Developer Shows Embeddings Beat LLMs for Emoji Search at Near-Zero Cost

Responding to Linear's LLM-powered emoji picker, Max Leiter argues most classification tasks are better served by embeddings, demonstrating near-zero-cost semantic emoji search with a local bge-small model.

2026-09-27 ~ 2026-09-27 · 2 related posts

• You don't need an LLM for good emoji search: embeddings run faster and cost $0 — maxleiter · 2026-09-27
• Semantic emoji search with on-device bge-small beats LLM at $0 per search — maxleiter · 2026-09-27