[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-why-rag-is-ending-for-agentic-ai-zh":3,"tags-why-rag-is-ending-for-agentic-ai-zh":35,"related-lang-why-rag-is-ending-for-agentic-ai-zh":46,"related-posts-why-rag-is-ending-for-agentic-ai-zh":50,"series-ai-agent-261fa342-f7bb-4330-a97a-a95f10ae3f94":87},{"id":4,"title":5,"content":6,"summary":7,"source":8,"source_url":9,"author":10,"image_url":11,"keywords":12,"language":19,"translated_content":10,"views":20,"is_premium":21,"created_at":22,"updated_at":22,"cover_image":11,"published_at":23,"rewrite_status":24,"rewrite_error":10,"rewritten_from_id":25,"slug":26,"category":27,"related_article_id":28,"status":29,"google_indexed_at":30,"x_posted_at":10,"tweet_text":10,"title_rewritten_at":10,"title_original":10,"key_takeaways":31,"topic_cluster_id":10,"embedding":10,"is_canonical_seed":21},"261fa342-f7bb-4330-a97a-a95f10ae3f94","為什麼 RAG 正在結束，agentic AI 需要的是編譯式知識層","\u003Cp data-speakable=\"summary\">\u003Ca href=\"\u002Fnews\u002Fwhy-rag-needs-self-healing-layer-zh\">RAG\u003C\u002Fa> 不適合當 \u003Ca href=\"\u002Ftag\u002Fagentic-ai\">agentic AI\u003C\u002Fa> 的預設架構，代理需要先編譯好的知識層，而不是每一步都重新檢索原文。\u003C\u002Fp>\u003Cp>我認為，\u003Ca href=\"\u002Ftag\u002Frag\">RAG\u003C\u002Fa> 正在成為 agentic AI 的過渡方案，真正會留下來的是編譯式知識層，而不是更強的向量搜尋。原因很直接：代理不是聊天機器人，不能靠使用者幫它整理上下文。Pinecone 的 Nexus 方向很能說明這件事，它談的不是把相似片段找得更準，而是把企業資料先整理成可重用的任務物件，並讓代理用 KnowQL 去要求輸出形狀、信心與延遲。這代表市場已經把問題從「找得到文本嗎」改成「能不能直接產出可執行知識」。\u003C\u002Fp>\u003Ch2>第一個論點：代理需要穩定的中間產物，不是每輪重搜\u003C\u002Fh2>\u003Cp>人類在看 RAG 回答時，還能自己比對、取捨、補缺；代理做不到。多步驟任務一旦每一步都要重新檢索，系統就會把同樣的上下文反覆拉進來，成本、延遲與錯誤率一起上升。Pinecone 對 context compiler 的描述，重點就在於先把資料編成持久的任務知識，再交給代理執行。這不是把搜尋做得更快，而是把「搜尋後回答」改成「先編譯，再行動」。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778105450961-s3p1.png\" alt=\"為什麼 RAG 正在結束，agentic AI 需要的是編譯式知識層\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>\u003Ca href=\"\u002Ftag\u002Ftoken\">token\u003C\u002Fa> 經濟學把差異放得更大。Pinecone 宣稱，一個財務分析任務從 280 萬 tokens 降到 4,000 tokens，縮水幅度接近 98%。就算這是供應商基準，實際上線未必完全複製，方向仍然非常清楚：代理如果每一步都拖著巨大上下文走，延遲、費用和推理漂移會一起惡化。編譯層把昂貴的理解前置，產出可重用的 artifact，這是架構級改變，不是 RAG 的小修小補。\u003C\u002Fp>\u003Ch2>第二個論點：企業場景要的是結構化、可追溯的知識\u003C\u002Fh2>\u003Cp>企業資料不是一堆可任意排序的段落，而是合約、政策、表格、工單、報表與日誌的混合體，還常常互相衝突、版本不一、欄位缺漏。傳統 RAG 把這些都當成文字去 embedding 和 ranking，做廣泛召回還行，但一碰到合規摘要、財務對帳、風險判讀，就很容易失手。Nexus 強調欄位級引用與 deterministic conflict resolution，說白了就是承認企業工作不是「找相似句子」而已，而是要有來源、結構與衝突處理。\u003C\u002Fp>\u003Cp>市場也在往這裡投票。Vent\u003Ca href=\"\u002Fnews\u002Ffigure-billion-month-tokenized-credit-breakout-zh\">ure\u003C\u002Fa>Beat 的 Q1 2026 Pulse survey 顯示，獨立向量資料庫的採用份額在下降，而 hybrid retrieval 的意圖已經升到 33.3%。這不是偏好微調，而是買方在說：只靠 retrieval 不夠，還要結構、規則與可驗證輸出。換句話說，向量索引可以留在底層，但產品敘事不該再是「我們有更好的 search」。真正該賣的是把資料編成代理能直接使用的知識物件。\u003C\u002Fp>\u003Ch2>反方可能怎麼說\u003C\u002Fh2>\u003Cp>最強的反方論點其實很合理：RAG 便宜、簡單、成熟，團隊知道怎麼切 chunk、做 embedding、取 top-k，再把內容丟給 \u003Ca href=\"\u002Fnews\u002Fwhy-open-source-llms-should-be-judged-by-workload-not-hype-zh\">LLM\u003C\u002Fa>。對多數還在做 PoC 的團隊來說，這條路最快，也最容易向內部說明。