Smash Notes
Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race, & the return of history

Dwarkesh Podcast

Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race, & the return of history

Chatted with my friend Leopold Aschenbrenner on the trillion dollar nationalized cluster, CCP espionage at AI labs, how unhobblings and scaling can lead to 2027 AGI, dangers of outsourcing clusters to Middle East, leaving OpenAI, and situational awareness. Watch on YouTube . Listen on Apple Podcasts , Spotify , or any other podcast platform. Read the full…

People

Notes

  1. What Is the Key Question for AI Progress? at 10:49

    Leopold Aschenbrenner says the key question for AI progress in the next few years is how hard it is to unlock the test-time compute overhang. Full answer

  2. How Does Automating AI Research Create Superintelligence? at 24:17

    Leopold Aschenbrenner says automating AI research takes the path from AGI to AI vastly smarter than humans within a year or a couple of years. Full answer

  3. When Will the World Notice Superintelligence Is Coming? at 30:41

    Leopold Aschenbrenner says the world still has a few more years of midgame before superintelligence feels obvious. Full answer

  4. What Did Leopold Aschenbrenner's Great-grandmother Live Through? at 37:08

    Leopold Aschenbrenner says his great-grandmother lived through the Nazi era, World War II, and most of her life under the East German communist dictatorship. Full answer

  5. How Much Electricity Does a 100 Gigawatt AI Cluster Use? at 40:47

    Leopold Aschenbrenner says a 100 gigawatt AI cluster uses over 20 percent of United States electricity production. Full answer

  6. Should Only Democracies Develop Artificial General Intelligence? at 52:59

    Leopold Aschenbrenner says a narrow coalition of democracies should be the group developing artificial general intelligence. Full answer

  7. How Easy Is It to Steal Unreleased AI Code? at 57:10

    Leopold Aschenbrenner says stealing important unreleased AI code is extremely easy. DeepMind does not claim that taking the code is hard. Full answer

  8. Why Is a Three-month AI Lead Dangerous? at 1:06:43

    Leopold Aschenbrenner says a neck-and-neck race with only a three-month lead is incredibly dangerous, while a 20 or 30 year technological lead is totally… Full answer

  9. Why Is AI Arms Control Unstable in a Close Race? at 1:20:00

    Leopold Aschenbrenner says an arms control agreement is very unstable when the race is close, because three months ahead is enough to win. Full answer

  10. What Is Leopold Aschenbrenner Waiting to Buy? at 1:25:26

    Dwarkesh Patel says Leopold Aschenbrenner told an influential person he is not going to spend any money on consumption until he is ready to buy galaxies. Full answer

  11. Why Not Wait to Nationalize AI? at 1:46:36

    An early nationalization is what would lock stuff down in time, and Leopold Aschenbrenner expects America to move only once events are already obvious, in a… Full answer

  12. Why Does AI Alignment Look Like a War? at 2:07:03

    Leopold Aschenbrenner says alignment during an intelligence explosion looks like a war in a fog of war, not years of bureaucracy that end in written standards. Full answer

  13. Why Did the Future Fund Collapse? at 2:25:54

    The Future Fund collapsed from one day to the next in November of 22, after it came out that the backer was a giant fraud. Full answer

  14. Why Was Leopold Aschenbrenner Fired From OpenAI? at 2:30:35

    Leopold Aschenbrenner says OpenAI told employees he was fired for leaking. Full answer

  15. Why Is There So Much Drama at OpenAI? at 2:42:48

    Leopold Aschenbrenner says the drama at OpenAI comes from a real belief that OpenAI is building AGI, then a failure to take the implications seriously or keep… Full answer

  16. Why Do Ideas Get Harder to Find? at 2:46:08

    Leopold Aschenbrenner says ideas get harder to find because each next discovery takes a log-scale jump in cumulative research effort. Full answer

  17. Can AI Research Explode Before Robots Do Physical Work? at 2:59:05

    Leopold Aschenbrenner says an AI research explosion can start before robots do ordinary physical work. Full answer

  18. Why Is a Private AI Race Tough for Alignment? at 3:39:53

    Leopold Aschenbrenner says the private world is extremely tough for alignment, even when labs nominally care. Full answer

  19. Is China Impenetrable to Outsiders? at 3:47:13

    Leopold Aschenbrenner says China is impenetrable to him. Full answer

  20. Did Aschenbrenner Publish Knowing China Would Read It? at 3:55:13

    Leopold Aschenbrenner published knowing people in the PRC would read the series. Full answer

  21. Why Did Mormon Fertility Rates Fall? at 4:03:48

    Mormon fertility rates fell once the community grew big enough to live in close contact with ordinary society, Leopold Aschenbrenner says. Full answer

  22. Why Does Capital Still Matter After Superintelligence? at 4:08:10

    Capital still matters after superintelligence because the decade after artificial general intelligence stays wild, and Leopold Aschenbrenner says there is a… Full answer

Related notes

Related episodes

Plays from this note through the rest of the episode.

Note

Weekly highlights

One letter a week.

The notes worth hearing, with the minute they were said.