EASY: Multi-LLM Protocol for Local and Cloud AI - MinionS

Learn about MinionS, a cost-efficient protocol that achieves proprietary cloud LLM performance at a fraction of the cost by integrating local PC computing with open source language models. This 25-minute video from Discover AI presents Stanford University research on collaboration between on-device and cloud language models. Explore how this innovative approach can significantly reduce costs while maintaining high-quality AI performance. The research is authored by Avanika Narayan, Dan Biderman, Sabri Eyuboglu, Avner May, Scott Linderman, James Zou, and Christopher Ré from Stanford University's Departments of Computer Science, Statistics, Biomedical Data Science, and the Wu Tsai Neurosciences Institute, in collaboration with Together AI.