHeadings":14},[842,846,850,854],[843,844,845],"Claim","Relationship to evidence","Illustrative distance",[847,848,849],"The sensor reports 38.2 °C.","Direct reading","Low",[851,852,853],"The system is running hotter than yesterday.","Requires comparison with yesterday's reading","Moderate",[855,856,857],"The cooling system is failing.","Requires causal inference and alternative explanations","High",{},{"id":860,"data":861,"type":240,"tunes":863},"eYar059d69",{"text":862},"The original sensor value directly supports the first claim. It does not by itself prove the third.",{},{"id":865,"data":866,"type":281,"tunes":867},"eCnFlByvHt",{"code":328},{},{"id":869,"data":870,"type":42,"tunes":872},"aXGkK6BpkF",{"text":871,"level":234},"Where the Example Stops Working",{},{"id":874,"data":875,"type":240,"tunes":877},"lAwb9QZ_yB",{"text":876},"Evidence Distance is not simply a count of sentences or processing steps.",{},{"id":879,"data":880,"type":240,"tunes":882},"-SzYko9oIH",{"text":881},"One mathematically valid transformation may preserve evidence almost perfectly, while a single speculative interpretation may create a much larger epistemic jump.",{},{"id":884,"data":885,"type":240,"tunes":887},"XOQloBndLw",{"text":886},"The concept therefore includes both distance by transformation and distance by inference.",{},{"id":889,"data":890,"type":42,"tunes":892},"EOXbS4TAOR",{"text":891,"level":234},"Direct Answer",{},{"id":894,"data":895,"type":240,"tunes":897},"HcVL8vGM74",{"text":896},"Evidence Distance increases when a claim depends on more transformations, more assumptions, weaker provenance or stronger inference than the underlying evidence directly warrants.",{},{"id":899,"data":900,"type":281,"tunes":901},"_f7WXjUOsS",{"code":363},{},{"id":903,"data":904,"type":240,"tunes":906},"zUEC-cDhqm",{"text":905},"A longer distance does not automatically make a claim false. It means the reasoning chain requires more validation.",{},{"id":908,"data":909,"type":42,"tunes":911},"ix7TUGNz_W",{"text":910,"level":234},"Why This Matters for AI",{},{"id":913,"data":914,"type":240,"tunes":916},"wrAy5PLtLB",{"text":915},"LLMs are particularly good at producing fluent bridges between pieces of information. That capability is useful, but fluency can hide how far the conclusion has moved beyond the directly retrieved evidence.",{},{"id":918,"data":919,"type":240,"tunes":921},"jJseLuTWdv",{"text":920},"An answer can therefore look well sourced while containing claims whose support exists only through several unstated inferential steps.",{},{"id":923,"data":924,"type":394,"tunes":930},"ox8fok9dzs",{"link":925,"meta":926},"https:\u002F\u002Fstajic.de\u002Fblog\u002Ffrom-research-protocol-to-a-general-ai-reasoning-framework",{"image":927,"title":928,"description":929},{"url":391},"From Research Protocol to a General AI Reasoning Framework","Reliable reasoning requires explicit separation of evidence, assumptions, hypotheses and conclusions.",{},{"id":932,"data":933,"type":42,"tunes":935},"6ipse6YG1U",{"text":934,"level":234},"Four Components of Evidence Distance",{},{"id":937,"data":938,"type":42,"tunes":940},"P4xQ7o96dR",{"text":939,"level":259},"1. Transformation Distance",{},{"id":942,"data":943,"type":240,"tunes":945},"dvN8TUNEYG",{"text":944},"How many times has the original evidence been