✨
AI Summary
- Stefano Ermon discusses Inception's diffusion architecture for AI inference, addressing hardware and latency bottlenecks.
- He explains diffusion models' advantages over autoregressive LLMs for text and code generation.
- Parallel token generation in diffusion offers superior inference scaling for AI applications.
Guests on This Episode
IC
Inception Co
1 podcast appearance
More from No Priors: AI, Machine Learning, Tech, & Startups
