[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"portal-settings:stajic:fr":3,"public-menus:all":38,"post:how-far-is-an-ai-claim-from-the-evidence-evidence-distance:fr":221,"related:post:how-far-is-an-ai-claim-from-the-evidence-evidence-distance:fr:1":1554},{"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","fr","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":1553},{"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","À quelle distance se trouve une affirmation d’IA des preuves ? — Distance aux preuves","how-far-is-an-ai-claim-from-the-evidence-evidence-distance","{\"time\":1792096967668,\"blocks\":[{\"id\":\"U-5E-c-b7e\",\"type\":\"header\",\"data\":{\"text\":\"Question\",\"level\":2},\"tunes\":{}},{\"id\":\"FRol8Nr8db\",\"type\":\"paragraph\",\"data\":{\"text\":\"Deux réponses d'IA peuvent toutes deux citer des sources et différer néanmoins grandement dans la manière dont ces sources étayent directement les affirmations formulées. Comment décrire cette différence ?\"},\"tunes\":{}},{\"id\":\"3RFUF9JBWv\",\"type\":\"paragraph\",\"data\":{\"text\":\"Cet article utilise le terme Distance de preuve pour désigner la distance entre une affirmation et les preuves directement inspectables qui la soutiennent.\"},\"tunes\":{}},{\"id\":\"K2DtqExn8Q\",\"type\":\"quote\",\"data\":{\"text\":\"La Distance de preuve est le nombre et la force des transformations, hypothèses et étapes inférentielles qui séparent une affirmation des preuves pouvant la vérifier directement.\",\"caption\":\"Définition de travail — un concept de raisonnement introduit ici, et non une métrique standardisée.\",\"alignment\":\"left\"},\"tunes\":{}},{\"id\":\"mP_9dls1Mj\",\"type\":\"tableOfContents\",\"data\":{\"title\":\"Contents\",\"minLevel\":2,\"maxLevel\":3},\"tunes\":{}},{\"id\":\"LZ8gF5eBXn\",\"type\":\"header\",\"data\":{\"text\":\"Ce que cela signifie vraiment\",\"level\":2},\"tunes\":{}},{\"id\":\"UrcbcNY4cd\",\"type\":\"paragraph\",\"data\":{\"text\":\"Une citation prouve qu'une source a été référencée. Elle ne prouve pas automatiquement que la source soutient directement la phrase située à côté de la citation.\"},\"tunes\":{}},{\"id\":\"_uiRHanfHm\",\"type\":\"paragraph\",\"data\":{\"text\":\"Entre une observation originale et une réponse finale d'IA, il peut y avoir extraction, analyse syntaxique, découpage, récupération, résumé, interprétation, agrégation et inférence.\"},\"tunes\":{}},{\"id\":\"QjSOFgTbYt\",\"type\":\"code\",\"data\":{\"code\":\"Primary evidence\\n   ↓\\nExtracted data\\n   ↓\\nRetrieved passage\\n   ↓\\nSummary\\n   ↓\\nInterpretation\\n   ↓\\nInference\\n   ↓\\nFinal claim\"},\"tunes\":{}},{\"id\":\"tyKk70azTz\",\"type\":\"paragraph\",\"data\":{\"text\":\"Chaque étape peut être légitime. Mais chaque étape peut aussi introduire une distorsion.\"},\"tunes\":{}},{\"id\":\"c8cUBsGP9u\",\"type\":\"header\",\"data\":{\"text\":\"Exemple le plus simple\",\"level\":2},\"tunes\":{}},{\"id\":\"8WEHk8YTRh\",\"type\":\"paragraph\",\"data\":{\"text\":\"Supposons qu'un capteur de température indique 38,2 °C.\"},\"tunes\":{}},{\"id\":\"Us80vMItcP\",\"type\":\"table\",\"data\":{\"withHeadings\":true,\"stretched\":false,\"content\":[[\"Affirmation\",\"Relation aux preuves\",\"Distance illustrative\"],[\"Le capteur indique 38,2 °C.\",\"Lecture directe\",\"Faible\"],[\"Le système tourne plus chaud qu'hier.\",\"Nécessite une comparaison avec la lecture d'hier\",\"Modérée\"],[\"Le système de refroidissement est en panne.\",\"Nécessite une inférence causale et des explications alternatives\",\"Élevée\"]]},\"tunes\":{}},{\"id\":\"tb7oLoEZKG\",\"type\":\"paragraph\",\"data\":{\"text\":\"La valeur originale du capteur soutient directement la première affirmation. Elle ne prouve pas à elle seule la troisième.\"},\"tunes\":{}},{\"id\":\"C-eR0_3B6z\",\"type\":\"code\",\"data\":{\"code\":\"Evidence exists\\n        ≠\\nEvidence directly supports every derived claim\"},\"tunes\":{}},{\"id\":\"UEzosoy4ro\",\"type\":\"header\",\"data\":{\"text\":\"Où l'exemple cesse de fonctionner\",\"level\":2},\"tunes\":{}},{\"id\":\"DOoK_hTK4X\",\"type\":\"paragraph\",\"data\":{\"text\":\"La Distance de preuve n'est pas simplement un décompte de phrases ou d'étapes de traitement.\"},\"tunes\":{}},{\"id\":\"ErwUwhFbBy\",\"type\":\"paragraph\",\"data\":{\"text\":\"Une transformation mathématiquement valide peut préserver les preuves presque parfaitement, tandis qu'une seule interprétation spéculative peut créer un saut épistémique bien plus grand.\"},\"tunes\":{}},{\"id\":\"sJzX6y5YMv\",\"type\":\"paragraph\",\"data\":{\"text\":\"Le concept inclut donc à la fois la distance par transformation et la distance par inférence.\"},\"tunes\":{}},{\"id\":\"UofGG__fx9\",\"type\":\"header\",\"data\":{\"text\":\"Réponse directe\",\"level\":2},\"tunes\":{}},{\"id\":\"TW_Em9qWkv\",\"type\":\"paragraph\",\"data\":{\"text\":\"La distance de preuve augmente lorsqu'une affirmation dépend de plus de transformations, de plus d'hypothèses, d'une provenance plus faible ou d'une inférence plus forte que ce que les preuves sous-jacentes justifient directement.\"},\"tunes\":{}},{\"id\":\"fIpOMoYVBB\",\"type\":\"code\",\"data\":{\"code\":\"Short distance:\\nsource → claim\\n\\nLonger distance:\\nsource → extraction → summary → assumption → inference → claim\"},\"tunes\":{}},{\"id\":\"oowoCkwwjx\",\"type\":\"paragraph\",\"data\":{\"text\":\"Une distance plus longue ne rend pas automatiquement une affirmation fausse. Cela signifie que la chaîne de raisonnement nécessite davantage de validation.\"},\"tunes\":{}},{\"id\":\"VCPnCKQdl0\",\"type\":\"header\",\"data\":{\"text\":\"Pourquoi cela importe pour l'IA\",\"level\":2},\"tunes\":{}},{\"id\":\"94myaIM3xs\",\"type\":\"paragraph\",\"data\":{\"text\":\"Les LLM sont particulièrement doués pour produire des ponts fluides entre des éléments d'information. Cette capacité est utile, mais la fluidité peut masquer à quel point la conclusion s'est éloignée des preuves directement récupérées.\"},\"tunes\":{}},{\"id\":\"-DV37S4Dr5\",\"type\":\"paragraph\",\"data\":{\"text\":\"Une réponse peut donc sembler bien sourcée tout en contenant des affirmations dont le soutien n'existe qu'à travers plusieurs étapes inférentielles non explicitées.\"},\"tunes\":{}},{\"id\":\"kKkwbPzcM3\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Ffr\u002Fblog\u002Ffrom-research-protocol-to-a-general-ai-reasoning-framework\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Du protocole de recherche à un cadre général de raisonnement pour l'IA\",\"description\":\"Un raisonnement fiable exige une séparation explicite entre les preuves, les hypothèses, les conjectures et les conclusions.\"}},\"tunes\":{}},{\"id\":\"rs-zSvojU_\",\"type\":\"header\",\"data\":{\"text\":\"Quatre composantes de la distance de preuve\",\"level\":2},\"tunes\":{}},{\"id\":\"FD1yhHdlVS\",\"type\":\"header\",\"data\":{\"text\":\"1. Distance de transformation\",\"level\":3},\"tunes\":{}},{\"id\":\"MIEYJoRSAO\",\"type\":\"paragraph\",\"data\":{\"text\":\"Combien de fois les preuves originales ont-elles été converties, résumées, normalisées ou reformatées avant d'atteindre le modèle ?\"},\"tunes\":{}},{\"id\":\"aWBnyLO25p\",\"type\":\"header\",\"data\":{\"text\":\"2. Distance inférentielle\",\"level\":3},\"tunes\":{}},{\"id\":\"t8-EW59loR\",\"type\":\"paragraph\",\"data\":{\"text\":\"Combien d'étapes de raisonnement sont nécessaires avant que les preuves ne soutiennent la conclusion finale ?\"},\"tunes\":{}},{\"id\":\"5Bc1midVS1\",\"type\":\"header\",\"data\":{\"text\":\"3. Distance de provenance\",\"level\":3},\"tunes\":{}},{\"id\":\"pKUrZ6_qy7\",\"type\":\"paragraph\",\"data\":{\"text\":\"L'énoncé récupéré peut-il encore être retracé jusqu'à une source originale identifiable, ou est-il passé par des résumés et des copies dont l'origine est incertaine ?\"},\"tunes\":{}},{\"id\":\"LoFigVOcrl\",\"type\":\"header\",\"data\":{\"text\":\"4. Distance d'applicabilité\",\"level\":3},\"tunes\":{}},{\"id\":\"ZlgawMn_4r\",\"type\":\"paragraph\",\"data\":{\"text\":\"Combien de raisonnement est nécessaire pour établir que les preuves s'appliquent à cette version exacte, à cet utilisateur, à cette juridiction, à cette période ou à cette situation ?