Llama 2 lead and Llama 3 post-training lead Thomas Scialom of Meta/FAIR, on the Chinchilla trap, why Synthetic Data and RLHF works, and how Llama4's focus on Agents will lead us to Open Source AGI.| www.latent.space
How we can use AI for as a "partner in thought", losing faith in long context windows for improved reasoning, and why we should stop anthropomorphizing LLMs| www.latent.space