Multi Agent Architecture
topic on 3 shows · 4 statements across 3 episodes
Latent Space 
A Product Market Fit Show 
Big Technology
4 statements about Multi Agent Architecture, every show
Xanthos: Multi-agent AI needs giant models for orchestration, smaller models below
“At the top level, you usually have the most capable model, usually the most expensive also, and the biggest model, and then the underlying tasks can be performed by maybe either fast closed source models, or even open source models that you posturing for the t…”
Shah: Early multi-agent AI systems got stuck endlessly thanking each other
“When we first started deploying this multi-agent, multi-step architecture, we would get caught up in something called a gratitude loop, and the gratitude loop was that the agents would just pat each other on the back and say, great job, but they would not go t…”
Davis: Multi-agent research consumes 15x baseline tokens versus 4x for single-agent
“So based off of a basic conversation, a single agent architecture for research is around four X, the number of tokens needed to achieve a research output. When you use multi-agent architectures, it's actually 15 X the number of tokens.”
Davis: Multi-Agent Systems Struggle in Coding Due to Strong Sub-Task Coupling
“When using a multi-agent architecture versus a single agent for coding. And the main issue is that these are dependent upon each other. They're strongly coupled in the sense that when I take an action, that action then is impacts all the other sub-agents if it…”