[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-llms-now-patch-streaming-algorithms-in-minutes-not-months":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},5167,"llms-now-patch-streaming-algorithms-in-minutes-not-months","LLMs Now Patch Streaming Algorithms in Minutes, Not Months","ReForge uses an LLM to continuously patch video streaming algorithms as networks change, testing every edit against every past scenario before it ships.","Researchers built an AI system that keeps rewriting the algorithm your video app uses to pick streaming quality, and proves each change won't break anything it already handled.\n\nThe algorithm in question is ABR, or adaptive bitrate streaming - the logic that decides whether your video buffers at 1080p or drops to 480p when your connection wobbles. Engineers used to spend months hand-tuning ABR for a given network, and even then it only worked for the conditions it was built on. A new framework called ReForge puts a large language model in a loop: each round it reads where the current rules are failing, proposes one small tweak to a compact rulebook that routes decisions to a pool of pre-trained policies, then replays that tweak against every network scenario the system has seen so far before accepting it. Tested across nine real network families arriving in sequence - 3G, then 4G, then 5G - a handful of edits per new network lifted the average quality score from 1.23 to 1.74, beating the best single fixed policy at 1.66 and reaching 94% of a theoretical best case.\n\nThat's notable less for the raw numbers than for what's being automated: verification, not just generation. Anyone can ask an LLM to write a heuristic; the harder problem is proving a new tweak doesn't quietly break the networks it already handled, and ReForge's replay-and-check loop is a real attempt at that guardrail. It also generalized to network types it never trained on directly, with one family's score jumping from 0.30 to 0.80 - a sign the approach learns something more transferable than a lookup table.\n\nStreaming algorithms have always been a moving target chasing whatever network conditions exist on release day. Whether an LLM checking its own homework against history is enough to keep up long-term is a question only production traffic, not a nine-family benchmark, can really answer.","[\"ai\",\"streaming\",\"networking\",\"llm-agents\"]","2026-08-18T04:00:00.000Z","2026-08-18T07:56:19.463Z","2026-08-18T07:56:31.271Z","published",null,[],"ai",[24,26,27,28],"streaming","networking","llm-agents",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2608.15138",0,{"sections":35},[36,40,44,49,54,59,64,69,74,78,83,88,93,98],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",3293,"2026-08-20T04:00:00.000Z",{"name":41,"slug":42,"count":43,"latest_published_at":39},"Security","security",435,{"name":45,"slug":46,"count":47,"latest_published_at":48},"Policy","policy",210,"2026-08-19T09:32:27.000Z",{"name":50,"slug":51,"count":52,"latest_published_at":53},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":58},"Hardware","hardware",140,"2026-08-19T18:25:42.000Z",{"name":60,"slug":61,"count":62,"latest_published_at":63},"Consumer Tech","consumer-tech",95,"2026-08-18T16:05:00.000Z",{"name":65,"slug":66,"count":67,"latest_published_at":68},"Science","science",90,"2026-08-19T18:41:02.000Z",{"name":70,"slug":71,"count":72,"latest_published_at":73},"Software","software",73,"2026-08-18T07:51:50.000Z",{"name":75,"slug":76,"count":77,"latest_published_at":18},"Dev Tools","dev-tools",69,{"name":79,"slug":80,"count":81,"latest_published_at":82},"Startups","startups",47,"2026-08-19T19:13:46.000Z",{"name":84,"slug":85,"count":86,"latest_published_at":87},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":89,"slug":90,"count":91,"latest_published_at":92},"General","general",33,"2026-08-18T22:18:13.000Z",{"name":94,"slug":95,"count":96,"latest_published_at":97},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":99,"slug":100,"count":101,"latest_published_at":102},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]