[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-new-robot-ai-model-plans-32-times-faster-than-rivals":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},11309,"new-robot-ai-model-plans-32-times-faster-than-rivals","New Robot AI Model Plans 32 Times Faster Than Rivals","LeWAM merges perception and action planning into one AI model, tripling robot task success rates and planning far faster than prior approaches.","A new AI system lets robots predict the consequences of their actions and decide what to do next using a single model, instead of two bolted-together ones.\n\nResearchers describe LeWAM, a joint-embedding predictive architecture, or JEPA, that learns a compact model of the world and how actions change it, trained in one pass rather than in separate stages. Earlier JEPA-based robot planners trained perception and action-prediction separately, then added a control policy on top, leaving a gap between what the model understood and what it could act on. LeWAM closes that gap by training next-state prediction and action prediction together, with a constraint on the latent space that lets the same model either run a policy directly or simulate many possible actions to pick the best one. On contact-rich manipulation tasks, success rate rose from 28.6% to 89.7%; on long-horizon tasks, from 10.9% to 46.2%; and planning ran 32.3 times faster than prior JEPA approaches.\n\nRobot planning has long been bottlenecked by slow, decoupled pipelines: understand the world, then separately figure out what to do about it. A single model that does both, fast enough to weigh options in real time, is the kind of unglamorous plumbing fix that moves robotics forward more than another flashy demo.\n\nWorth noting: these numbers come from the paper's own benchmarks against its own baselines, not independent replication, and \"competitive with established methods\" on behavior cloning is a modest claim dressed up as a headline result.","[\"ai\",\"robotics\",\"world-models\",\"jepa\"]","2026-10-12T04:00:00.000Z","2026-10-12T07:18:00.288Z","2026-10-12T07:18:06.360Z","published",null,[],"ai",[24,26,27,28],"robotics","world-models","jepa",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.12818",0,{"sections":35},[36,39,44,49,54,59,64,69,74,79,84,89,94,99],{"name":37,"slug":24,"count":38,"latest_published_at":18},"AI",6966,{"name":40,"slug":41,"count":42,"latest_published_at":43},"Security","security",952,"2026-10-12T04:54:12.000Z",{"name":45,"slug":46,"count":47,"latest_published_at":48},"Policy","policy",496,"2026-10-11T20:06:43.000Z",{"name":50,"slug":51,"count":52,"latest_published_at":53},"Deals","deals",488,"2026-10-12T10:32:15.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":58},"Hardware","hardware",234,"2026-10-11T17:20:00.000Z",{"name":60,"slug":61,"count":62,"latest_published_at":63},"Science","science",196,"2026-10-11T15:34:40.000Z",{"name":65,"slug":66,"count":67,"latest_published_at":68},"Consumer Tech","consumer-tech",191,"2026-10-12T11:10:16.000Z",{"name":70,"slug":71,"count":72,"latest_published_at":73},"Startups","startups",120,"2026-10-09T17:02:31.000Z",{"name":75,"slug":76,"count":77,"latest_published_at":78},"Software","software",115,"2026-10-11T07:34:45.000Z",{"name":80,"slug":81,"count":82,"latest_published_at":83},"Dev Tools","dev-tools",111,"2026-10-12T11:22:24.000Z",{"name":85,"slug":86,"count":87,"latest_published_at":88},"General","general",70,"2026-10-11T14:15:00.000Z",{"name":90,"slug":91,"count":92,"latest_published_at":93},"Gaming","gaming",59,"2026-10-09T11:43:43.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"Reviews","reviews",34,"2026-10-08T14:00:22.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"How-To","how-to",8,"2026-10-05T09:00:00.000Z"]