[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-new-framework-trains-ai-routers-and-agents-together":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":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},6542,"new-framework-trains-ai-routers-and-agents-together","New Framework Trains AI Routers and Agents Together","CERA-MoA is a research framework where the router that assigns queries and the AI agents it manages train together instead of in separate stages.","A new academic framework lets teams of AI agents and the system that assigns them work learn from each other instead of training separately.\n\nMixture-of-agents systems split tasks among multiple specialized large language models, with a router deciding which agents handle which queries. Until now, that router and the agents it manages were typically trained in separate phases, so the router couldn't adapt as agents got better or worse at specific tasks. Researchers describe CERA-MoA, a framework where the router and the agents train together using reinforcement learning. It uses a familiarity estimator that reads an agent's mid-layer hidden states to judge whether that agent knows a topic, skipping the cost of running full queries through every agent just to find out.\n\nThat estimator lets CERA-MoA activate only the smallest useful set of agents for a given query, and it steers training data toward the agents best positioned to improve at specific skills. The result, per the paper, is agents that specialize more cleanly and a system that outperforms both fixed-routing and fixed-workflow baselines in the domains tested.\n\nIt's a preprint with results described only as spanning \"various domains,\" so the efficiency and specialization gains are worth watching for independent replication before anyone builds a product on top of them.","[\"ai agents\",\"mixture-of-agents\",\"reinforcement learning\",\"arxiv\"]","2026-09-17T04:00:00.000Z","2026-09-17T23:20:47.149Z","2026-09-17T23:20:59.077Z","published",null,[],"ai",[26,27,28,29],"ai agents","mixture-of-agents","reinforcement learning","arxiv",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.18779",0,{"sections":36},[37,41,45,50,55,59,63,68,73,77,82,87,92,97],{"name":38,"slug":24,"count":39,"latest_published_at":40},"AI",3852,"2026-09-17T08:27:09.000Z",{"name":42,"slug":43,"count":44,"latest_published_at":18},"Security","security",648,{"name":46,"slug":47,"count":48,"latest_published_at":49},"Policy","policy",338,"2026-09-11T04:00:00.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":18},"Hardware","hardware",154,{"name":60,"slug":61,"count":62,"latest_published_at":18},"Science","science",114,{"name":64,"slug":65,"count":66,"latest_published_at":67},"Consumer Tech","consumer-tech",99,"2026-09-09T17:27:33.000Z",{"name":69,"slug":70,"count":71,"latest_published_at":72},"Software","software",75,"2026-09-10T20:41:21.000Z",{"name":74,"slug":75,"count":76,"latest_published_at":18},"Dev Tools","dev-tools",73,{"name":78,"slug":79,"count":80,"latest_published_at":81},"Startups","startups",55,"2026-09-09T23:14:29.000Z",{"name":83,"slug":84,"count":85,"latest_published_at":86},"Gaming","gaming",43,"2026-09-10T12:18:06.000Z",{"name":88,"slug":89,"count":90,"latest_published_at":91},"General","general",41,"2026-09-08T01:57:23.000Z",{"name":93,"slug":94,"count":95,"latest_published_at":96},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":98,"slug":99,"count":100,"latest_published_at":101},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]