\"},\"tunes\":{}},{\"id\":\"nHeOEmCdVh\",\"type\":\"header\",\"data\":{\"text\":\"La distance de preuve dans le RAG\",\"level\":2},\"tunes\":{}},{\"id\":\"8abRGgv8hH\",\"type\":\"paragraph\",\"data\":{\"text\":\"Le RAG peut raccourcir une distance importante : la distance entre un modèle et les connaissances externes.\"},\"tunes\":{}},{\"id\":\"6yz0sMBlWb\",\"type\":\"paragraph\",\"data\":{\"text\":\"Il ne raccourcit pas automatiquement la distance entre un passage récupéré et l'affirmation finale.\"},\"tunes\":{}},{\"id\":\"42VR-3_5jt\",\"type\":\"code\",\"data\":{\"code\":\"RAG solves:\\nmodel ↔ external information\\n\\nRAG does not automatically solve:\\nevidence ↔ justified conclusion\"},\"tunes\":{}},{\"id\":\"1A4cY9J9jY\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Ffr\u002Fblog\u002Fwhere-does-an-llm-get-its-data-rag-data-sources-in-python\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Où un LLM obtient-il ses données ? Sources de données RAG en Python\",\"description\":\"La récupération apporte des informations externes dans le contexte. La distance de preuve demande ce qui se passe entre ces informations et l'affirmation finalement produite.\"}},\"tunes\":{}},{\"id\":\"QQueuXodER\",\"type\":\"header\",\"data\":{\"text\":\"Lien avec le déclencheur de récupération\",\"level\":2},\"tunes\":{}},{\"id\":\"nk5TrUeNNX\",\"type\":\"paragraph\",\"data\":{\"text\":\"Le déclencheur de récupération demande si les connaissances internes du modèle sont suffisantes.\"},\"tunes\":{}},{\"id\":\"QnB7VjCfv5\",\"type\":\"paragraph\",\"data\":{\"text\":\"La distance de preuve pose une question ultérieure : une fois les preuves récupérées, dans quelle mesure soutiennent-elles directement ce que le système veut affirmer ?\"},\"tunes\":{}},{\"id\":\"noRfYy8wMl\",\"type\":\"code\",\"data\":{\"code\":\"Need evidence?\\n   ↓\\nRetrieval Trigger\\n   ↓\\nRetrieve\\n   ↓\\nHow directly does evidence support the claim?\\n   ↓\\nEvidence Distance\"},\"tunes\":{}},{\"id\":\"piD0jF1B6J\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Ffr\u002Fblog\u002Fwhen-should-an-ai-stop-trusting-its-own-knowledge-the-retrieval-trigger\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Quand une IA doit-elle cesser de faire confiance à ses propres connaissances ? — Le déclencheur de récupération\",\"description\":\"Le déclencheur décide quand des preuves sont nécessaires. La distance de preuve évalue dans quelle mesure l'affirmation finale s'éloigne de ces preuves.\"}},\"tunes\":{}},{\"id\":\"GFydqD6ip4\",\"type\":\"header\",\"data\":{\"text\":\"Un diagnostic pratique\",\"level\":2},\"tunes\":{}},{\"id\":\"FZ9UC4yBV7\",\"type\":\"paragraph\",\"data\":{\"text\":\"Pour une affirmation importante générée par IA, reconstruisez la chaîne de soutien à rebours.\"},\"tunes\":{}},{\"id\":\"YaajwIDBtn\",\"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\":\"pG_Nt5kmIx\",\"type\":\"paragraph\",\"data\":{\"text\":\"Si la chaîne se rompt, le système a rencontré un problème de traçabilité des preuves.\"},\"tunes\":{}},{\"id\":\"8iAFVihOfK\",\"type\":\"paragraph\",\"data\":{\"text\":\"Si la chaîne reste intacte mais nécessite plusieurs hypothèses fortes, la distance de preuve reste élevée.\"},\"tunes\":{}},{\"id\":\"Fx1qSCq_7Y\",\"type\":\"header\",\"data\":{\"text\":\"Exemple : données produit\",\"level\":2},\"tunes\":{}},{\"id\":\"YGRwsd45ed\",\"type\":\"paragraph\",\"data\":{\"text\":\"Supposons qu'une spécification de fabricant indique qu'un ordinateur portable pèse 1,48 kg.\"},\"tunes\":{}},{\"id\":\"-DXLwSAT5Q\",\"type\":\"code\",\"data\":{\"code\":\"Manufacturer specification\\n→ 1.48 kg\\n→ \\\"The laptop weighs 1.48 kg\\\"\\n\\nLow Evidence Distance\"},\"tunes\":{}},{\"id\":\"4lAeLZC-XI\",\"type\":\"paragraph\",\"data\":{\"text\":\"Considérons maintenant :\"},\"tunes\":{}},{\"id\":\"pPoYPaJVfl\",\"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\":\"1xahoQfBjz\",\"type\":\"paragraph\",\"data\":{\"text\":\"La spécification est réelle. La recommandation est encore à plusieurs étapes de raisonnement de celle-ci.\"},\"tunes\":{}},{\"id\":\"8Tv9MutAra\",\"type\":\"header\",\"data\":{\"text\":\"Exemple : Preuve scientifique\",\"level\":2},\"tunes\":{}},{\"id\":\"80fVFG8bKv\",\"type\":\"paragraph\",\"data\":{\"text\":\"Un article peut rapporter une association observée. Une réponse d'IA peut incorrectement la compresser en une déclaration causale.\"},\"tunes\":{}},{\"id\":\"ei64yymEDG\",\"type\":\"code\",\"data\":{\"code\":\"Observed association\\n        ↓\\nInterpretation\\n        ↓\\nUnstated causal assumption\\n        ↓\\n\\\"X causes Y\\\"\"},\"tunes\":{}},{\"id\":\"xelk4Y9tBM\",\"type\":\"paragraph\",\"data\":{\"text\":\"Le problème n'est pas qu'aucune preuve n'existe. Le problème est que la déclaration finale franchit une distance probante plus grande que ce que l'étude soutient directement.\"},\"tunes\":{}},{\"id\":\"NgfumURDIl\",\"type\":\"header\",\"data\":{\"text\":\"Connexion au contexte prêt pour les agents\",\"level\":2},\"tunes\":{}},{\"id\":\"qDryGUUuXn\",\"type\":\"paragraph\",\"data\":{\"text\":\"Les objets d'information structurés peuvent réduire la distance probante évitable en préservant l'identité de la source, les unités, les horodatages, les conflits et la validité au lieu de forcer le modèle à les reconstruire à partir d'un texte aplati.\"},\"tunes\":{}},{\"id\":\"6ayRZdBizj\",\"type\":\"linkTool\",\"data\":{\"link\":\"https:\u002F\u002Fstajic.de\u002Ffr\u002Fportfolio\u002Ffrom-raw-data-to-verifiable-ai-context-a-practical-architecture-for-agent-ready-information\",\"meta\":{\"image\":{\"url\":\"\"},\"title\":\"Des données brutes au contexte IA vérifiable : une architecture pratique pour une information prête pour les agents\",\"description\":\"Préserver la provenance et la validité avant la récupération réduit le nombre d'étapes de reconstruction non étayées nécessaires ultérieurement.\"}},\"tunes\":{}},{\"id\":\"aegCMRv-xN\",\"type\":\"header\",\"data\":{\"text\":\"Une échelle de travail simple\",\"level\":2},\"tunes\":{}},{\"id\":\"1ctmp1RO_6\",\"type\":\"table\",\"data\":{\"withHeadings\":true,\"stretched\":false,\"content\":[[\"Niveau\",\"Description\"],[\"ED0\",\"Fait source directement rapporté\"],[\"ED1\",\"Transformation déterministe simple ou comparaison directe\"],[\"ED2\",\"Une étape d'interprétation significative\"],[\"ED3\",\"Hypothèses multiples ou sources combinées\"],[\"ED4\",\"Conclusion fortement causale, prédictive ou prescriptive au-delà des preuves directes\"]]},\"tunes\":{}},{\"id\":\"5nGMsGnY2n\",\"type\":\"paragraph\",\"data\":{\"text\":\"Cette échelle est illustrative, pas une norme scientifique. Son but est de forcer une question explicite : Combien avons-nous dû ajouter entre la preuve et l'affirmation ?\"},\"tunes\":{}},{\"id\":\"4K08dsmvIE\",\"type\":\"header\",\"data\":{\"text\":\"Idées fausses\",\"level\":2},\"tunes\":{}},{\"id\":\"uGEJNyXcv3\",\"type\":\"paragraph\",\"data\":{\"text\":\"Une citation signifie que la distance probante est faible. Non. Une citation peut se trouver à côté d'une affirmation qui dépasse substantiellement ce que la source citée établit.\"},\"tunes\":{}},{\"id\":\"ysL8AGwcVK\",\"type\":\"paragraph\",\"data\":{\"text\":\"Une longue distance probante signifie que l'affirmation est fausse. Non. Des conclusions complexes peuvent être bien étayées, mais elles nécessitent une chaîne de raisonnement plus solide.\"},\"tunes\":{}},{\"id\":\"S7ybJ-eBFz\",\"type\":\"paragraph\",\"data\":{\"text\":\"Les sources primaires éliminent l'inférence. Non. Les preuves primaires peuvent encore nécessiter une interprétation.\"},\"tunes\":{}},{\"id\":\"z8t-Sl3iqs\",\"type\":\"paragraph\",\"data\":{\"text\":\"Plus de sources réduisent automatiquement la distance. Dix faibles répétitions de la même affirmation non étayée ne créent pas de preuve directe.