Dwarkesh Podcast

Leopold Aschenbrenner - China/US Super Intelligence Race, 2027 AGI, & The Return of History

916 snips
Jun 4, 2024
In a riveting discussion, Leopold Aschenbrenner, a valedictorian from Columbia and ex-member of OpenAI's super alignment team, explores the fierce AI race between the U.S. and China. He warns against the perils of outsourcing AI infrastructure and lays out ambitious predictions of achieving AGI by 2027. Leopold shares insights on the geopolitical implications of AI, the importance of safeguarding technology from espionage, and the lessons we can learn from history to enhance national security in this fast-evolving landscape.
Ask episode
AI Snips
Chapters
Books
Transcript
Episode notes
INSIGHT

AI's Industrial Nature

  • AI development is an industrial process requiring massive infrastructure like clusters and power plants.
  • Training compute for large AI systems has grown exponentially for a decade.
INSIGHT

Trillion-Dollar Cluster

  • Leopold projects a "trillion-dollar cluster" by 2030, consuming over 20% of US electricity.
  • He believes a 10-gigawatt cluster, the size of some US states' power consumption, is feasible by 2028.
INSIGHT

AGI Timeline

  • Leopold predicts AGI around the 10-gigawatt cluster range, possibly by 2027.
  • He believes models will be smarter than most college grads by 2025/26 and expert-level by 2027/28.
Get the Snipd Podcast app to discover more snips from this episode
Get the app