converted, summarized, normalized or reformatted before reaching the model?",{},{"id":947,"data":948,"type":42,"tunes":950},"dXofACF9_x",{"text":949,"level":259},"2. Inferential Distance",{},{"id":952,"data":953,"type":240,"tunes":955},"kkEzV3ydId",{"text":954},"How many reasoning steps are required before the evidence supports the final conclusion?",{},{"id":957,"data":958,"type":42,"tunes":960},"PPYB0RuX5_",{"text":959,"level":259},"3. Provenance Distance",{},{"id":962,"data":963,"type":240,"tunes":965},"cAVBYebJr7",{"text":964},"Can the retrieved statement still be traced to an identifiable original source, or has it passed through summaries and copies whose origin is unclear?",{},{"id":967,"data":968,"type":42,"tunes":970},"254N_99Z7X",{"text":969,"level":259},"4. Applicability Distance",{},{"id":972,"data":973,"type":240,"tunes":975},"to3KWv5WSK",{"text":974},"How much reasoning is required to establish that the evidence applies to this exact version, user, jurisdiction, time period or situation?",{},{"id":977,"data":978,"type":42,"tunes":980},"3WnVPQ6wqw",{"text":979,"level":234},"Evidence Distance in RAG",{},{"id":982,"data":983,"type":240,"tunes":985},"r8s-75EBhA",{"text":984},"RAG can shorten one important distance: the distance between a model and external knowledge.",{},{"id":987,"data":988,"type":240,"tunes":990},"I_8h52A3Yc",{"text":989},"It does not automatically shorten the distance between a retrieved passage and the final claim.",{},{"id":992,"data":993,"type":281,"tunes":994},"jDDRUu2FUK",{"code":459},{},{"id":996,"data":997,"type":394,"tunes":1003},"rm14qIVB1M",{"link":998,"meta":999},"https:\u002F\u002Fstajic.de\u002Fblog\u002Fwhere-does-an-llm-get-its-data-rag-data-sources-in-python",{"image":1000,"title":1001,"description":1002},{"url":391},"Where Does an LLM Get Its Data? RAG Data Sources in Python","Retrieval brings external information into context. Evidence Distance asks what happens between that information and the claim eventually produced.",{},{"id":1005,"data":1006,"type":42,"tunes":1008},"CRUZDXYloa",{"text":1007,"level":234},"Connection to the Retrieval Trigger",{},{"id":1010,"data":1011,"type":240,"tunes":1013},"-TPiYJ569c",{"text":1012},"The Retrieval Trigger asks whether internal model knowledge is sufficient.",{},{"id":1015,"data":1016,"type":240,"tunes":1018},"_VjLi5fs4x",{"text":1017},"Evidence Distance asks a later question: after evidence has been retrieved, how directly does it support what the system wants to claim?",{},{"id":1020,"data":1021,"type":281,"tunes":1022},"EImPw6i617",{"code":488},{},{"id":1024,"data":1025,"type":394,"tunes":1031},"ZqyhB2bFtR",{"link":1026,"meta":1027},"https:\u002F\u002Fstajic.de\u002Fblog\u002Fwhen-should-an-ai-stop-trusting-its-own-knowledge-the-retrieval-trigger",{"image":1028,"title":1029,"description":1030},{"url":391},"When Should an AI Stop Trusting Its Own Knowledge? — The Retrieval Trigger","The trigger decides when evidence is needed. Evidence Distance evaluates how far the eventual claim moves beyond that evidence.",{},{"id":1033,"data":1034,"type":42,"tunes":1036},"ST2vnm6vwF",{"text":1035,"level":234},"A Practical Diagnostic",{},{"id":1038,"data":1039,"type":240,"tunes":1041},"qyoHcDcTKb",{"text":1040},"For an important AI-generated claim, reconstruct the support chain