From Single AI to Multi AI Systems
summary
Argues that real-world systems move from single models to cooperating and competing agents and gives concrete wins: a robot-soccer grand challenge aiming to beat the World Cup winner by 2050, deployed BDI agents that run adherence pilots for diabetes and telecom lifecycle management, city traffic lights trained with decentralized reinforcement learning that show 30% traffic reduction and 10% pollution cuts in trials, population-scale micro‑simulations for pandemic planning, and industry needs like contract net protocols and task allocation rather than naive LLM-wrapped agents.
AI-written summary · about
chapters (12)
- 0:00Introduction and welcome
- 1:24From single to multi AI systems
- 4:56AI Grand Challenge in robot soccer
- 8:33Healthcare agent systems for chronic disease
- 9:37Traffic light coordination with agents
- 11:43Societal simulations with intentional agents
- 13:53Applications in finance and trading
- 14:36What defines a multi-agent system
- 17:03Team performance modeling in agents
- 18:27Dos and don'ts in multi-agent systems
- 22:36Quantum computing and AI future
- 23:44Q&A on quantum computing
related talks
50mScalable Innovation: Exploring Multi-Agent Systems
27mMulti-Agent AI: Transforming Business Operations and Value Creation
32mThe Road to General AI
26mAI Agents and the Future of Human-AI Collaborative Intelligence
34mAI Agents: Transforming Autonomous Decision-Making
59mThe evolution in AI thinking and products of the next decade
27mHow to Build AI Agents and Keep Your Sanity
45mIllusion of Life