Meta's GenRecs Tokenization lifts ad conversions +0.35% with semantic vocabulary
Meta_Engineers · x · 2026-10-30
Meta details GenRecs Tokenization: instead of learning each ad one ID at a time (new ads start with no signal), every ad, product, Page and pixel is represented with tokens from a stable semantic vocabulary, so a newly launched trail shoe inherits knowledge of last season's. Paired with the Generative Ads Model (GEM) and Adaptive Ranking Model, it delivered a +0.35% conversion lift.
More from Infra
- llmman: run Claude Code, Codex and other agents on any local model with one command — weskambale · 2026-10-30
- How Instagram cut feed load time by 50% and Discord stores 100B+ messages — _jaydeepkarale · 2026-10-30
- YC-backed VirajAero launches turbine that powers datacenters with 40% more power from same fuel — Scobleizer · 2026-10-30
- Proof of Agent rolls out 15-minute checkpoints anchored to Solana and Base, capping validation costs — adamamcbride · 2026-10-30
- Pooling $8K for a shared local LLM rig takes 3.3 years to beat a $20/mo ChatGPT sub — Any-Blacksmith-2054 · 2026-10-30
- 10 coding agents tested on local Qwen 3.8 27B: Pi wins on efficiency at 9.2k tokens per task — latent_lurker8587 · 2026-10-30