[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-wm-sar-cuts-the-context-ai-agents-need-to-fix-planning-errors":10,"sections":40},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":30,"tags":31,"sources":35,"feedback":39,"feedback_at":22,"cost_usd":39,"total_tokens":39},3406,"wm-sar-cuts-the-context-ai-agents-need-to-fix-planning-errors","WM-SAR Cuts the Context AI Agents Need to Fix Planning Errors","A new technique repairs broken AI planning graphs by isolating error-amplifying nodes, outperforming brute-force LLM correction on tight token budgets.","A research paper proposes a smarter way to fix mistakes inside long-running AI agent workflows — without replanning everything from scratch.\n\nAs AI agents grow from simple tool chains into persistent workflows spanning thousands of steps, errors no longer happen in isolation — they propagate through large planning graphs. The standard fix is expensive: rescan the whole graph, feed a large language model a giant context window, and hope it finds the problem. The authors of WM-SAR (World-Model Subgraph Amplification Repair) argue this approach is both computationally wasteful and counterproductive, because flooding an LLM with irrelevant graph nodes actively degrades its repair accuracy. Their method works backward from the pattern of error amplification — finding the nodes and edges that keep spreading a fault — and hands the LLM only that compact causal subgraph. In graph simulations and live LLM repair experiments, WM-SAR achieved near-full graph stabilization while using a fraction of the token budget that brute-force correctors required.\n\nThis matters because context cost is increasingly the ceiling on what agent systems can do in practice. Techniques that deliver the same or better correction quality with fewer tokens don't just save money — they make longer, more complex workflows feasible on models with bounded context windows. It also shifts the design question away from \"how big a context can we throw at this\" toward \"how precisely can we identify what actually needs fixing.\"\n\nThe broader pattern here echoes work on sparse attention and retrieval-augmented generation: the field keeps rediscovering that giving a model less, more relevant information tends to beat giving it everything and hoping for the best.","[\"ai\",\"agents\",\"llm\",\"research\"]","2026-07-03T04:00:00.000Z","2026-07-03T05:17:30.310Z","2026-07-03T05:17:33.097Z","published",null,[24],{"id":25,"reviewer":26,"round":27,"reason":28,"status":29},"editor-r1","editor",1,"The headline and dek read as working placeholders — 'Fix the Root Cause, Not the Symptom' is too vague and informal for publication; rewrite the headline to name the technique and state the concrete claim (e.g. fewer tokens, smaller models), matching the specificity of the dek.","resolved","ai",[30,32,33,34],"agents","llm","research",[36],{"name":37,"url":38},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.01767",0,{"sections":41},[42,46,51,56,61,66,71,76,81,86,91,95,100,105],{"name":43,"slug":30,"count":44,"latest_published_at":45},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":47,"slug":48,"count":49,"latest_published_at":50},"Security","security",294,"2026-07-15T19:59:48.000Z",{"name":52,"slug":53,"count":54,"latest_published_at":55},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":57,"slug":58,"count":59,"latest_published_at":60},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":62,"slug":63,"count":64,"latest_published_at":65},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":67,"slug":68,"count":69,"latest_published_at":70},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":72,"slug":73,"count":74,"latest_published_at":75},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Dev Tools","dev-tools",59,"2026-07-07T04:00:00.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":90},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":92,"slug":93,"count":89,"latest_published_at":94},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":96,"slug":97,"count":98,"latest_published_at":99},"General","general",29,"2026-07-10T22:28:58.000Z",{"name":101,"slug":102,"count":103,"latest_published_at":104},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":106,"slug":107,"count":108,"latest_published_at":109},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]