backwards.",{},{"id":1043,"data":1044,"type":281,"tunes":1045},"e3fDmLk0fe",{"code":512},{},{"id":1047,"data":1048,"type":240,"tunes":1050},"iQ60c7NRaE",{"text":1049},"If the chain breaks, the system has encountered an evidence-traceability problem.",{},{"id":1052,"data":1053,"type":240,"tunes":1055},"YQAKlNZ2gM",{"text":1054},"If the chain remains intact but requires several strong assumptions, the Evidence Distance is still high.",{},{"id":1057,"data":1058,"type":42,"tunes":1060},"clN7jFkUBf",{"text":1059,"level":234},"Example: Product Data",{},{"id":1062,"data":1063,"type":240,"tunes":1065},"CRF-A9O7Ot",{"text":1064},"Suppose a manufacturer specification says a laptop weighs 1.48 kg.",{},{"id":1067,"data":1068,"type":281,"tunes":1069},"JF3z6aF2q_",{"code":537},{},{"id":1071,"data":1072,"type":240,"tunes":1074},"vTMOOgkLPb",{"text":1073},"Now consider:",{},{"id":1076,"data":1077,"type":281,"tunes":1078},"Tuhfzor1yp",{"code":547},{},{"id":1080,"data":1081,"type":240,"tunes":1083},"JHFRGIgUJD",{"text":1082},"The specification is real. The recommendation is still several reasoning steps away from it.",{},{"id":1085,"data":1086,"type":42,"tunes":1088},"2lKERCrnW1",{"text":1087,"level":234},"Example: Scientific Evidence",{},{"id":1090,"data":1091,"type":240,"tunes":1093},"HA8DBYwUuK",{"text":1092},"A paper may report an observed association. An AI answer can incorrectly compress that into a causal statement.",{},{"id":1095,"data":1096,"type":281,"tunes":1097},"D9C2dVB7j_",{"code":567},{},{"id":1099,"data":1100,"type":240,"tunes":1102},"CKNjK25pkL",{"text":1101},"The problem is not that no evidence exists. The problem is that the final statement crosses a larger evidential distance than the study directly supports.",{},{"id":1104,"data":1105,"type":42,"tunes":1107},"sR430N3cDd",{"text":1106,"level":234},"Connection to Agent-Ready Context",{},{"id":1109,"data":1110,"type":240,"tunes":1112},"v77YyG55ko",{"text":1111},"Structured information objects can reduce avoidable Evidence Distance by preserving source identity, units, timestamps, conflicts and validity instead of forcing the model to reconstruct them from flattened text.",{},{"id":1114,"data":1115,"type":394,"tunes":1121},"EX8darCIjd",{"link":1116,"meta":1117},"https:\u002F\u002Fstajic.de\u002Fportfolio\u002Ffrom-raw-data-to-verifiable-ai-context-a-practical-architecture-for-agent-ready-information",{"image":1118,"title":1119,"description":1120},{"url":391},"From Raw Data to Verifiable AI Context: A Practical Architecture for Agent-Ready Information","Preserving provenance and validity before retrieval reduces the number of unsupported reconstruction steps required later.",{},{"id":1123,"data":1124,"type":42,"tunes":1126},"dsfuoURj1e",{"text":1125,"level":234},"A Simple Working Scale",{},{"id":1128,"data":1129,"type":318,"tunes":1144},"BLi7Yth64O",{"content":1130,"stretched":43,"withHeadings":14},[1131,1134,1136,1138,1140,1142],[1132,1133],"Level","Description",[606,1135],"Directly reported source fact",[609,1137],"Simple deterministic transformation or direct comparison",[612,1139],"One meaningful interpretive step",[615,1141],"Multiple assumptions or combined sources",[618,1143],"Strong causal, predictive or prescriptive conclusion beyond direct evidence",{},{"id":1146,"data":1147,"type":240,"tunes":1149},"8y8bzCcHBs",{"text":1148},"This scale is