Understanding LLM Limitations
Pedro challenges the notion that LLMs are merely stochastic parrots, emphasizing their ability to generalize from vast data rather than just memorizing it. However, he points out a significant gap in their understanding of common sense, which even a toddler possesses. The discussion also highlights a deep schism in the AI community regarding the roles of language and vision in advancing AI capabilities.In this clip
From this podcast
AI + a16z
The Best Way to Achieve AGI Is to Invent It
Related Questions