Sunday, Aug 16, 2026

Top stories

GenRec: Netflix's LLM-native recommendation ranker

Netflix replaced its recommendation ranker's traditional feature-engineering stack with GenRec, an LLM-based system that scores titles from natural-language "context" — verbalized user history and catalog metadata — instead of thousands of hand-crafted features. A two-phase pipeline first adapts a foundation model on Netflix data, then post-trains it for ranking with a catalog-constrained scoring head (so it can't recommend a title that doesn't exist) and reward signals tuned for long-term satisfaction, not just click-through. It runs prefill-only on vLLM, scoring the whole catalog in a single forward pass without autoregressive decoding, and beat the production baseline offline (+1.6% Mean Reciprocal Rank) using roughly 40x fewer labeled examples, then held up with statistically significant gains in an online A/B test across about 10% of Netflix's traffic.

Sources & depth

CORS Chat

Simon Willison built CORS Chat, a browser-based chat UI, to test Qwen 3.8 27B running locally in LM Studio across an M5 MacBook Pro and an NVIDIA DGX Spark. It talks to any OpenAI-Responses-compatible endpoint that sets CORS headers — confirmed working against LM Studio's --cors flag and OpenRouter — so it's useful for juggling several concurrent chats against different local or hosted models, with conversations saved in the browser and exportable as JSON. The one distinctive touch: it renders SVG output progressively as tokens stream in, rather than waiting for the full response.

Sources & depth
  • simonwillison.netCORS Chat2026-08-15 20:19 IST
Also today