You don't need an LLM for good emoji search: embeddings run faster and cost $0
maxleiter · x · 2026-09-27
Max Leiter pushes back on the hype around using Jev for classification tasks like Linear's emoji picker, arguing most such use cases can be matched or beaten with plain embeddings.
- He built semantic emoji search two years ago without any LLM: bge-small-en-v1.5 runs in-browser (37 MB, cached after first download), $0 per search
- His demo compares three approaches live with latency and cost: on-device bge-small, text-embedding-3-large + Postgres pgvector, and Jev via Vercel AI Gateway
- His rule of thumb: use Jev when you need actual LLM reasoning behind a decision; for semantic matching and retrieval, embeddings are faster, cheaper, and occasionally better
- He concedes embeddings require more setup (a database plus an embedding provider), while Jev is easier to wire up
Related event: Developer Shows Embeddings Beat LLMs for Emoji Search at Near-Zero Cost(2 posts)→
More from coding & agent
- 100-Paper Review: Most Prompt-Injection Defenses Fail Against Adaptive Attacks — Ok-Lab-7347 · 2026-09-27
- Procedural Three.js Running Horse in One HTML File, via 67 Minutes of AI Thinking — Ok_Bag_7550 · 2026-09-27
- Motion design in pure code: Opus writes Canvas renderer and a Node synth soundtrack — mishig25 · 2026-09-27
- Antigravity Rewind: Open-Source One-Click Surgical Undo for Google Antigravity — Imbourass · 2026-09-27
- Open-Source Fooocus-Style Studio for Qwen-Image 2.1, Built Almost Entirely by Claude — Repulsive_Junket7919 · 2026-09-27
- TailwindCSS MCP server adds class lookup, CSS conversion and palette tools — modelcontextprotocol · 2026-09-27