Dexa/Dwarkesh Patel

Optimistic Research Directions

Shane Legg discusses various research directions at DeepMind, highlighting the promise of Deliberative Dialogue in scaling alignment for increasingly powerful systems. He also emphasizes the importance of improving episodic memory and assistant thinking, while addressing the need for models to be more concrete in applications that require factual accuracy.

In this clip
From this podcast
Dwarkesh Podcast
Shane Legg (DeepMind Founder) - 2028 AGI, New Architectures, Aligning Superhuman Models
Related Questions
- What do you think about the potential for Large Language Models (LLMs) to scale to Artificial General Intelligence (AGI) in the context of the episodes Why Your RAG Pipeline Is Broken, and How to Fix It with Jason Liu - 709 and Trusting Intuition in AI?