1 article tagged with this topic
Recommenders shift to unified sequence modeling. NVIDIA and Meta ship Meta's 2024 HSTU as turnkey inference — feeds may enter an LLM-like era.
