[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-panorama-speeds-up-vector-search-by-up-to-289x":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},4139,"panorama-speeds-up-vector-search-by-up-to-289x","PANORAMA Speeds Up Vector Search by Up to 28.9x","A new refinement technique for approximate nearest-neighbor search cuts query time by pruning candidates early, and it's already in the FAISS library.","A research technique called PANORAMA has been merged into FAISS, the widely used vector search library, promising to dramatically cut the time it takes to find similar items in high-dimensional datasets.\n\nThe bottleneck in most vector search pipelines isn't the initial candidate retrieval — it's verifying which candidates actually win. PANORAMA attacks that step by using PCA to compress where the signal actually lives in a vector, then computing partial distances incrementally. At each step it calculates a lower bound on the full distance and drops any candidate the moment that bound exceeds the current best result. A variance-shaping step keeps it compatible with Product Quantization indexes, which would otherwise break under PCA's uneven energy distribution. The result: end-to-end speedups of up to 28.9x, with gains that scale predictably based on how compressible a dataset's structure is.\n\nVector search is the unglamorous plumbing behind nearly every retrieval-augmented AI system — the thing that finds relevant chunks before a model ever sees them. Faster search at the same recall level means either lower infrastructure costs or more candidates checked in the same time budget, both of which matter at production scale. PANORAMA's integration across FAISS's main index families — IVFPQ\u002FFlat, HNSW, and Refine — means teams don't need to swap out their stack to benefit.\n\nSpeedups that scale with dataset structure are a harder sell on unstructured or low-spectral-decay corpora, so the 28.9x headline should be read as a ceiling, not a floor.","[\"vector-search\",\"machine-learning\",\"ai\",\"open-source\"]","2026-07-07T04:00:00.000Z","2026-07-07T18:02:30.771Z","2026-07-07T18:02:33.751Z","published",null,[],"ai",[26,27,24,28],"vector-search","machine-learning","open-source",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.00566",0,{"sections":35},[36,40,45,50,55,60,65,70,75,79,84,88,93,98],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":41,"slug":42,"count":43,"latest_published_at":44},"Security","security",294,"2026-07-15T19:59:48.000Z",{"name":46,"slug":47,"count":48,"latest_published_at":49},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":59},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":61,"slug":62,"count":63,"latest_published_at":64},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":18},"Dev Tools","dev-tools",59,{"name":80,"slug":81,"count":82,"latest_published_at":83},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":85,"slug":86,"count":82,"latest_published_at":87},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":89,"slug":90,"count":91,"latest_published_at":92},"General","general",29,"2026-07-10T22:28:58.000Z",{"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"]