illustrative, not a scientific standard. Its purpose is to force an explicit question: How much did we have to add between the evidence and the claim?",{},{"id":1151,"data":1152,"type":42,"tunes":1154},"aFK0YiAlYO",{"text":1153,"level":234},"Misconceptions",{},{"id":1156,"data":1157,"type":240,"tunes":1159},"07KSk6xDx-",{"text":1158},"A citation means Evidence Distance is low. No. A citation can sit beside a claim that substantially exceeds what the cited source establishes.",{},{"id":1161,"data":1162,"type":240,"tunes":1164},"t_cf-Zfan9",{"text":1163},"Long Evidence Distance means the claim is wrong. No. Complex conclusions can be well supported, but they require a stronger reasoning chain.",{},{"id":1166,"data":1167,"type":240,"tunes":1169},"Pu9PRX67Q6",{"text":1168},"Primary sources eliminate inference. No. Primary evidence can still require interpretation.",{},{"id":1171,"data":1172,"type":240,"tunes":1174},"oF1xNsOHHq",{"text":1173},"More sources automatically reduce distance. Ten weak repetitions of the same unsupported claim do not create direct evidence.",{},{"id":1176,"data":1177,"type":42,"tunes":1179},"IkolK5baph",{"text":1178,"level":234},"Edge Case: Derived Data",{},{"id":1181,"data":1182,"type":240,"tunes":1184},"jRH36fwdR3",{"text":1183},"Some derived values are more reliable than raw human descriptions. A correctly calculated total from verified transactions can be highly dependable even though it is one transformation away from the original records.",{},{"id":1186,"data":1187,"type":240,"tunes":1189},"SXbdKiNDst",{"text":1188},"Evidence Distance therefore measures epistemic separation, not a simplistic preference for untouched data.",{},{"id":1191,"data":1192,"type":42,"tunes":1194},"mUbvdnFChM",{"text":1193,"level":234},"Limitations",{},{"id":1196,"data":1197,"type":240,"tunes":1199},"N_EskxFOep",{"text":1198},"Evidence Distance is a conceptual diagnostic, not a universal quantitative score.",{},{"id":1201,"data":1202,"type":240,"tunes":1204},"nW_-E08So_",{"text":1203},"Different domains require different treatment of inference. Mathematical proof, clinical evidence, software logs, market analysis and historical interpretation cannot share one universal numerical threshold.",{},{"id":1206,"data":1207,"type":42,"tunes":1209},"_Bd95wXkri",{"text":1208,"level":234},"What Would Change This Answer?",{},{"id":1211,"data":1212,"type":240,"tunes":1214},"OqCaMc2_OX",{"text":1213},"Future AI systems may generate explicit provenance graphs, machine-checkable claim-evidence links and typed reasoning traces. That could make parts of Evidence Distance measurable automatically.",{},{"id":1216,"data":1217,"type":240,"tunes":1219},"58Hf3O9-Iz",{"text":1218},"The underlying problem remains: a claim can move farther from its evidence even while every intermediate sentence sounds plausible.",{},{"id":1221,"data":1222,"type":42,"tunes":1224},"8WOq5oE_e_",{"text":1223,"level":234},"Conclusion",{},{"id":1226,"data":1227,"type":240,"tunes":1229},"1pr1TfwnKc",{"text":1228},"The important distinction is not merely between sourced and unsourced answers.",{},{"id":1231,"data":1232,"type":240,"tunes":1234},"54CysT_NXa",{"text":1233},"It is between claims that remain close to directly inspectable evidence and claims that depend on increasingly long chains of interpretation and