[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-ai-agents-struggle-to-know-when-to-quit":10,"sections":35},{"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":24,"persona_id":22,"persona_name":22,"section":25,"tags":26,"sources":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},2436,"ai-agents-struggle-to-know-when-to-quit","AI Agents Struggle to Know When to Quit","A study of 13 LLM agent systems finds that knowing when to stop — not just whether to stop — is the harder problem in autonomous task completion.","Most AI agents have a quitting problem — and it cuts both ways.\n\nResearchers tested 13 large language model agent systems and two agent scaffolds across more than 28,000 tasks spanning web shopping, terminal environments, and question answering. The core question: can agents recognize when a goal is unachievable and stop trying? The answer is complicated. Some agents never abstained when they should have; others eventually quit, but only after burning through unnecessary steps. The trickiest cases were tasks that looked feasible on the surface until the environment revealed a dead end — say, a search that returns no valid match.\n\nThe timing problem matters more than it might seem. An agent that keeps calling tools on an unsolvable task is not just wasting compute — it is actively doing the wrong thing, potentially making irreversible moves or surfacing misleading partial results to users who assume the agent would have stopped if nothing was working. The researchers also found a counterintuitive result: larger, more capable models sometimes abstained *less* timely than smaller ones, suggesting that raw capability does not fix the calibration gap.\n\nTo address this, the team introduces CONVOLVE, a context engineering method that distills past interaction trajectories into reusable stopping rules without retraining the underlying model. On the WebShop benchmark, it more than doubled Llama-3.3-70B's timely abstention recall rate, from 26.7 to 57.4. That is a meaningful jump from a prompt-level intervention — though it also illustrates how much ground agents still need to cover before autonomous tool use can be trusted in production. The broader industry push toward agentic AI has largely focused on expanding what agents *can* do; this research is a useful reminder that knowing what not to do is just as important.","[\"ai\",\"agents\",\"llm\",\"research\"]","2026-06-30T04:00:00.000Z","2026-06-30T05:17:59.799Z","2026-06-30T05:18:09.258Z","published",null,[],"https:\u002F\u002Fcdn.xyz.onl\u002Farticle-images\u002Fai-agents-struggle-to-know-when-to-quit.webp","ai",[25,27,28,29],"agents","llm","research",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2606.28733",0,{"sections":36},[37,41,46,51,56,61,66,71,76,81,86,90,95,100],{"name":38,"slug":25,"count":39,"latest_published_at":40},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":42,"slug":43,"count":44,"latest_published_at":45},"Security","security",294,"2026-07-15T19:59:48.000Z",{"name":47,"slug":48,"count":49,"latest_published_at":50},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":52,"slug":53,"count":54,"latest_published_at":55},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":57,"slug":58,"count":59,"latest_published_at":60},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":62,"slug":63,"count":64,"latest_published_at":65},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":67,"slug":68,"count":69,"latest_published_at":70},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":72,"slug":73,"count":74,"latest_published_at":75},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Dev Tools","dev-tools",59,"2026-07-07T04:00:00.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":87,"slug":88,"count":84,"latest_published_at":89},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":91,"slug":92,"count":93,"latest_published_at":94},"General","general",29,"2026-07-10T22:28:58.000Z",{"name":96,"slug":97,"count":98,"latest_published_at":99},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":101,"slug":102,"count":103,"latest_published_at":104},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]