[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-new-ai-harness-lets-models-edit-images-mid-reasoning":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},9138,"new-ai-harness-lets-models-edit-images-mid-reasoning","New AI Harness Lets Models Edit Images Mid-Reasoning","Researchers unveiled Mosaic, a toolkit for multi-image AI that crops and compares photos, then trained an 8B agent to wield it with reinforcement learning.","A new research harness lets AI models crop, zoom, and compare images instead of just describing them in words.\n\nResearchers introduce Mosaic, a general-purpose multi-image harness that gives multi-modal AI models ten composable image operations to actively construct visual intermediates while reasoning, rather than relying solely on text-based chain-of-thought. They compared five re-representation strategies, some textual and some visual, across existing multi-image benchmarks plus a new benchmark called MosaicBench, built specifically to test fine-grained, grounding-heavy visual reasoning. The results show the better strategy depends heavily on the task: visual re-representation wins on jobs that need precise visual evidence, like hypothesis testing, precision comparisons, and orientation-sensitive reasoning, while tasks about higher-level meaning show smaller, less consistent gains from either approach. Building on that finding, the team trained an 8-billion-parameter model, MosaicAgent-8B, to use Mosaic via reinforcement learning, rewarding only correct answers and proper output format, without any step-by-step demonstrations of how to use the tools.\n\nMost multi-image AI tools today reason entirely in text, describing what's in each photo and comparing the descriptions, which breaks down when the answer hinges on a visual detail no caption fully captures. Mosaic's task-dependent results argue against defaulting to text-only chain-of-thought for visual agents, and the fact that MosaicAgent-8B learned to chain multiple image operations on its own, with no demonstrations and no tool-specific rewards, suggests reinforcement learning can teach genuinely useful tool-use habits rather than just pattern-matched responses.\n\nIt's a reminder that teaching a model to look more carefully, rather than just prompting it to think longer, might be the bigger unlock for visual reasoning.","[\"ai\",\"computer vision\",\"ai agents\",\"research\"]","2026-10-01T04:00:00.000Z","2026-10-01T22:00:43.019Z","2026-10-01T22:00:45.336Z","published",null,[],"ai",[24,26,27,28],"computer vision","ai agents","research",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.39363",0,{"sections":35},[36,39,43,47,52,57,61,66,71,75,80,85,90,95],{"name":37,"slug":24,"count":38,"latest_published_at":18},"AI",5572,{"name":40,"slug":41,"count":42,"latest_published_at":18},"Security","security",815,{"name":44,"slug":45,"count":46,"latest_published_at":18},"Policy","policy",430,{"name":48,"slug":49,"count":50,"latest_published_at":51},"Deals","deals",298,"2026-09-30T21:00:26.000Z",{"name":53,"slug":54,"count":55,"latest_published_at":56},"Hardware","hardware",196,"2026-09-30T13:00:00.000Z",{"name":58,"slug":59,"count":60,"latest_published_at":18},"Science","science",163,{"name":62,"slug":63,"count":64,"latest_published_at":65},"Consumer Tech","consumer-tech",149,"2026-09-30T22:57:11.000Z",{"name":67,"slug":68,"count":69,"latest_published_at":70},"Dev Tools","dev-tools",93,"2026-10-01T02:30:48.000Z",{"name":72,"slug":73,"count":69,"latest_published_at":74},"Software","software","2026-09-30T21:41:11.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Startups","startups",84,"2026-09-30T20:39:09.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Gaming","gaming",51,"2026-09-30T16:24:30.000Z",{"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",6,"2026-06-16T09:00:00.000Z"]