assumption.",{},{"id":1236,"data":1237,"type":281,"tunes":1238},"Qrjo3WEUSL",{"code":714},{},{"id":1240,"data":1241,"type":240,"tunes":1243},"6ir8aIAfd3",{"text":1242},"Evidence Distance gives us a language for describing that space.",{},{"id":1245,"data":1246,"type":42,"tunes":1248},"vPhupXcbeN",{"text":1247,"level":234},"Primary Sources",{},{"id":1250,"data":1251,"type":240,"tunes":1252},"BaO4JmWRQL",{"text":729},{},{"id":1254,"data":1255,"type":240,"tunes":1256},"eg4Z4G6Cgs",{"text":734},{},{"id":1258,"data":1259,"type":240,"tunes":1261},"BJH7zUd4T8",{"text":1260},"W3C, \u003Ca href=\"https:\u002F\u002Fwww.w3.org\u002FTR\u002Fprov-overview\u002F\" target=\"_blank\">PROV Overview\u003C\u002Fa>. A standardized model for representing provenance relationships.",{},"Not all supported answers are equally close to their evidence. Evidence Distance is a practical way to describe how many transformations, summaries, assumptions and inferential steps separate an AI claim from directly inspectable evidence.",{"lang":7,"title":224,"content":226,"contentJson":1264,"excerpt":742},{"time":228,"blocks":1265,"version":741},[1266,1269,1272,1275,1278,1281,1284,1287,1290,1293,1296,1299,1302,1310,1313,1316,1319,1322,1325,1328,1331,1334,1337,1340,1343,1346,1349,1354,1357,1360,1363,1366,1369,1372,1375,1378,1381,1384,1387,1390,1393,1398,1401,1404,1407,1410,1415,1418,1421,1424,1427,1430,1433,1436,1439,1442,1445,1448,1451,1454,1457,1460,1463,1466,1471,1474,1484,1487,1490,1493,1496,1499,1502,1505,1508,1511,1514,1517,1520,1523,1526,1529,1532,1535,1538,1541,1544,1547,1550,1553],{"id":231,"data":1267,"type":42,"tunes":1268},{"text":233,"level":234},{},{"id":237,"data":1270,"type":240,"tunes":1271},{"text":239},{},{"id":243,"data":1273,"type":240,"tunes":1274},{"text":245},{},{"id":248,"data":1276,"type":253,"tunes":1277},{"text":250,"caption":251,"alignment":252},{},{"id":256,"data":1279,"type":260,"tunes":1280},{"title":258,"maxLevel":259,"minLevel":234},{},{"id":263,"data":1282,"type":42,"tunes":1283},{"text":265,"level":234},{},{"id":268,"data":1285,"type":240,"tunes":1286},{"text":270},{},{"id":273,"data":1288,"type":240,"tunes":1289},{"text":275},{},{"id":278,"data":1291,"type":281,"tunes":1292},{"code":280},{},{"id":284,"data":1294,"type":240,"tunes":1295},{"text":286},{},{"id":289,"data":1297,"type":42,"tunes":1298},{"text":291,"level":234},{},{"id":294,"data":1300,"type":240,"tunes":1301},{"text":296},{},{"id":299,"data":1303,"type":318,"tunes":1309},{"content":1304,"stretched":43,"withHeadings":14},[1305,1306,1307,1308],[303,304,305],[307,308,309],[311,312,313],[315,316,317],{},{"id":321,"data":1311,"type":240,"tunes":1312},{"text":323},{},{"id":326,"data":1314,"type":281,"tunes":1315},{"code":328},{},{"id":331,"data":1317,"type":42,"tunes":1318},{"text":333,"level":234},{},{"id":336,"data":1320,"type":240,"tunes":1321},{"text":338},{},{"id":341,"data":1323,"type":240,"tunes":1324},{"text":343},{},{"id":346,"data":1326,"type":240,"tunes":1327},{"text":348},{},{"id":351,"data":1329,"type":42,"tunes":1330},{"text":353,"level":234},{},{"id":356,"data":1332,"type":240,"tunes":1333},{"text":358},{},{"id":361,"data":1335,"type":281,"tunes":1336},{"code":363},{},{"id":366,"data":1338,"type":240,"tunes":1339},{"text":368},{},{"id":371,"data":1341,"type":42,"tunes":1342},{"text":373,"level":234},{},{"id":376,"data