\"},\"tunes\":{}},{\"id\":\"RDzGjjxuuc\",\"type\":\"header\",\"data\":{\"text\":\"Cas limite : Données dérivées\",\"level\":2},\"tunes\":{}},{\"id\":\"eItaEDKK3R\",\"type\":\"paragraph\",\"data\":{\"text\":\"Certaines valeurs dérivées sont plus fiables que les descriptions humaines brutes. Un total correctement calculé à partir de transactions vérifiées peut être hautement fiable, même s'il ne résulte que d'une seule transformation à partir des enregistrements originaux.\"},\"tunes\":{}},{\"id\":\"vBdc8hymZy\",\"type\":\"paragraph\",\"data\":{\"text\":\"La Distance aux Preuves mesure donc la séparation épistémique, et non une préférence simpliste pour les données non modifiées.\"},\"tunes\":{}},{\"id\":\"RXQXotzXRw\",\"type\":\"header\",\"data\":{\"text\":\"Limites\",\"level\":2},\"tunes\":{}},{\"id\":\"JjvwlkIJOr\",\"type\":\"paragraph\",\"data\":{\"text\":\"La Distance aux Preuves est un outil de diagnostic conceptuel, et non un score quantitatif universel.\"},\"tunes\":{}},{\"id\":\"HFJlUDTxF6\",\"type\":\"paragraph\",\"data\":{\"text\":\"Différents domaines nécessitent un traitement différent de l'inférence. La preuve mathématique, les preuves cliniques, les journaux logiciels, l'analyse de marché et l'interprétation historique ne peuvent pas partager un seuil numérique universel unique.\"},\"tunes\":{}},{\"id\":\"XGjXl2Z8WW\",\"type\":\"header\",\"data\":{\"text\":\"Qu'est-ce qui changerait cette réponse ?\",\"level\":2},\"tunes\":{}},{\"id\":\"SlqRtFaMg3\",\"type\":\"paragraph\",\"data\":{\"text\":\"Les futurs systèmes d'IA pourraient générer des graphes de provenance explicites, des liens affirmation-preuve vérifiables par machine et des traces de raisonnement typées. Cela pourrait rendre certaines parties de la Distance aux Preuves mesurables automatiquement.\"},\"tunes\":{}},{\"id\":\"RNgkjK7ENz\",\"type\":\"paragraph\",\"data\":{\"text\":\"Le problème sous-jacent demeure : une affirmation peut s'éloigner de ses preuves même si chaque phrase intermédiaire semble plausible.\"},\"tunes\":{}},{\"id\":\"gCzER2dcxH\",\"type\":\"header\",\"data\":{\"text\":\"Conclusion\",\"level\":2},\"tunes\":{}},{\"id\":\"TsiQdQEYrr\",\"type\":\"paragraph\",\"data\":{\"text\":\"La distinction importante ne se limite pas à celle entre les réponses sourcées et non sourcées.\"},\"tunes\":{}},{\"id\":\"ou4HQModNd\",\"type\":\"paragraph\",\"data\":{\"text\":\"Elle se situe entre les affirmations qui restent proches de preuves directement inspectables et celles qui dépendent de chaînes d'interprétation et d'hypothèses de plus en plus longues.\"},\"tunes\":{}},{\"id\":\"8X-yxETmWe\",\"type\":\"code\",\"data\":{\"code\":\"Evidence\\n   ↓\\nEvidence Distance\\n   ↓\\nReasoning\\n   ↓\\nAnswer Validity Boundary\"},\"tunes\":{}},{\"id\":\"vEy-4dZBXE\",\"type\":\"paragraph\",\"data\":{\"text\":\"La Distance aux Preuves nous donne un langage pour décrire cet espace.\"},\"tunes\":{}},{\"id\":\"NaB35Wz01G\",\"type\":\"header\",\"data\":{\"text\":\"Sources primaires\",\"level\":2},\"tunes\":{}},{\"id\":\"H25HPflL_j\",\"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\":\"ray0tEnwZO\",\"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\":\"uXwmbDFVoi\",\"type\":\"paragraph\",\"data\":{\"text\":\"W3C, \u003Ca href=\\\"https:\u002F\u002Fwww.w3.org\u002FTR\u002Fprov-overview\u002F\\\" target=\\\"_blank\\\">PROV Overview\u003C\u002Fa>. Un modèle standardisé pour représenter les relations de provenance.\"},\"tunes\":{}}],\"version\":\"2.31.6\"}",{"time":228,"blocks":229,"version":741},1792096967668,[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},"U-5E-c-b7e",{"text":233,"level":234},"Question",2,{},{"id":237,"data":238,"type":240,"tunes":241},"FRol8Nr8db",{"text":239},"Deux réponses d'IA peuvent toutes deux citer des sources et différer néanmoins grandement dans la manière dont ces sources étayent directement les affirmations formulées. Comment décrire cette différence ?","paragraph",{},{"id":243,"data":244,"type":240,"tunes":246},"3RFUF9JBWv",{"text":245},"Cet article utilise le terme Distance de preuve pour désigner la distance entre une affirmation et les preuves directement inspectables qui la soutiennent.",{},{"id":248,"data":249,"type":253,"tunes":254},"K2DtqExn8Q",{"text":250,"caption":251,"alignment":252},"La Distance de preuve est le nombre et la force des transformations, hypothèses et étapes inférentielles qui séparent une affirmation des preuves pouvant la vérifier directement.","Définition de travail — un concept de raisonnement introduit ici, et non une métrique standardisée.","left","quote",{},{"id":256,"data":257,"type":260,"tunes":261},"mP_9dls1Mj",{"title":258,"maxLevel":259,"minLevel":234},"Contents",3,"tableOfContents",{},{"id":263,"data":264,"type":42,"tunes":266},"LZ8gF5eBXn",{"text":265,"level":234},"Ce que cela signifie vraiment",{},{"id":268,"data":269,"type":240,"tunes":271},"UrcbcNY4cd",{"text":270},"Une citation prouve qu'une source a été référencée. Elle ne prouve pas automatiquement que la source soutient directement la phrase située à côté de la citation.",{},{"id":273,"data":274,"type":240,"tunes":276},"_uiRHanfHm",{"text":275},"Entre une observation originale et une réponse finale d'IA, il peut y avoir extraction, analyse syntaxique, découpage, récupération, résumé, interprétation, agrégation et inférence.",{},{"id":278,"data":279,"type":281,"tunes":282},"QjSOFgTbYt",{"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},"tyKk70azTz",{"text":286},"Chaque étape peut être légitime. Mais chaque étape peut aussi introduire une distorsion.",{},{"id":289,"data":290,"type":42,"tunes":292},"c8cUBsGP9u",{"text":291,"level":234},"Exemple le plus simple",{},{"id":294,"data":295,"type":240,"tunes":297},"8WEHk8YTRh",{"text":296},"Supposons qu'un capteur de température indique 38,2 °C.",{},{"id":299,"data":300,"type":318,"tunes":319},"Us80vMItcP",{"content":301,"stretched":43,"withHeadings":14},[302,306,310,314],[303,304,305],"Affirmation","Relation aux preuves","Distance illustrative",[307,308,309],"Le capteur indique 38,2 °C.","Lecture directe","Faible",[311,312,313],"Le système tourne plus chaud qu'hier.","Nécessite une comparaison avec la lecture d'hier","Modérée",[315,316,317],"Le système de refroidissement est en panne.","Nécessite une inférence causale et des explications alternatives","Élevée","table",{},{"id":321,"data":322,"type":240,"tunes":324},"tb7oLoEZKG",{"text":323},"La valeur originale du capteur soutient directement la première affirmation. Elle ne prouve pas à elle seule la troisième.",{},{"id":326,"data":327,"type":281,"tunes":329},"C-eR0_3B6z",{"code":328},"Evidence exists\n        ≠\nEvidence directly supports every derived claim",{},{"id":331,"data":332,"type":42,"tunes":334},"UEzosoy4ro",{"text":333,"level":234},"Où l'exemple cesse de fonctionner",{},{"id":336,"data":337,"type":240,"tunes":339},"DOoK_hTK4X",{"text":338},"La Distance de preuve n'est pas simplement un décompte de phrases ou d'étapes de traitement.",{},{"id":341,"data":342,"type":240,"tunes":344},"ErwUwhFbBy",{"text":343},"Une transformation mathématiquement valide peut préserver les preuves presque parfaitement, tandis qu'une seule interprétation spéculative peut créer un saut épistémique bien plus grand.",{},{"id":346,"data":347,"type":240,"tunes":349},"sJzX6y5YMv",{"text":348},"Le concept inclut donc à la fois la distance par transformation et la distance par inférence.",{},{"id":351,"data":352,"type":42,"tunes":354},"UofGG__fx9",{"text":353,"level":234},"Réponse directe",{},{"id":356,"data":357,"type":240,"tunes":359},"TW_Em9qWkv",{"text":358},"La distance de preuve augmente lorsqu'une affirmation dépend de plus de transformations, de plus d'hypothèses, d'une provenance plus faible ou d'une inférence plus forte que ce que les preuves sous-jacentes justifient directement.",{},{"id":361,"data":362,"type":281,"tunes":364},"fIpOMoYVBB",{"code":363},"Short distance:\nsource → claim\n\nLonger distance:\nsource → extraction → summary → assumption → inference → claim",{},{"id":366,"data":367,"type":240,"tunes":369},"oowoCkwwjx",{"text":368},"Une distance plus longue ne rend pas automatiquement une affirmation fausse. Cela signifie que la chaîne de raisonnement nécessite davantage de