"Upstream" with Erik Torenberg cover image

E39: Balaji Srinivasan and Nathan Labenz on AI's Future, Human-AI Symbiosis, and AI Control

"Upstream" with Erik Torenberg

00:00

The Challenges of Robustness in AI and the Concept of Open Weights

The lack of a robust architecture in AI and the challenges in guaranteeing robustness, particularly in language models, are discussed. The concept of open weights as a means of ensuring safety is explored, along with the research being done on proto scaling laws for adversarial robustness. The limitations of current AI models and the potential of powerful AI systems utilizing a mixture of different architectures are also highlighted.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app