Much of the literature on AI timelines focuses on the 50% probability threshold. I believe It might be valuable if this literature would In addition also focus on larger thresholds, say a likelihood of eighty% for the event of a certain engineering. LeCun is also the director of AI investigat