Co-Evolution: How the Open Source Intelligence Stack Compounds
summary
Argues that AI progress comes from synchronized updates across models, training frameworks, inference engines, RL systems, hardware and cloud, and that open source is the coordination mechanism making that possible. Cites concrete stack pieces hosted by the foundation: PyTorch, vLLM, Ray and DeepSpeed; gives community scale—12,000+ contributors, 2,000 organizations, people from over 45 countries—and product moves: ExecuTorch becoming part of PyTorch Core and a new Healion DSL for fast kernels. Notes industry speed: thousands of specialized models in production and models released every 5 hours.
AI-written summary · about
chapters (8)
- 0:00PyTorch Europe conference
- 1:49PyTorch Foundation projects
- 3:56Community diversity insights
- 4:57Hardware-software co-evolution
- 7:07PyTorch in AI acceleration
- 8:51ExecuTorch announcement
- 9:33Healion in PyTorch Foundation
- 11:25Call for open collaboration
related talks
16mKeynote: Open Source for AI and AI for Open Source
10mMeet the Community Behind the Open Source AI Stack
11mKeynote: Open Source Stack Empowers the AI Era
17mOpen Innovation: Unlocking AI's Potential
11mOpen Source for the AI Era
25mOpen AI (Two Words): The Only Path Forward for AI
42mHow Open Standards Can Heal the Open Wound of AI
13mKeynote: Introductory Remarks