AI Inference Evolution
George discusses the imbalance between training and inference in AI, advocating for more focus on inference. He explores the potential of on-device fine-tuning for models like those used at comma. The conversation delves into privacy concerns and the idea of individualized intelligence clusters with stringent security measures.In this clip
From this podcast
Latent Space - The AI Engineer Podcast (Video Podcast)
Ep 18: Petaflops to the People — with George Hotz of tinycorp
Related Questions
How is inference used in machine learning in the context of the episode AI is more than GenAI and the clip Model Training Insights?
What's holding back self-driving cars as discussed in the episode George Hotz: Hacking the Simulation & Learning to Drive with Neural Nets | Lex Fridman Podcast #132 and the clip Autonomous Driving Insights?
How is inference used in machine learning, as discussed in the episode Dileep George: Brain-Inspired AI | Lex Fridman Podcast #115 and the clip Graphical Models Explained?