Me, Myself, and AI Challenging the Average With Open-Source AI: Hugging Face’s Thomas Wolf
Sep 16, 2025
Thomas Wolf, cofounder and chief science officer of Hugging Face, shares insights into making AI accessible for everyone. He discusses the tensions between open-source and closed-source models, emphasizing collaboration and innovation. The conversation dives into the evolution of hardware, highlighting shifts from CPUs to GPUs and the new metrics in chip design. Wolf also stresses the importance of engaging young minds with technology, promoting creativity and critical thinking, while envisioning a future where AI supports education.
AI Snips
Chapters
Transcript
Episode notes
Open Models Enable Local Control
- Hugging Face hosts open-source AI models, datasets, and low-code spaces used by roughly 10 million builders.
- Users can download models to run locally for privacy, speed, or specialized hardware needs.
Long-Term Edge Of Open Source
- Open source often wins in the long run but closed-source can iterate faster short-term.
- Both will coexist because each has trade-offs in speed, capital, and accessibility.
Share Full Recipes To Scale Impact
- Publish models with full recipes, data, and training details to accelerate community adoption.
- Share detailed docs and reproducible steps so others can build on your work.