Could LLMs Be The Route To Superintelligence? — With Mustafa Suleyman

Microsoft's push towards humanist superintelligence is explored with Mustafa Suleyman, CEO of Microsoft AI, who defines superintelligence and explains its potential benefits, such as cheap, abundant medical and legal expertise. Suleyman argues for a verticalized approach to AI development, focusing on domain-specific models to ensure safety and control. He addresses concerns about the limitations of current methods, particularly LLMs, but expresses optimism about ongoing progress through innovations in areas like recurrency and memory. The conversation also covers Microsoft's new agreement with OpenAI, which allows Microsoft to pursue AGI independently, and the economic viability of building proprietary models in a commoditizing market. Finally, the discussion touches on the societal implications of AI companions and the evolving definition of what it means to be human.