I'm updating my priors on AI hype: while I still think we're far from true AGI, I'm starting to believe that current models are more than just narrow tricks and may have real-world impacts sooner rather than later - my model suggests we're 2-5 years from serious economic