validation.",{},{"id":371,"data":372,"type":42,"tunes":374},"VCPnCKQdl0",{"text":373,"level":234},"Pourquoi cela importe pour l'IA",{},{"id":376,"data":377,"type":240,"tunes":379},"94myaIM3xs",{"text":378},"Les LLM sont particulièrement doués pour produire des ponts fluides entre des éléments d'information. Cette capacité est utile, mais la fluidité peut masquer à quel point la conclusion s'est éloignée des preuves directement récupérées.",{},{"id":381,"data":382,"type":240,"tunes":384},"-DV37S4Dr5",{"text":383},"Une réponse peut donc sembler bien sourcée tout en contenant des affirmations dont le soutien n'existe qu'à travers plusieurs étapes inférentielles non explicitées.",{},{"id":386,"data":387,"type":394,"tunes":395},"kKkwbPzcM3",{"link":388,"meta":389},"https:\u002F\u002Fstajic.de\u002Ffr\u002Fblog\u002Ffrom-research-protocol-to-a-general-ai-reasoning-framework",{"image":390,"title":392,"description":393},{"url":391},"","Du protocole de recherche à un cadre général de raisonnement pour l'IA","Un raisonnement fiable exige une séparation explicite entre les preuves, les hypothèses, les conjectures et les conclusions.","linkTool",{},{"id":397,"data":398,"type":42,"tunes":400},"rs-zSvojU_",{"text":399,"level":234},"Quatre composantes de la distance de preuve",{},{"id":402,"data":403,"type":42,"tunes":405},"FD1yhHdlVS",{"text":404,"level":259},"1. Distance de transformation",{},{"id":407,"data":408,"type":240,"tunes":410},"MIEYJoRSAO",{"text":409},"Combien de fois les preuves originales ont-elles été converties, résumées, normalisées ou reformatées avant d'atteindre le modèle ?",{},{"id":412,"data":413,"type":42,"tunes":415},"aWBnyLO25p",{"text":414,"level":259},"2. Distance inférentielle",{},{"id":417,"data":418,"type":240,"tunes":420},"t8-EW59loR",{"text":419},"Combien d'étapes de raisonnement sont nécessaires avant que les preuves ne soutiennent la conclusion finale ?",{},{"id":422,"data":423,"type":42,"tunes":425},"5Bc1midVS1",{"text":424,"level":259},"3. Distance de provenance",{},{"id":427,"data":428,"type":240,"tunes":430},"pKUrZ6_qy7",{"text":429},"L'énoncé récupéré peut-il encore être retracé jusqu'à une source originale identifiable, ou est-il passé par des résumés et des copies dont l'origine est incertaine ?",{},{"id":432,"data":433,"type":42,"tunes":435},"LoFigVOcrl",{"text":434,"level":259},"4. Distance d'applicabilité",{},{"id":437,"data":438,"type":240,"tunes":440},"ZlgawMn_4r",{"text":439},"Combien de raisonnement est nécessaire pour établir que les preuves s'appliquent à cette version exacte, à cet utilisateur, à cette juridiction, à cette période ou à cette situation ?",{},{"id":442,"data":443,"type":42,"tunes":445},"nHeOEmCdVh",{"text":444,"level":234},"La distance de preuve dans le RAG",{},{"id":447,"data":448,"type":240,"tunes":450},"8abRGgv8hH",{"text":449},"Le RAG peut raccourcir une distance importante : la distance entre un modèle et les connaissances externes.",{},{"id":452,"data":453,"type":240,"tunes":455},"6yz0sMBlWb",{"text":454},"Il ne raccourcit pas automatiquement la distance entre un passage récupéré et l'affirmation finale.",{},{"id":457,"data":458,"type":281,"tunes":460},"42VR-3_5jt",{"code":459},"RAG solves:\nmodel ↔ external information\n\nRAG does not automatically solve:\nevidence ↔ justified conclusion",{},{"id":462,"data":463,"type":394,"tunes":469},"1A4cY9J9jY",{"link":464,"meta":465},"https:\u002F\u002Fstajic.de\u002Ffr\u002Fblog\u002Fwhere-does-an-llm-get-its-data-rag-data-sources-in-python",{"image":466,"title":467,"description":468},{"url":391},"Où un LLM obtient-il ses données ? Sources de données RAG en Python","La récupération apporte des informations externes dans le contexte. La distance de preuve demande ce qui se passe entre ces informations et l'affirmation finalement produite.",{},{"id":471,"data":472,"type":42,"tunes":474},"QQueuXodER",{"text":473,"level":234},"Lien avec le déclencheur de récupération",{},{"id":476,"data":477,"type":240,"tunes":479},"nk5TrUeNNX",{"text":478},"Le déclencheur de récupération demande si les connaissances internes du modèle sont suffisantes.",{},{"id":481,"data":482,"type":240,"tunes":484},"QnB7VjCfv5",{"text":483},"La distance de preuve pose une question ultérieure : une fois les preuves récupérées, dans quelle mesure soutiennent-elles directement ce que le système veut affirmer ?",{},{"id":486,"data":487,"type":281,"tunes":489},"noRfYy8wMl",{"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},"piD0jF1B6J",{"link":493,"meta":494},"https:\u002F\u002Fstajic.de\u002Ffr\u002Fblog\u002Fwhen-should-an-ai-stop-trusting-its-own-knowledge-the-retrieval-trigger",{"image":495,"title":496,"description":497},{"url":391},"Quand une IA doit-elle cesser de faire confiance à ses propres connaissances ? — Le déclencheur de récupération","Le déclencheur décide quand des preuves sont nécessaires. La distance de preuve évalue dans quelle mesure l'affirmation finale s'éloigne de ces preuves.",{},{"id":500,"data":501,"type":42,"tunes":503},"GFydqD6ip4",{"text":502,"level":234},"Un diagnostic pratique",{},{"id":505,"data":506,"type":240,"tunes":508},"FZ9UC4yBV7",{"text":507},"Pour une affirmation importante générée par IA, reconstruisez la chaîne de soutien à rebours.",{},{"id":510,"data":511,"type":281,"tunes":513},"YaajwIDBtn",{"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},"pG_Nt5kmIx",{"text":517},"Si la chaîne se rompt, le système a rencontré un problème de traçabilité des preuves.",{},{"id":520,"data":521,"type":240,"tunes":523},"8iAFVihOfK",{"text":522},"Si la chaîne reste intacte mais nécessite plusieurs hypothèses fortes, la distance de preuve reste élevée.",{},{"id":525,"data":526,"type":42,"tunes":528},"Fx1qSCq_7Y",{"text":527,"level":234},"Exemple : données produit",{},{"id":530,"data":531,"type":240,"tunes":533},"YGRwsd45ed",{"text":532},"Supposons qu'une spécification de fabricant indique qu'un ordinateur portable pèse 1,48 kg.",{},{"id":535,"data":536,"type":281,"tunes":538},"-DXLwSAT5Q",{"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},"4lAeLZC-XI",{"text":542},"Considérons maintenant :",{},{"id":545,"data":546,"type":281,"tunes":548},"pPoYPaJVfl",{"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},"1xahoQfBjz",{"text":552},"La spécification est réelle. La recommandation est encore à plusieurs étapes de raisonnement de celle-ci.",{},{"id":555,"data":556,"type":42,"tunes":558},"8Tv9MutAra",{"text":557,"level":234},"Exemple : Preuve scientifique",{},{"id":560,"data":561,"type":240,"tunes":563},"80fVFG8bKv",{"text":562},"Un article peut rapporter une association observée. Une réponse d'IA peut incorrectement la compresser en une déclaration causale.",{},{"id":565,"data":566,"type":281,"tunes":568},"ei64yymEDG",{"code":567},"Observed association\n        ↓\nInterpretation\n        ↓\nUnstated causal assumption\n        ↓\n\"X causes Y\"",{},{"id":570,"data":571,"type":240,"tunes":573},"xelk4Y9tBM",{"text":572},"Le problème n'est pas qu'aucune preuve n'existe. Le problème est que la déclaration finale franchit une distance probante plus grande que ce que l'étude soutient directement.",{},{"id":575,"data":576,"type":42,"tunes":578},"NgfumURDIl",{"text":577,"level":234},"Connexion au contexte prêt pour les agents",{},{"id":580,"data":581,"type":240,"tunes":583},"qDryGUUuXn",{"text":582},"Les objets d'information structurés peuvent réduire la distance probante évitable en préservant l'identité de la source, les unités, les horodatages, les conflits et la validité au lieu de forcer le modèle à les reconstruire à partir d'un texte aplati.",{},{"id":585,"data":586,"type":394,"tunes":592},"6ayRZdBizj",{"link":587,"meta":588},"https:\u002F\u002Fstajic.de\u002Ffr\u002Fportfolio\u002Ffrom-raw-data-to-verifiable-ai-context-a-practical-architecture-for-agent-ready-information",{"image":589,"title":590,"description":591},{"url":391},"Des données brutes au contexte IA vérifiable : une architecture pratique pour une information prête pour les agents","Préserver la provenance et la validité avant la récupération réduit le nombre d'étapes de reconstruction non étayées nécessaires