Artificial Intelligence Masterclass My P(DOOM) is now 12.70%! Down from 30%! Bayesian Networks and Wisdom of the Crowd
Jun 10, 2025
Dive into the intriguing world of Bayesian networks as they help reduce the Probability of Doom through visual decision-making tools. The risks of uncontrollable superintelligence are explored, with a focus on essential criteria that could signal danger. Additionally, the complexities surrounding artificial superintelligence and its alignment with human goals are discussed, backed by audience survey insights. This thought-provoking conversation sheds light on the future challenges posed by advanced AI.
AI Snips
Chapters
Transcript
Episode notes
Revised P(DOOM)
- David Shapiro revised his P(DOOM) to 12.7% using a Bayesian network.
- This framework assesses the probability of doom based on four factors.
Framework Development
- Shapiro developed the framework through debates and discussions.
- He specifically mentions a valuable conversation with Igor.
Wisdom of the Crowd
- Shapiro used "wisdom of the crowd" via audience polling to estimate probabilities.
- Direct polling suggested a 37.5% chance of doom, higher than his calculated 12.7%.