相較之下，編譯式知識層多了預處理、任務化 artifact、刷新邏輯與新的查詢語意，工程成本更高，短期看起來不像是最划算的選擇。\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778105452105-kero.png\" alt=\"為什麼 RAG 正在結束，agentic AI 需要的是編譯式知識層\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>另一個反方顧慮也成立：如果編譯太深，知識層可能過度貼合某些任務，失去彈性。RAG 的優點就是把原始材料留在旁邊，資料變動時也比較容易補救；編譯後的 artifact 若不同步，就會產生漂移。對變動快、風險低、或探索性很強的場景，RAG 的確比較耐用。\u003C\u002Fp>\u003Cp>但這些理由只能證明 RAG 還有用，不能證明它還應該是 agentic AI 的主架構。當任務要求可重複執行、來源可追蹤、跨步驟可重用時，反覆檢索的簡單性會被持續放大的成本吞掉。RAG 最適合的是原型、開放式探索與低風險助理；一旦要做真正的代理工作，編譯式知識層才是最低可行架構。\u003C\u002Fp>\u003Ch2>你能做什麼\u003C\u002Fh2>\u003Cp>如果你在做 \u003Ca href=\"\u002Ftag\u002Fagent\">agent\u003C\u002Fa>，請把 retrieval 從核心產品位置移開，改成把知識層設計成「為行動而編譯」的系統。工程師應該把來源攝取、artifact 生成、代理查詢介面分開，並把刷新與衝突處理當成一等公民。PM 不要只看 retrieval accuracy，要看任務完成率、token burn、衝突處理與引用品質。創辦人則應該直接把產品敘事轉向 durable knowledge objects 和 declarative agent queries，因為市場正在往這裡走。向量搜尋可以留，但不要再把 index 誤認成 intelligence layer。\u003C\u002Fp>","RAG 不再適合作為 agentic AI 的預設架構，因為代理需要可重用、可驗證的編譯式知識層，而不是每一步都重新檢索原始文本。","rocketnews.com","https:\u002F\u002Frocketnews.com\u002F2026\u002F05\u002Fthe-rag-era-is-ending-for-agentic-ai-a-new-compilation-stage-knowledge-layer-is-what-comes-next\u002F",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778105450961-s3p1.png",[13,14,15,16,17,18],"RAG","agentic AI","knowledge layer","vector database","context compiler","Pinecone Nexus","zh",2,false,"2026-05-06T22:10:28.570639+00:00","2026-05-06T22:10:28.372+00:00","done","625dd6a5-d6ab-457e-8b1e-8cfe457c9b03","why-rag-is-ending-for-agentic-ai-zh","ai-agent","71f9b52e-54c0-4df7-acbc-3edf5628a0b7","published","2026-05-07T09:00:18.649+00:00",[32,33,34],"RAG 仍可用於原型與低風險助理，但不適合當 agentic AI 的預設架構。","代理需要先編譯好的、可重用且可追溯的知識物件，而不是每一步重新檢索原文。","工程與產品指標應從檢索準確率，轉向任務完成率、token 成本、衝突處理與引用品質。",[36,38,40,42,44],{"name":13,"slug":37},"rag",{"name":17,"slug":39},"context-compiler",{"name":15,"slug":41},"knowledge-layer",{"name":16,"slug":43},"vector-database",{"name":14,"slug":45},"agentic-ai",{"id":28,"slug":47,"title":48,"language":49},"why-rag-is-ending-for-agentic-ai-en","Why RAG is ending for agentic AI","en",[51,57,63,69,75,81],{"id":52,"slug":53,"title":54,"cover_image":55,"image_url":55,"created_at":56,"category":27},"38406a12-f833-4c69-ae22-99c31f03dd52","switch-ai-outputs-markdown-to-html-zh","怎麼把 AI 輸出改成 HTML","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778743243861-8901.png","2026-05-14T07:20:21.545364+00:00",{"id":58,"slug":59,"title":60,"cover_image":61,"image_url":61,"created_at":62,"category":27},"c7c69fe4-97e3-4edf-a9d6-a79d0c4495b4","anthropic-cat-wu-proactive-ai-assistants-zh","Cat Wu 談 Claude 的主動式 AI","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778735455993-gnw7.png","2026-05-14T05:10:30.453046+00:00",{"id":64,"slug":65,"title":66,"cover_image":67,"image_url":67,"created_at":68,"category":27},"e1d6acda-fa49-4514-aa75-709504be9f93","how-to-run-hermes-agent-on-discord-zh","如何在 Discord 執行 Hermes