ultérieurement.",{},{"id":594,"data":595,"type":42,"tunes":597},"aegCMRv-xN",{"text":596,"level":234},"Une échelle de travail simple",{},{"id":599,"data":600,"type":318,"tunes":620},"1ctmp1RO_6",{"content":601,"stretched":43,"withHeadings":14},[602,605,608,611,614,617],[603,604],"Niveau","Description",[606,607],"ED0","Fait source directement rapporté",[609,610],"ED1","Transformation déterministe simple ou comparaison directe",[612,613],"ED2","Une étape d'interprétation significative",[615,616],"ED3","Hypothèses multiples ou sources combinées",[618,619],"ED4","Conclusion fortement causale, prédictive ou prescriptive au-delà des preuves directes",{},{"id":622,"data":623,"type":240,"tunes":625},"5nGMsGnY2n",{"text":624},"Cette échelle est illustrative, pas une norme scientifique. Son but est de forcer une question explicite : Combien avons-nous dû ajouter entre la preuve et l'affirmation ?",{},{"id":627,"data":628,"type":42,"tunes":630},"4K08dsmvIE",{"text":629,"level":234},"Idées fausses",{},{"id":632,"data":633,"type":240,"tunes":635},"uGEJNyXcv3",{"text":634},"Une citation signifie que la distance probante est faible. Non. Une citation peut se trouver à côté d'une affirmation qui dépasse substantiellement ce que la source citée établit.",{},{"id":637,"data":638,"type":240,"tunes":640},"ysL8AGwcVK",{"text":639},"Une longue distance probante signifie que l'affirmation est fausse. Non. Des conclusions complexes peuvent être bien étayées, mais elles nécessitent une chaîne de raisonnement plus solide.",{},{"id":642,"data":643,"type":240,"tunes":645},"S7ybJ-eBFz",{"text":644},"Les sources primaires éliminent l'inférence. Non. Les preuves primaires peuvent encore nécessiter une interprétation.",{},{"id":647,"data":648,"type":240,"tunes":650},"z8t-Sl3iqs",{"text":649},"Plus de sources réduisent automatiquement la distance. Dix faibles répétitions de la même affirmation non étayée ne créent pas de preuve directe.",{},{"id":652,"data":653,"type":42,"tunes":655},"RDzGjjxuuc",{"text":654,"level":234},"Cas limite : Données dérivées",{},{"id":657,"data":658,"type":240,"tunes":660},"eItaEDKK3R",{"text":659},"Certaines valeurs dérivées sont plus fiables que les descriptions humaines brutes. Un total correctement calculé à partir de transactions vérifiées peut être hautement fiable, même s'il ne résulte que d'une seule transformation à partir des enregistrements originaux.",{},{"id":662,"data":663,"type":240,"tunes":665},"vBdc8hymZy",{"text":664},"La Distance aux Preuves mesure donc la séparation épistémique, et non une préférence simpliste pour les données non modifiées.",{},{"id":667,"data":668,"type":42,"tunes":670},"RXQXotzXRw",{"text":669,"level":234},"Limites",{},{"id":672,"data":673,"type":240,"tunes":675},"JjvwlkIJOr",{"text":674},"La Distance aux Preuves est un outil de diagnostic conceptuel, et non un score quantitatif universel.",{},{"id":677,"data":678,"type":240,"tunes":680},"HFJlUDTxF6",{"text":679},"Différents domaines nécessitent un traitement différent de l'inférence. La preuve mathématique, les preuves cliniques, les journaux logiciels, l'analyse de marché et l'interprétation historique ne peuvent pas partager un seuil numérique universel unique.",{},{"id":682,"data":683,"type":42,"tunes":685},"XGjXl2Z8WW",{"text":684,"level":234},"Qu'est-ce qui changerait cette réponse ?",{},{"id":687,"data":688,"type":240,"tunes":690},"SlqRtFaMg3",{"text":689},"Les futurs systèmes d'IA pourraient générer des graphes de provenance explicites, des liens affirmation-preuve vérifiables par machine et des traces de raisonnement typées. Cela pourrait rendre certaines parties de la Distance aux Preuves mesurables automatiquement.",{},{"id":692,"data":693,"type":240,"tunes":695},"RNgkjK7ENz",{"text":694},"Le problème sous-jacent demeure : une affirmation peut s'éloigner de ses preuves même si chaque phrase intermédiaire semble plausible.",{},{"id":697,"data":698,"type":42,"tunes":700},"gCzER2dcxH",{"text":699,"level":234},"Conclusion",{},{"id":702,"data":703,"type":240,"tunes":705},"TsiQdQEYrr",{"text":704},"La distinction importante ne se limite pas à celle entre les réponses sourcées et non sourcées.",{},{"id":707,"data":708,"type":240,"tunes":710},"ou4HQModNd",{"text":709},"Elle se situe entre les affirmations qui restent proches de preuves directement inspectables et celles qui dépendent de chaînes d'interprétation et d'hypothèses de plus en plus longues.",{},{"id":712,"data":713,"type":281,"tunes":715},"8X-yxETmWe",{"code":714},"Evidence\n   ↓\nEvidence Distance\n   ↓\nReasoning\n   ↓\nAnswer Validity Boundary",{},{"id":717,"data":718,"type":240,"tunes":720},"vEy-4dZBXE",{"text":719},"La Distance aux Preuves nous donne un langage pour décrire cet espace.",{},{"id":722,"data":723,"type":42,"tunes":725},"NaB35Wz01G",{"text":724,"level":234},"Sources primaires",{},{"id":727,"data":728,"type":240,"tunes":730},"H25HPflL_j",{"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},"ray0tEnwZO",{"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},"uXwmbDFVoi",{"text":739},"W3C, \u003Ca href=\"https:\u002F\u002Fwww.w3.org\u002FTR\u002Fprov-overview\u002F\" target=\"_blank\">PROV Overview\u003C\u002Fa>. Un modèle standardisé pour représenter les relations de provenance.",{},"2.31.6","Toutes les réponses étayées ne sont pas également proches de leurs preuves. La distance aux preuves est un moyen pratique de décrire combien de transformations, de résumés, d'hypothèses et d'étapes inferentielles séparent une affirmation de l'IA de preuves directement inspectables.","\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,"Risques, contrôles et preuves","risks-and-controls",{"id":764,"name":765,"slug":766},49,"Contrôles et preuves","controls",{"id":768,"login":769,"email":391,"displayName":770},"20","rooth8233","Aleksandar Stajić",[772,1260],{"lang":773,"title":774,"content":775,"contentJson":776,"excerpt":1259},"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,[779,783,788,793,799,803,808,813,818,822,827,832,837,858,863,867,872,877,882,887,892,897,901,906,911,916,921,930,935,940,945,950,955,960,965,970,975,980,985,990,994,1003,1008,1013,1018,1022,1031,1036,1041,1045,1050,1055,1060,1065,1069,1074,1078,1083,1088,1093,1097,1102,1107,1112,1121,1126,1143,1148,1153,1158,1163,1168,1173,1178,1183,1188,1193,1198,1203,1208,1213,1218,1222,1227,1232,1236,1241,1246,1250,1254],{"id":780,"data":781,"type":42,"tunes":782},"g-2Ciyj8LJ",{"text":233,"level":234},{},{"id":784,"data":785,"type":240,"tunes":787},"SQWD7UraiX",{"text":786},"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":789,"data":790,"type":240,"tunes":792},"Sj1fx2865j",{"text":791},"This article uses the term Evidence Distance for the distance between a claim and the directly inspectable evidence that supports it.",{},{"id":794,"data":795,"type":253,"tunes":798},"d5th-A-mq5",{"text":796,"caption":797,"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":800,"data":801,"type":260,"tunes":802},"X81jXC72FC",{"title":258,"maxLevel":259,"minLevel":234},{},{"id":804,"data":805,"type":42,"tunes":807},"_pZcnPEoil",{"text":806,"level":234},"What This Really Means",{},{"id":809,"data":810,"type":240,"tunes":812},"F5odO7xyha",{"text":811},"A citation proves that a source was referenced. It does not automatically prove that the source directly supports the sentence beside the citation.",{},{"id":814,"data":815,"type":240,"tunes":817},"8Gqlg-ArJS",{"text":816},"Between an original observation and a final AI answer there may be extraction, parsing, chunking, retrieval, summarization, interpretation, aggregation and inference.",{},{"id":819,"data":820,"type":281,"tunes":821},"lXkL5LV9ab",{"code":280},{},{"id":823,"data":824,"type":240,"tunes":826},"Z-8GPQIpCI",{"text":825},"Every step can be legitimate. But every step can also introduce distortion.",{},{"id":828,"data":829,"type":42,"tunes":831},"aXMwTGZ8Ia",{"text":830,"level":234},"Simplest Example",{},{"id":833,"data":834,"type":240,"tunes":836},"bZlU7KD3M2",{"text":835},"Suppose a temperature sensor reports 38.2 °C.",{},{"id":838,"data":839,"type":318,"tunes":857},"SP87UUPaU3",{"content":840,"stretched":43,"withHeadings":14},[841,845,849,853],[842,843,844],"Claim","Relationship