The ODSC Ai X Podcast features Eric Xing discussing a shift in artificial intelligence from mere task execution to creating systems that understand and adapt. This transformation involves three foundational concepts: world models, agentic AI, and AI-driven biology. World models simulate not only physical environments but also abstract concepts like human thought. Xing emphasizes that these models aim to reason and prepare agents for decision-making in uncertain contexts, rather than generate flashy outputs. This new approach is fundamental for developing general intelligence in AI systems.
A world model is a simulator of all possibilities in the real world - conditioned on the past and what you do with it.
The goal is to reason about what might happen before acting in the real world.
Collection
[
|
...
]