":1344,"type":240,"tunes":1345},{"text":378},{},{"id":381,"data":1347,"type":240,"tunes":1348},{"text":383},{},{"id":386,"data":1350,"type":394,"tunes":1353},{"link":388,"meta":1351},{"image":1352,"title":392,"description":393},{"url":391},{},{"id":397,"data":1355,"type":42,"tunes":1356},{"text":399,"level":234},{},{"id":402,"data":1358,"type":42,"tunes":1359},{"text":404,"level":259},{},{"id":407,"data":1361,"type":240,"tunes":1362},{"text":409},{},{"id":412,"data":1364,"type":42,"tunes":1365},{"text":414,"level":259},{},{"id":417,"data":1367,"type":240,"tunes":1368},{"text":419},{},{"id":422,"data":1370,"type":42,"tunes":1371},{"text":424,"level":259},{},{"id":427,"data":1373,"type":240,"tunes":1374},{"text":429},{},{"id":432,"data":1376,"type":42,"tunes":1377},{"text":434,"level":259},{},{"id":437,"data":1379,"type":240,"tunes":1380},{"text":439},{},{"id":442,"data":1382,"type":42,"tunes":1383},{"text":444,"level":234},{},{"id":447,"data":1385,"type":240,"tunes":1386},{"text":449},{},{"id":452,"data":1388,"type":240,"tunes":1389},{"text":454},{},{"id":457,"data":1391,"type":281,"tunes":1392},{"code":459},{},{"id":462,"data":1394,"type":394,"tunes":1397},{"link":464,"meta":1395},{"image":1396,"title":467,"description":468},{"url":391},{},{"id":471,"data":1399,"type":42,"tunes":1400},{"text":473,"level":234},{},{"id":476,"data":1402,"type":240,"tunes":1403},{"text":478},{},{"id":481,"data":1405,"type":240,"tunes":1406},{"text":483},{},{"id":486,"data":1408,"type":281,"tunes":1409},{"code":488},{},{"id":491,"data":1411,"type":394,"tunes":1414},{"link":493,"meta":1412},{"image":1413,"title":496,"description":497},{"url":391},{},{"id":500,"data":1416,"type":42,"tunes":1417},{"text":502,"level":234},{},{"id":505,"data":1419,"type":240,"tunes":1420},{"text":507},{},{"id":510,"data":1422,"type":281,"tunes":1423},{"code":512},{},{"id":515,"data":1425,"type":240,"tunes":1426},{"text":517},{},{"id":520,"data":1428,"type":240,"tunes":1429},{"text":522},{},{"id":525,"data":1431,"type":42,"tunes":1432},{"text":527,"level":234},{},{"id":530,"data":1434,"type":240,"tunes":1435},{"text":532},{},{"id":535,"data":1437,"type":281,"tunes":1438},{"code":537},{},{"id":540,"data":1440,"type":240,"tunes":1441},{"text":542},{},{"id":545,"data":1443,"type":281,"tunes":1444},{"code":547},{},{"id":550,"data":1446,"type":240,"tunes":1447},{"text":552},{},{"id":555,"data":1449,"type":42,"tunes":1450},{"text":557,"level":234},{},{"id":560,"data":1452,"type":240,"tunes":1453},{"text":562},{},{"id":565,"data":1455,"type":281,"tunes":1456},{"code":567},{},{"id":570,"data":1458,"type":240,"tunes":1459},{"text":572},{},{"id":575,"data":1461,"type":42,"tunes":1462},{"text":577,"level":234},{},{"id":580,"data":1464,"type":240,"tunes":1465},{"text":582},{},{"id":585,"data":1467,"type":394,"tunes":1470},{"link":587,"meta":1468},{"image":1469,"title":590,"description":591},{"url":391},{},{"id":594,"data":1472,"type":42,"tunes":1473},{"text":596,"level":234},{},{"id":599,"data":1475,"type":318,"tunes":1483},{"content":1476,"stretched":43,"withHeadings":14},[1477,1478,1479,1480,1481,1482],[603,604],[606,607],[609,610],[612,613],[615,616],[618,619],{},{"id":622,"data":1485,"type":240,"tunes":1486},{"text":624},{},{"id":627,"data":1488,"type":42,"tunes":1489},{"text":629,"level":234},{},{"id":632,"data":1491,"type":240