to evidence","Illustrative distance",[846,847,848],"The sensor reports 38.2 °C.","Direct reading","Low",[850,851,852],"The system is running hotter than yesterday.","Requires comparison with yesterday's reading","Moderate",[854,855,856],"The cooling system is failing.","Requires causal inference and alternative explanations","High",{},{"id":859,"data":860,"type":240,"tunes":862},"eYar059d69",{"text":861},"The original sensor value directly supports the first claim. It does not by itself prove the third.",{},{"id":864,"data":865,"type":281,"tunes":866},"eCnFlByvHt",{"code":328},{},{"id":868,"data":869,"type":42,"tunes":871},"aXGkK6BpkF",{"text":870,"level":234},"Where the Example Stops Working",{},{"id":873,"data":874,"type":240,"tunes":876},"lAwb9QZ_yB",{"text":875},"Evidence Distance is not simply a count of sentences or processing steps.",{},{"id":878,"data":879,"type":240,"tunes":881},"-SzYko9oIH",{"text":880},"One mathematically valid transformation may preserve evidence almost perfectly, while a single speculative interpretation may create a much larger epistemic jump.",{},{"id":883,"data":884,"type":240,"tunes":886},"XOQloBndLw",{"text":885},"The concept therefore includes both distance by transformation and distance by inference.",{},{"id":888,"data":889,"type":42,"tunes":891},"EOXbS4TAOR",{"text":890,"level":234},"Direct Answer",{},{"id":893,"data":894,"type":240,"tunes":896},"HcVL8vGM74",{"text":895},"Evidence Distance increases when a claim depends on more transformations, more assumptions, weaker provenance or stronger inference than the underlying evidence directly warrants.",{},{"id":898,"data":899,"type":281,"tunes":900},"_f7WXjUOsS",{"code":363},{},{"id":902,"data":903,"type":240,"tunes":905},"zUEC-cDhqm",{"text":904},"A longer distance does not automatically make a claim false. It means the reasoning chain requires more validation.",{},{"id":907,"data":908,"type":42,"tunes":910},"ix7TUGNz_W",{"text":909,"level":234},"Why This Matters for AI",{},{"id":912,"data":913,"type":240,"tunes":915},"wrAy5PLtLB",{"text":914},"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":917,"data":918,"type":240,"tunes":920},"jJseLuTWdv",{"text":919},"An answer can therefore look well sourced while containing claims whose support exists only through several unstated inferential steps.",{},{"id":922,"data":923,"type":394,"tunes":929},"ox8fok9dzs",{"link":924,"meta":925},"https:\u002F\u002Fstajic.de\u002Fblog\u002Ffrom-research-protocol-to-a-general-ai-reasoning-framework",{"image":926,"title":927,"description":928},{"url":391},"From Research Protocol to a General AI Reasoning Framework","Reliable reasoning requires explicit separation of evidence, assumptions, hypotheses and conclusions.",{},{"id":931,"data":932,"type":42,"tunes":934},"6ipse6YG1U",{"text":933,"level":234},"Four Components of Evidence Distance",{},{"id":936,"data":937,"type":42,"tunes":939},"P4xQ7o96dR",{"text":938,"level":259},"1. Transformation Distance",{},{"id":941,"data":942,"type":240,"tunes":944},"dvN8TUNEYG",{"text":943},"How many times has the original evidence been converted, summarized, normalized or reformatted before reaching the model?",{},{"id":946,"data":947,"type":42,"tunes":949},"dXofACF9_x",{"text":948,"level":259},"2. Inferential Distance",{},{"id":951,"data":952,"type":240,"tunes":954},"kkEzV3ydId",{"text":953},"How many reasoning steps are required before the evidence supports the final conclusion?",{},{"id":956,"data":957,"type":42,"tunes":959},"PPYB0RuX5_",{"text":958,"level":259},"3. Provenance Distance",{},{"id":961,"data":962,"type":240,"tunes":964},"cAVBYebJr7",{"text":963},"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":966,"data":967,"type":42,"tunes":969},"254N_99Z7X",{"text":968,"level":259},"4. Applicability Distance",{},{"id":971,"data":972,"type":240,"tunes":974},"to3KWv5WSK",{"text":973},"How much reasoning is required to establish that the evidence applies to this exact version, user, jurisdiction, time period or situation?",{},{"id":976,"data":977,"type":42,"tunes":979},"3WnVPQ6wqw",{"text":978,"level":234},"Evidence Distance in RAG",{},{"id":981,"data":982,"type":240,"tunes":984},"r8s-75EBhA",{"text":983},"RAG can shorten one important distance: the distance between a model and external knowledge.",{},{"id":986,"data":987,"type":240,"tunes":989},"I_8h52A3Yc",{"text":988},"It does not automatically shorten the distance between a retrieved passage and the final claim.",{},{"id":991,"data":992,"type":281,"tunes":993},"jDDRUu2FUK",{"code":459},{},{"id":995,"data":996,"type":394,"tunes":1002},"rm14qIVB1M",{"link":997,"meta":998},"https:\u002F\u002Fstajic.de\u002Fblog\u002Fwhere-does-an-llm-get-its-data-rag-data-sources-in-python",{"image":999,"title":1000,"description":1001},{"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":1004,"data":1005,"type":42,"tunes":1007},"CRUZDXYloa",{"text":1006,"level":234},"Connection to the Retrieval Trigger",{},{"id":1009,"data":1010,"type":240,"tunes":1012},"-TPiYJ569c",{"text":1011},"The Retrieval Trigger asks whether internal model knowledge is sufficient.",{},{"id":1014,"data":1015,"type":240,"tunes":1017},"_VjLi5fs4x",{"text":1016},"Evidence Distance asks a later question: after evidence has been retrieved, how directly does it support what the system wants to claim?",{},{"id":1019,"data":1020,"type":281,"tunes":1021},"EImPw6i617",{"code":488},{},{"id":1023,"data":1024,"type":394,"tunes":1030},"ZqyhB2bFtR",{"link":1025,"meta":1026},"https:\u002F\u002Fstajic.de\u002Fblog\u002Fwhen-should-an-ai-stop-trusting-its-own-knowledge-the-retrieval-trigger",{"image":1027,"title":1028,"description":1029},{"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":1032,"data":1033,"type":42,"tunes":1035},"ST2vnm6vwF",{"text":1034,"level":234},"A Practical Diagnostic",{},{"id":1037,"data":1038,"type":240,"tunes":1040},"qyoHcDcTKb",{"text":1039},"For an important AI-generated claim, reconstruct the support chain backwards.",{},{"id":1042,"data":1043,"type":281,"tunes":1044},"e3fDmLk0fe",{"code":512},{},{"id":1046,"data":1047,"type":240,"tunes":1049},"iQ60c7NRaE",{"text":1048},"If the chain breaks, the system has encountered an evidence-traceability problem.",{},{"id":1051,"data":1052,"type":240,"tunes":1054},"YQAKlNZ2gM",{"text":1053},"If the chain remains intact but requires several strong assumptions, the Evidence Distance is still high.",{},{"id":1056,"data":1057,"type":42,"tunes":1059},"clN7jFkUBf",{"text":1058,"level":234},"Example: Product Data",{},{"id":1061,"data":1062,"type":240,"tunes":1064},"CRF-A9O7Ot",{"text":1063},"Suppose a manufacturer specification says a laptop weighs 1.48 kg.",{},{"id":1066,"data":1067,"type":281,"tunes":1068},"JF3z6aF2q_",{"code":537},{},{"id":1070,"data":1071,"type":240,"tunes":1073},"vTMOOgkLPb",{"text":1072},"Now consider:",{},{"id":1075,"data":1076,"type":281,"tunes":1077},"Tuhfzor1yp",{"code":547},{},{"id":1079,"data":1080,"type":240,"tunes":1082},"JHFRGIgUJD",{"text":1081},"The specification is real. The recommendation is still several reasoning steps away from it.",{},{"id":1084,"data":1085,"type":42,"tunes":1087},"2lKERCrnW1",{"text":1086,"level":234},"Example: Scientific Evidence",{},{"id":1089,"data":1090,"type":240,"tunes":1092},"HA8DBYwUuK",{"text":1091},"A paper may report an observed association. An AI answer can incorrectly compress that into a causal statement.",{},{"id":1094,"data":1095,"type":281,"tunes":1096},"D9C2dVB7j_",{"code":567},{},{"id":1098,"data":1099,"type":240,"tunes":1101},"CKNjK25pkL",{"text":1100},"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":1103,"data":1104,"type":42,"tunes":1106},"sR430N3cDd",{"text":1105,"level":234},"Connection to Agent-Ready Context",{},{"id":1108,"data":1109,"type":240,"tunes":1111},"v77YyG55ko",{"text":1110},"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":1113,"data":1114,"type":394,"tunes":1120},"EX8darCIjd",{"link":1115,"meta":1116},"https:\u002F\u002Fstajic.de\u002Fportfolio\u002Ffrom-raw-data-to-verifiable-ai-context-a-practical-architecture-for-agent-ready-information",{"image":1117,"title":1118,"description":1119},{"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":1122,"data":1123,"type":42,"tunes":1125},"dsfuoURj1e",{"text":1124,"level":234},"A