Agent","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778724655796-cjul.png","2026-05-14T02:10:34.362605+00:00",{"id":70,"slug":71,"title":72,"cover_image":73,"image_url":73,"created_at":74,"category":27},"4104fa5f-d95f-45c5-9032-99416cf0365c","why-ragflow-is-the-right-open-source-rag-engine-to-self-host-zh","為什麼 RAGFlow 是最適合自架的開源 RAG 引擎","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778674262278-1630.png","2026-05-13T12:10:23.762632+00:00",{"id":76,"slug":77,"title":78,"cover_image":79,"image_url":79,"created_at":80,"category":27},"7095f05c-34f5-469f-a044-2525d2010ce9","how-to-add-temporal-rag-in-production-zh","如何在正式環境加入 Temporal RAG","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778667053844-osvs.png","2026-05-13T10:10:30.930982+00:00",{"id":82,"slug":83,"title":84,"cover_image":85,"image_url":85,"created_at":86,"category":27},"10479c95-53c6-4723-9aaa-2fde5fb19ee7","github-agentic-workflows-ai-github-actions-zh","GitHub 把 AI 代理放進 Actions","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778551884342-8io7.png","2026-05-12T02:11:02.069769+00:00",[88,93,98,103,108,113,118,123,128,133],{"id":89,"slug":90,"title":91,"created_at":92},"4ae1e197-1d3d-4233-8733-eafe9cb6438b","claude-now-uses-your-pc-to-finish-tasks-zh","Claude 開始幫你操作電腦","2026-03-26T07:20:48.457387+00:00",{"id":94,"slug":95,"title":96,"created_at":97},"5bede67f-e21c-413d-9ab8-54a3c3d26227","googles-2026-ai-agent-report-decoded-zh","Google 2026 AI Agent 報告解讀","2026-03-26T11:15:22.651956+00:00",{"id":99,"slug":100,"title":101,"created_at":102},"2987d097-563f-46c7-b76f-b558d8ef7c2b","kimi-k25-review-stronger-still-not-legend-zh","Kimi K2.5 評測：更強，但還不是神作","2026-03-27T07:15:55.277513+00:00",{"id":104,"slug":105,"title":106,"created_at":107},"95c9053b-e3f4-4cb5-aace-5c54f4c9e044","claude-code-controls-mac-desktop-zh","Claude Code 也能操控 Mac 了","2026-03-28T03:01:58.58121+00:00",{"id":109,"slug":110,"title":111,"created_at":112},"dc58e153-e3a8-4c06-9b96-1aa64eabbf5f","cloudflare-100x-faster-ai-agent-sandbox-zh","Cloudflare 的 AI 沙箱跑超快","2026-03-28T03:09:44.142236+00:00",{"id":114,"slug":115,"title":116,"created_at":117},"1c8afc56-253f-47a2-979f-1065ff072f2a","openai-backs-isara-agent-swarm-bet-zh","OpenAI 挺 Isara 的 agent swarm …","2026-03-28T03:15:27.513155+00:00",{"id":119,"slug":120,"title":121,"created_at":122},"7379b422-576e-45df-ad5a-d57a0d9dd467","openai-plan-automated-ai-researcher-zh","OpenAI 想做自動化 AI 研究員","2026-03-28T03:17:42.090548+00:00",{"id":124,"slug":125,"title":126,"created_at":127},"48c9889e-86df-450b-a356-e4a4b7c83c5b","harness-engineering-ai-agent-reliability-2026-zh","駕馭工程：從「馬具」到「作業系統」，AI Agent 可靠性的終極密碼","2026-03-31T06:42:53.556721+00:00",{"id":129,"slug":130,"title":131,"created_at":132},"e41546b8-ba9e-455f-9159-88d4614ad711","openai-codex-plugin-claude-code-zh","OpenAI 把 Codex 放進 Claude Code","2026-04-01T09:21:54.687617+00:00",{"id":134,"slug":135,"title":136,"created_at":137},"96d8e8c8-1edd-475d-9145-b1e7a1b02b65","mcp-explained-from-prompts-to-production-zh","MCP 怎麼把提示詞變工作流","2026-04-01T09:24:39.321274+00:00"]