,"tunes":1492},{"text":634},{},{"id":637,"data":1494,"type":240,"tunes":1495},{"text":639},{},{"id":642,"data":1497,"type":240,"tunes":1498},{"text":644},{},{"id":647,"data":1500,"type":240,"tunes":1501},{"text":649},{},{"id":652,"data":1503,"type":42,"tunes":1504},{"text":654,"level":234},{},{"id":657,"data":1506,"type":240,"tunes":1507},{"text":659},{},{"id":662,"data":1509,"type":240,"tunes":1510},{"text":664},{},{"id":667,"data":1512,"type":42,"tunes":1513},{"text":669,"level":234},{},{"id":672,"data":1515,"type":240,"tunes":1516},{"text":674},{},{"id":677,"data":1518,"type":240,"tunes":1519},{"text":679},{},{"id":682,"data":1521,"type":42,"tunes":1522},{"text":684,"level":234},{},{"id":687,"data":1524,"type":240,"tunes":1525},{"text":689},{},{"id":692,"data":1527,"type":240,"tunes":1528},{"text":694},{},{"id":697,"data":1530,"type":42,"tunes":1531},{"text":699,"level":234},{},{"id":702,"data":1533,"type":240,"tunes":1534},{"text":704},{},{"id":707,"data":1536,"type":240,"tunes":1537},{"text":709},{},{"id":712,"data":1539,"type":281,"tunes":1540},{"code":714},{},{"id":717,"data":1542,"type":240,"tunes":1543},{"text":719},{},{"id":722,"data":1545,"type":42,"tunes":1546},{"text":724,"level":234},{},{"id":727,"data":1548,"type":240,"tunes":1549},{"text":729},{},{"id":732,"data":1551,"type":240,"tunes":1552},{"text":734},{},{"id":737,"data":1554,"type":240,"tunes":1555},{"text":739},{},"Post erfolgreich abgerufen",{"items":1558,"source":1615,"manualIds":1616,"manualMatchedIds":1617},[1559,1566,1573,1580,1587,1594,1601,1608],{"id":1560,"slug":1561,"title":1562,"excerpt":1563,"featuredImage":1564,"publishedAt":1565},"481","generative-ai-explained-models-retrieval-tools-and-applications-are-not-the-same-thing","L'IA generativa spiegata: modelli, recupero, strumenti e applicazioni non sono la stessa cosa","L'IA generativa è più di un modello. Scopri come modelli, recupero, strumenti, contesto, runtime e applicazioni si integrano nei sistemi di IA in produzione.","\u002Fuploads\u002F2026\u002F10\u002Fgenerative-ai-explained-models-retrieval-tools-and-applications-are-not-the-same-thing-1791475411822-pp0dvz.webp","2026-10-08T12:00:00.000Z",{"id":1567,"slug":1568,"title":1569,"excerpt":1570,"featuredImage":1571,"publishedAt":1572},"381","enterprise-grade-multi-tenant-architecture-for-an-international-platform","Architettura Multi-Tenant di Livello Enterprise per una Piattaforma Internazionale","Loving Rocks è una piattaforma per matrimoni di livello enterprise progettata con una vera architettura multi-tenant, database isolati per tenant e internazionalizzazione integrata per scalabilità globale, sicurezza e stabilità operativa a lungo termine.","\u002Fuploads\u002F2026\u002F01\u002Fenterprise-grade-multi-tenant-architecture-for-an-international-platform-1769789121298-b6v7ak.webp","2026-01-30T12:04:00.000Z",{"id":1574,"slug":1575,"title":1576,"excerpt":1577,"featuredImage":1578,"publishedAt":1579},"494","air-gapped-ai-how-ai-systems-work-without-internet-or-cloud-access","AI Air-Gapped: Come funzionano i sistemi di IA senza Internet o accesso al cloud","L'IA air-gapped esegue modelli, RAG e applicazioni di IA all'interno di un dominio di sicurezza isolato senza dipendenze da internet o dal cloud. Scopri come modelli, dati, aggiornamenti e strumenti operano