Simple Working Scale",{},{"id":1127,"data":1128,"type":318,"tunes":1142},"BLi7Yth64O",{"content":1129,"stretched":43,"withHeadings":14},[1130,1132,1134,1136,1138,1140],[1131,604],"Level",[606,1133],"Directly reported source fact",[609,1135],"Simple deterministic transformation or direct comparison",[612,1137],"One meaningful interpretive step",[615,1139],"Multiple assumptions or combined sources",[618,1141],"Strong causal, predictive or prescriptive conclusion beyond direct evidence",{},{"id":1144,"data":1145,"type":240,"tunes":1147},"8y8bzCcHBs",{"text":1146},"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":1149,"data":1150,"type":42,"tunes":1152},"aFK0YiAlYO",{"text":1151,"level":234},"Misconceptions",{},{"id":1154,"data":1155,"type":240,"tunes":1157},"07KSk6xDx-",{"text":1156},"A citation means Evidence Distance is low. No. A citation can sit beside a claim that substantially exceeds what the cited source establishes.",{},{"id":1159,"data":1160,"type":240,"tunes":1162},"t_cf-Zfan9",{"text":1161},"Long Evidence Distance means the claim is wrong. No. Complex conclusions can be well supported, but they require a stronger reasoning chain.",{},{"id":1164,"data":1165,"type":240,"tunes":1167},"Pu9PRX67Q6",{"text":1166},"Primary sources eliminate inference. No. Primary evidence can still require interpretation.",{},{"id":1169,"data":1170,"type":240,"tunes":1172},"oF1xNsOHHq",{"text":1171},"More sources automatically reduce distance. Ten weak repetitions of the same unsupported claim do not create direct evidence.",{},{"id":1174,"data":1175,"type":42,"tunes":1177},"IkolK5baph",{"text":1176,"level":234},"Edge Case: Derived Data",{},{"id":1179,"data":1180,"type":240,"tunes":1182},"jRH36fwdR3",{"text":1181},"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":1184,"data":1185,"type":240,"tunes":1187},"SXbdKiNDst",{"text":1186},"Evidence Distance therefore measures epistemic separation, not a simplistic preference for untouched data.",{},{"id":1189,"data":1190,"type":42,"tunes":1192},"mUbvdnFChM",{"text":1191,"level":234},"Limitations",{},{"id":1194,"data":1195,"type":240,"tunes":1197},"N_EskxFOep",{"text":1196},"Evidence Distance is a conceptual diagnostic, not a universal quantitative score.",{},{"id":1199,"data":1200,"type":240,"tunes":1202},"nW_-E08So_",{"text":1201},"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":1204,"data":1205,"type":42,"tunes":1207},"_Bd95wXkri",{"text":1206,"level":234},"What Would Change This Answer?",{},{"id":1209,"data":1210,"type":240,"tunes":1212},"OqCaMc2_OX",{"text":1211},"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":1214,"data":1215,"type":240,"tunes":1217},"58Hf3O9-Iz",{"text":1216},"The underlying problem remains: a claim can move farther from its evidence even while every intermediate sentence sounds plausible.",{},{"id":1219,"data":1220,"type":42,"tunes":1221},"8WOq5oE_e_",{"text":699,"level":234},{},{"id":1223,"data":1224,"type":240,"tunes":1226},"1pr1TfwnKc",{"text":1225},"The important distinction is not merely between sourced and unsourced answers.",{},{"id":1228,"data":1229,"type":240,"tunes":1231},"54CysT_NXa",{"text":1230},"It is between claims that remain close to directly inspectable evidence and claims that depend on increasingly long chains of interpretation and assumption.",{},{"id":1233,"data":1234,"type":281,"tunes":1235},"Qrjo3WEUSL",{"code":714},{},{"id":1237,"data":1238,"type":240,"tunes":1240},"6ir8aIAfd3",{"text":1239},"Evidence Distance gives us a language for describing that space.",{},{"id":1242,"data":1243,"type":42,"tunes":1245},"vPhupXcbeN",{"text":1244,"level":234},"Primary Sources",{},{"id":1247,"data":1248,"type":240,"tunes":1249},"BaO4JmWRQL",{"text":729},{},{"id":1251,"data":1252,"type":240,"tunes":1253},"eg4Z4G6Cgs",{"text":734},{},{"id":1255,"data":1256,"type":240,"tunes":1258},"BJH7zUd4T8",{"text":1257},"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":1261,"excerpt":742},{"time":228,"blocks":1262,"version":741},[1263,1266,1269,1272,1275,1278,1281,1284,1287,1290,1293,1296,1299,1307,1310,1313,1316,1319,1322,1325,1328,1331,1334,1337,1340,1343,1346,1351,1354,1357,1360,1363,1366,1369,1372,1375,1378,1381,1384,1387,1390,1395,1398,1401,1404,1407,1412,1415,1418,1421,1424,1427,1430,1433,1436,1439,1442,1445,1448,1451,1454,1457,1460,1463,1468,1471,1481,1484,1487,1490,1493,1496,1499,1502,1505,1508,1511,1514,1517,1520,1523,1526,1529,1532,1535,1538,1541,1544,1547,1550],{"id":231,"data":1264,"type":42,"tunes":1265},{"text":233,"level":234},{},{"id":237,"data":1267,"type":240,"tunes":1268},{"text":239},{},{"id":243,"data":1270,"type":240,"tunes":1271},{"text":245},{},{"id":248,"data":1273,"type":253,"tunes":1274},{"text":250,"caption":251,"alignment":252},{},{"id":256,"data":1276,"type":260,"tunes":1277},{"title":258,"maxLevel":259,"minLevel":234},{},{"id":263,"data":1279,"type":42,"tunes":1280},{"text":265,"level":234},{},{"id":268,"data":1282,"type":240,"tunes":1283},{"text":270},{},{"id":273,"data":1285,"type":240,"tunes":1286},{"text":275},{},{"id":278,"data":1288,"type":281,"tunes":1289},{"code":280},{},{"id":284,"data":1291,"type":240,"tunes":1292},{"text":286},{},{"id":289,"data":1294,"type":42,"tunes":1295},{"text":291,"level":234},{},{"id":294,"data":1297,"type":240,"tunes":1298},{"text":296},{},{"id":299,"data":1300,"type":318,"tunes":1306},{"content":1301,"stretched":43,"withHeadings":14},[1302,1303,1304,1305],[303,304,305],[307,308,309],[311,312,313],[315,316,317],{},{"id":321,"data":1308,"type":240,"tunes":1309},{"text":323},{},{"id":326,"data":1311,"type":281,"tunes":1312},{"code":328},{},{"id":331,"data":1314,"type":42,"tunes":1315},{"text":333,"level":234},{},{"id":336,"data":1317,"type":240,"tunes":1318},{"text":338},{},{"id":341,"data":1320,"type":240,"tunes":1321},{"text":343},{},{"id":346,"data":1323,"type":240,"tunes":1324},{"text":348},{},{"id":351,"data":1326,"type":42,"tunes":1327},{"text":353,"level":234},{},{"id":356,"data":1329,"type":240,"tunes":1330},{"text":358},{},{"id":361,"data":1332,"type":281,"tunes":1333},{"code":363},{},{"id":366,"data":1335,"type":240,"tunes":1336},{"text":368},{},{"id":371,"data":1338,"type":42,"tunes":1339},{"text":373,"level":234},{},{"id":376,"data":1341,"type":240,"tunes":1342},{"text":378},{},{"id":381,"data":1344,"type":240,"tunes":1345},{"text":383},{},{"id":386,"data":1347,"type":394,"tunes":1350},{"link":388,"meta":1348},{"image":1349,"title":392,"description":393},{"url":391},{},{"id":397,"data":1352,"type":42,"tunes":1353},{"text":399,"level":234},{},{"id":402,"data":1355,"type":42,"tunes":1356},{"text":404,"level":259},{},{"id":407,"data":1358,"type":240,"tunes":1359},{"text":409},{},{"id":412,"data":1361,"type":42,"tunes":1362},{"text":414,"level":259},{},{"id":417,"data":1364,"type":240,"tunes":1365},{"text":419},{},{"id":422,"data":1367,"type":42,"tunes":1368},{"text":424,"level":259},{},{"id":427,"data":1370,"type":240,"tunes":1371},{"text":429},{},{"id":432,"data":1373,"type":42,"tunes":1374},{"text":434,"level":259},{},{"id":437,"data":1376,"type":240,"tunes":1377},{"text":439},{},{"id":442,"data":1379,"type":42,"tunes":1380},{"text":444,"level":234},{},{"id":447,"data":1382,"type":240,"tunes":1383},{"text":449},{},{"id":452,"data":1385,"type":240,"tunes":1386},{"text":454},{},{"id":457,"data":1388,"type":281,"tunes":1389},{"code":459},{},{"id":462,"data":1391,"type":394,"tunes":1394},{"link":464,"meta":1392},{"image":1393,"title":467,"description":468},{"url":391},{},{"id":471,"data":1396,"type":42,"tunes":1397},{"text":473,"level":234},{},{"id":476,"data":1399,"type":240,"tunes":1400},{"text":478},{},{"id":481,"data":1402,"type":240,"tunes":1403},{"text":483},{},{"id":486,"data":1405,"type":281,"tunes":1406},{"code":488},{},{"id":491,"data":1408,"type":394,"tunes":1411},{"link":493,"meta":1409},{"image":1410,"title":496,"description":497},{"url":391},{},{"id":500,"data":1413,"type":42,"tunes