Deep dive into Google Discover: personalization outweighs topic relevance in feed ranking
gaganghotra_ · x · 2026-08-19
1492.vision published a deep dive into Google Discover's recommendation algorithm, based on a recent job posting and two years of real-world feed observation. The research identifies the system's four pillars: retrieval, prediction, ranking, and embedding.
Key Findings:
- Two Independent Axes: Approximately nine prediction scores collapse into two dimensions: Attention (article property) and Engagement (reader-content pair property).
- Personalization Dominance: Reader-source affinity learned by the model drives amplification about 8x more than the explicit 'Follow' signal when topic potential is equal.
- User Embeddings: Users are represented by families of named embeddings (e.g., Astria, Hobbes, Deep Now) to enable precise matching.
More from Research
- Google's Willow chip cuts quantum errors 40x, raising Q-Day concerns for Bitcoin — jamestagg · 2026-10-16
- A Cookbook for Machine Learning: compiling ML goals into tractable optimization — goyal__pramod · 2026-10-16
- Transluce vets: a year of neuron autointerp work still can't explain complex behaviors — aryaman2020 · 2026-10-16
- Transformer trained on flying bat recordings yields hippocampus 'digital twin' revealing hidden structure — patrickmineault · 2026-10-16
- One year of Recursive Language Models: GPT-5-mini beats GPT-5 by 110% on 132k-token tasks, cheaper too — a1zhang · 2026-10-16
- MIT PhD Sharut Gupta joins DeepMind; Isola hails her representational alignment defense talk — phillip_isola · 2026-10-16