Eugene Yan proposes a shift away from purely chat-based interfaces for interacting with Large Language Models (LLMs). He argues that users often prefer visual and interactive elements like clicks and filters, similar to existing e-commerce and recommendation platforms. Yan suggests that LLMs could leverage user context, such as past behavior and preferences, to provide more personalized responses without requiring extensive chat input. His prototype demonstrates a system that combines recommendation techniques with LLMs, using item embeddings and a product graph to retrieve and rank relevant items before offering LLM-assisted chat for further refinement. AI
IMPACT Suggests LLM interfaces could evolve beyond chat to incorporate user context and visual interactions, potentially improving usability.
RANK_REASON Opinion piece discussing user experience for LLMs.
- Eugene Yan
- FastAPI
- Google Maps
- Javascript
- Jinja Template Engine
- LLMs
- Midjourney
- Netflix
- Python
- UCSD Book Graph
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →