Frontiers of Agentic AI
summary
An open-source platform AG2 (fka AutoGen) grew to tens of thousands of users and top GitHub rank within months and powers projects from Google agent platforms to Walmart recommendations; small tools like Flammo and research pipelines automate hyperparameter tuning and paper-to-code. Multi-agent systems like Denario automate scientific workflows and have produced hundreds of papers with at least one accepted; MassGen (three months old) runs parallel agents with shared hubs, tools and adaptive coordination and raises a GBQA physics score from 38% to 44% by having agents converge (3 of 4) on a correct solution.
AI-written summary · about
chapters (15)
- 0:00Agentic AI journey
- 1:26Birth of Autogen AG2
- 3:12Agentic AI in industries
- 4:35Agent evolution dimensions
- 4:55Interface advances
- 8:44Open Agents networks
- 10:11Agent coding power
- 11:59Dinaric automating discovery
- 14:06Research to real-world code
- 15:30Multi-agent architecture
- 16:32Multi-agent scaling
- 18:59MasterGen collaboration
- 21:53Self-correcting agents
- 22:14Live multi-agent demo
- 24:00Conclusion
related talks
45mBuild Agentic AI Apps with the Autogen framework
34mBuilding Multi-Agent Generative-AI Applications with AutoGen
01h 27mBuilding Vertical Agent with AG2 AgentOS
10mKeynote: Next-Gen AI Applications via Multi-Agent Conversation
33mAutoGen: A Multi-Agent AI Framework
20mGenAI agents - Evolution of automation
26mAI powered automation & multi-agent orchestration- 10mKeynote: Next-Gen AI Applications via Multi-Agent Conversation