":1414},{"text":502,"level":234},{},{"id":505,"data":1416,"type":240,"tunes":1417},{"text":507},{},{"id":510,"data":1419,"type":281,"tunes":1420},{"code":512},{},{"id":515,"data":1422,"type":240,"tunes":1423},{"text":517},{},{"id":520,"data":1425,"type":240,"tunes":1426},{"text":522},{},{"id":525,"data":1428,"type":42,"tunes":1429},{"text":527,"level":234},{},{"id":530,"data":1431,"type":240,"tunes":1432},{"text":532},{},{"id":535,"data":1434,"type":281,"tunes":1435},{"code":537},{},{"id":540,"data":1437,"type":240,"tunes":1438},{"text":542},{},{"id":545,"data":1440,"type":281,"tunes":1441},{"code":547},{},{"id":550,"data":1443,"type":240,"tunes":1444},{"text":552},{},{"id":555,"data":1446,"type":42,"tunes":1447},{"text":557,"level":234},{},{"id":560,"data":1449,"type":240,"tunes":1450},{"text":562},{},{"id":565,"data":1452,"type":281,"tunes":1453},{"code":567},{},{"id":570,"data":1455,"type":240,"tunes":1456},{"text":572},{},{"id":575,"data":1458,"type":42,"tunes":1459},{"text":577,"level":234},{},{"id":580,"data":1461,"type":240,"tunes":1462},{"text":582},{},{"id":585,"data":1464,"type":394,"tunes":1467},{"link":587,"meta":1465},{"image":1466,"title":590,"description":591},{"url":391},{},{"id":594,"data":1469,"type":42,"tunes":1470},{"text":596,"level":234},{},{"id":599,"data":1472,"type":318,"tunes":1480},{"content":1473,"stretched":43,"withHeadings":14},[1474,1475,1476,1477,1478,1479],[603,604],[606,607],[609,610],[612,613],[615,616],[618,619],{},{"id":622,"data":1482,"type":240,"tunes":1483},{"text":624},{},{"id":627,"data":1485,"type":42,"tunes":1486},{"text":629,"level":234},{},{"id":632,"data":1488,"type":240,"tunes":1489},{"text":634},{},{"id":637,"data":1491,"type":240,"tunes":1492},{"text":639},{},{"id":642,"data":1494,"type":240,"tunes":1495},{"text":644},{},{"id":647,"data":1497,"type":240,"tunes":1498},{"text":649},{},{"id":652,"data":1500,"type":42,"tunes":1501},{"text":654,"level":234},{},{"id":657,"data":1503,"type":240,"tunes":1504},{"text":659},{},{"id":662,"data":1506,"type":240,"tunes":1507},{"text":664},{},{"id":667,"data":1509,"type":42,"tunes":1510},{"text":669,"level":234},{},{"id":672,"data":1512,"type":240,"tunes":1513},{"text":674},{},{"id":677,"data":1515,"type":240,"tunes":1516},{"text":679},{},{"id":682,"data":1518,"type":42,"tunes":1519},{"text":684,"level":234},{},{"id":687,"data":1521,"type":240,"tunes":1522},{"text":689},{},{"id":692,"data":1524,"type":240,"tunes":1525},{"text":694},{},{"id":697,"data":1527,"type":42,"tunes":1528},{"text":699,"level":234},{},{"id":702,"data":1530,"type":240,"tunes":1531},{"text":704},{},{"id":707,"data":1533,"type":240,"tunes":1534},{"text":709},{},{"id":712,"data":1536,"type":281,"tunes":1537},{"code":714},{},{"id":717,"data":1539,"type":240,"tunes":1540},{"text":719},{},{"id":722,"data":1542,"type":42,"tunes":1543},{"text":724,"level":234},{},{"id":727,"data":1545,"type":240,"tunes":1546},{"text":729},{},{"id":732,"data":1548,"type":240,"tunes":1549},{"text":734},{},{"id":737,"data":1551,"type":240,"tunes":1552},{"text":739},{},"Post erfolgreich abgerufen",{"items":1555,"source":1612,"manualIds":1613,"manualMatchedIds":1614},[1556,1563,1570,1577,1584,1591,1598,1605],{"id":1557,"slug":1558,"title":1559,"excerpt":1560,"featuredImage":1561,"publishedAt":1562},"481","generative-ai-explained-models-retrieval-tools-and-applications-are-not-the-same-thing","L'IA générative expliquée : modèles, recherche, outils et applications ne sont pas la même chose","L'IA générative est plus qu'un modèle. Découvrez comment les modèles, la récupération, les outils, le contexte, les environnements d'exécution et les applications s'articulent dans les systèmes d'IA en production.","\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":1564,"slug":1565,"title":1566,"excerpt":1567,"featuredImage":1568,"publishedAt":1569},"482","adr-vs-nfr-architecture-decisions-and-system-quality-are-not-the-same-thing","ADR vs NFR : les décisions d'architecture et la qualité du système ne sont pas la même chose","ADR vs NFR expliqués : découvrez comment les exigences de qualité système orientent les décisions d'architecture, comment les ADR consignent les compromis et pourquoi la validation reste distincte.","\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":1571,"slug":1572,"title":1573,"excerpt":1574,"featuredImage":1575,"publishedAt":1576},"381","enterprise-grade-multi-tenant-architecture-for-an-international-platform","Enterprise-Grade Multi-Tenant Architecture for an International Platform","Loving Rocks is an enterprise-grade wedding platform designed with a true multi-tenant architecture, isolated databases per tenant, and built-in internationalization for global scalability, security, and long-term operational stability.","\u002Fuploads\u002F2026\u002F01\u002Fenterprise-grade-multi-tenant-architecture-for-an-international-platform-1769789121298-b6v7ak.webp","2026-01-30T12:04:00.000Z",{"id":1578,"slug":1579,"title":1580,"excerpt":1581,"featuredImage":1582,"publishedAt":1583},"492","mcp-explained-what-it-connects-what-it-does-not-do-and-where-it-fits","MCP expliqué : ce qu'il connecte, ce qu'il ne fait pas et où il s'intègre","Le protocole de contexte de modèle connecte les applications d’IA à des outils, des ressources et des invites externes par le biais d’une frontière standard client-serveur. Découvrez ce que fait le MCP, ce qu’il ne fait pas et où il s’intègre dans l’architecture des agents.","\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":1585,"slug":1586,"title":1587,"excerpt":1588,"featuredImage":1589,"publishedAt":1590},"495","sovereign-ai-control-of-models-data-infrastructure-and-dependencies","IA souveraine : contrôle des modèles, des données, des infrastructures et des dépendances","L'IA souveraine concerne le contrôle effectif sur les modèles, les données, l'infrastructure, les logiciels, les opérations et les dépendances stratégiques — et non simplement l'endroit où un modèle d'IA est hébergé.","\u002Fuploads\u002F2026\u002F10\u002Fsovereign-ai-control-of-models-data-infrastructure-and-dependencies-1791488833132-niy85x.webp","2026-10-08T15:45:00.000Z",{"id":1592,"slug":1593,"title":1594,"excerpt":1595,"featuredImage":1596,"publishedAt":1597},"494","air-gapped-ai-how-ai-systems-work-without-internet-or-cloud-access","IA en environnement isolé : comment les systèmes d’IA fonctionnent sans accès à Internet ni au cloud","L'IA en environnement isolé exécute des modèles, du RAG et des applications d'IA à l'intérieur d'un domaine de sécurité isolé, sans dépendance à Internet ni au cloud. Découvrez comment les modèles, les données, les mises à jour et les outils fonctionnent hors ligne.","\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":1599,"slug":1600,"title":1601,"excerpt":1602,"featuredImage":1603,"publishedAt":1604},"498","can-evidence-survive-the-ai-pipeline-the-evidence-preservation-chain","Les preuves peuvent-elles survivre au pipeline d’IA ? — La chaîne de préservation des preuves","Une preuve peut être correcte à la source et devenir néanmoins peu fiable après l'analyse syntaxique, la normalisation, le découpage, la récupération et l'assemblage du contexte. La chaîne de préservation des preuves vérifie si l'identité, la provenance, la portée et le sens survivent à l'ensemble du pipeline d'IA.","\u002Fuploads\u002F2026\u002F10\u002Fcan-evidence-survive-the-ai-pipeline-the-evidence-preservation-chain-1792096396174-urdg9z.webp","2026-10-15T16:31:00.000Z",{"id":1606,"slug":1607,"title":1608,"excerpt":1609,"featuredImage":1610,"publishedAt":1611},"491","ai-governance-models-data-permissions-risk-and-auditability","Gouvernance de l'IA : modèles, données, autorisations, risques et auditabilité","La gouvernance de l'IA définit qui peut approuver, exploiter, modifier et auditer les systèmes d'IA à travers les modèles, les fournisseurs, les données, les autorisations, les risques, l'évaluation et l'ensemble du cycle de vie.","\u002Fuploads\u002F2026\u002F10\u002Fai-governance-models-data-permissions-risk-and-auditability-1791485901301-g60xyu.webp","2026-10-08T12:56:00.000Z","fallback",[],[]]