[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-new-ai-optimizer-beats-rivals-at-fixing-data-pipelines":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},11308,"new-ai-optimizer-beats-rivals-at-fixing-data-pipelines","New AI Optimizer Beats Rivals at Fixing Data Pipelines","CoCoEvolve, an AI optimizer that reviews a system's full history instead of just its latest failure, posted the best scores yet on several data benchmarks.","A new evolutionary optimizer just posted some of the biggest benchmark jumps seen in AI system tuning.\n\nResearchers introduced CoCoEvolve (Cortex Code Evolve), a system that mutates AI pipelines by pulling evidence from an entire evolutionary history and probing the live system, rather than reacting to a single isolated failure. They tested it on three kinds of production AI systems: data agents, LLM-powered data pipelines, and data-engineering workflows. On the DABStep benchmark, it hit 92.8% held-out hard accuracy, well past Meta-Harness's 78.4% and GEPA's 30.4% under matched conditions. On data-eng-bench, it pushed the test-set solve rate from a 59.1% baseline to 77.4%, beating the next-best method's 73.6%. It also topped GEPA by 11 points on HotpotQA.\n\nThe pitch here is about where optimization breaks down, not just how high the scores go. Most prior optimizers look only at a candidate's most recent failure when deciding how to fix it, which is fine for a single prompt but falls apart on production systems full of hidden, overlapping interactions. Letting the mutator probe the live system and draw on the full history instead of the latest frontier is the actual mechanism behind these gains, and the 62-point gap over GEPA on DABStep suggests the frontier-only approach has a real ceiling, not just room for tuning.\n\nThe paper's own ablations show every component contributes something, which is the detail worth checking before anyone calls this a universal fix rather than a strong result on three specific benchmark families.","[\"ai\",\"evolutionary-optimization\",\"data-pipelines\",\"benchmarks\"]","2026-10-12T04:00:00.000Z","2026-10-12T07:15:24.916Z","2026-10-12T07:15:30.078Z","published",null,[],"ai",[24,26,27,28],"evolutionary-optimization","data-pipelines","benchmarks",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.12794",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"]