offline.","\u002Fuploads\u002F2026\u002F10\u002Fair-gapped-ai-how-ai-systems-work-without-internet-or-cloud-access-1791487983978-e6xqf0.webp","2026-10-08T11:32:00.000Z",{"id":1581,"slug":1582,"title":1583,"excerpt":1584,"featuredImage":1585,"publishedAt":1586},"495","sovereign-ai-control-of-models-data-infrastructure-and-dependencies","IA sovrana: controllo di modelli, dati, infrastrutture e dipendenze","L'IA sovrana riguarda il controllo effettivo su modelli, dati, infrastrutture, software, operazioni e dipendenze strategiche — non semplicemente dove è ospitato un modello di IA.","\u002Fuploads\u002F2026\u002F10\u002Fsovereign-ai-control-of-models-data-infrastructure-and-dependencies-1791488833132-niy85x.webp","2026-10-08T15:45:00.000Z",{"id":1588,"slug":1589,"title":1590,"excerpt":1591,"featuredImage":1592,"publishedAt":1593},"491","ai-governance-models-data-permissions-risk-and-auditability","Governance dell'IA: modelli, dati, autorizzazioni, rischio e verificabilità","La governance dell'IA definisce chi può approvare, gestire, modificare e verificare i sistemi di IA attraverso modelli, fornitori, dati, autorizzazioni, rischi, valutazione e l'intero ciclo di vita.","\u002Fuploads\u002F2026\u002F10\u002Fai-governance-models-data-permissions-risk-and-auditability-1791485901301-g60xyu.webp","2026-10-08T12:56:00.000Z",{"id":1595,"slug":1596,"title":1597,"excerpt":1598,"featuredImage":1599,"publishedAt":1600},"482","adr-vs-nfr-architecture-decisions-and-system-quality-are-not-the-same-thing","ADR vs NFR: decisioni architetturali e qualità del sistema non sono la stessa cosa","ADR vs NFR spiegati: scopri come i requisiti di qualità del sistema guidano le decisioni architetturali, come gli ADR registrano i compromessi e perché la validazione rimane separata.","\u002Fuploads\u002F2026\u002F10\u002Fadr-vs-nfr-architecture-decisions-and-system-quality-are-not-the-same-thing-1791475921511-6zgen1.webp","2026-10-08T12:11:00.000Z",{"id":1602,"slug":1603,"title":1604,"excerpt":1605,"featuredImage":1606,"publishedAt":1607},"492","mcp-explained-what-it-connects-what-it-does-not-do-and-where-it-fits","MCP spiegato: cosa collega, cosa non fa e dove si colloca","Il Model Context Protocol collega le applicazioni di intelligenza artificiale a strumenti, risorse e prompt esterni attraverso un confine standard client-server. Scopri cosa fa MCP, cosa non fa e dove si colloca nell'architettura degli agenti.","\u002Fuploads\u002F2026\u002F10\u002Fmcp-explained-what-it-connects-what-it-does-not-do-and-where-it-fits-1791486640275-7ub1cq.webp","2026-10-08T15:09:00.000Z",{"id":1609,"slug":1610,"title":1611,"excerpt":1612,"featuredImage":1613,"publishedAt":1614},"498","can-evidence-survive-the-ai-pipeline-the-evidence-preservation-chain","Può la prova sopravvivere alla pipeline dell'IA? — La catena di conservazione delle prove","Le prove possono essere corrette alla fonte e diventare comunque inaffidabili dopo l'analisi, la normalizzazione, la suddivisione in blocchi, il recupero e l'assemblaggio del contesto. La Catena di Preservazione delle Prove tiene traccia se identità, provenienza, ambito e significato sopravvivono all'intera pipeline di IA.","\u002Fuploads\u002F2026\u002F10\u002Fcan-evidence-survive-the-ai-pipeline-the-evidence-preservation-chain-1792096396174-urdg9z.webp","2026-10-15T16:31:00.000Z","fallback",[],[]]