[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-a-new-method-teaches-robot-ai-what-memories-matter":10,"sections":34},{"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":24,"tags":25,"sources":29,"feedback":33,"feedback_at":22,"cost_usd":33,"total_tokens":33},9694,"a-new-method-teaches-robot-ai-what-memories-matter","A New Method Teaches Robot AI What Memories Matter","A new training method called Divide-and-Remember helps robot AI models recall only the history details that matter for long, multi-step tasks.","Researchers have a new way to make robot AI models remember only what they need to finish long, multi-step tasks.\n\nVision-language-action models, the systems that let robots turn a camera feed and a text instruction into motion, tend to fail when the current camera frame does not tell the whole story. If a robot needs to recall that it already picked up a cup three steps ago, a model with no memory just guesses. Existing fixes decide what to keep by fixed rules, such as saving frames with big pixel changes, and researchers say that approach produces inconsistent results across tasks. A new paper reframes the problem as optimization: the ideal memory keeps only the history information relevant to the current action that is not already visible in the current frame. The proposed method, Divide-and-Remember, recursively breaks a long history into small groups, each narrowed down by a lightweight, shared selector trained end to end, so it scales to long histories without adding much compute. On RoboMME, a benchmark of 16 long-horizon manipulation tasks testing when, where, what, and how to act, the method hit state-of-the-art average success using a memory budget of just 64 tokens, with matching gains in real robot tests.\n\nThat 64-token budget is the real story here. Robot policies have been stuck choosing between ignoring history entirely or dragging along bloated context windows that slow everything down. If a smarter selection rule, not more compute, is what was holding long-horizon robot tasks back, that changes where research money should go next.\n\nStill, 16 curated benchmark tasks and a handful of real-robot trials are a long way from a warehouse floor, and this is one lab's result awaiting outside replication.","[\"ai\",\"robotics\",\"research\",\"memory\"]","2026-10-02T04:00:00.000Z","2026-10-03T06:40:10.830Z","2026-10-03T06:40:16.438Z","published",null,[],"ai",[24,26,27,28],"robotics","research","memory",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.00982",0,{"sections":35},[36,39,43,47,52,56,60,65,70,75,80,85,90,95],{"name":37,"slug":24,"count":38,"latest_published_at":18},"AI",5976,{"name":40,"slug":41,"count":42,"latest_published_at":18},"Security","security",842,{"name":44,"slug":45,"count":46,"latest_published_at":18},"Policy","policy",438,{"name":48,"slug":49,"count":50,"latest_published_at":51},"Deals","deals",317,"2026-10-01T22:00:00.000Z",{"name":53,"slug":54,"count":55,"latest_published_at":18},"Hardware","hardware",199,{"name":57,"slug":58,"count":59,"latest_published_at":18},"Science","science",173,{"name":61,"slug":62,"count":63,"latest_published_at":64},"Consumer Tech","consumer-tech",155,"2026-10-01T19:54:10.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Dev Tools","dev-tools",96,"2026-10-01T16:57:03.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Software","software",93,"2026-09-30T21:41:11.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Startups","startups",90,"2026-10-01T21:55:22.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Gaming","gaming",53,"2026-10-02T02:50:39.000Z",{"name":86,"slug":87,"count":88,"latest_published_at":89},"General","general",50,"2026-09-30T21:37:54.000Z",{"name":91,"slug":92,"count":93,"latest_published_at":94},"Reviews","reviews",31,"2026-09-28T14:31:34.000Z",{"name":96,"slug":97,"count":98,"latest_published_at":99},"How-To","how-to",7,"2026-10-01T09:00:00.000Z"]