Shane Legg (DeepMind Founder) - 2028 AGI, New Architectures, Aligning Superhuman Models
Dwarkesh Podcast
Limitations of Language Models in Measuring Human Cognition
Large language models used in research papers have limitations in measuring aspects of human cognition. They overlook abilities like comprehending streaming video and having episodic memory. These models primarily focus on benchmarks like MMLEU, but fail to capture the holistic cognitive capabilities of humans.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
