On rare occasions I post about something other than fish or water. I have a MSCS, like to maintain a current understanding of technology, and lately been spending an inordinate amount of time diving as deep as I can into the economic and investing aspects of AI. Recently, Paul Kedrosky has emerged as a compelling advocate of the AI bubble position, at least over the near term. This morning I asked Gemini “what are the errors in paul kedrosky’s analysis of an AI bubble” and was surprised by how quickly a detailed response was provided. LLMs clearly have limitations and uncertainties, especially from an economics and investor perspective, but they have become huge time savers for the sort of research and thinking I do. An LLM is not intelligent and could not have formulated Kedrosky’s argument but its ability to quickly provide an outline of a counterargument based on available sources in an obscure area was impressive. I still don’t understand how an AI company makes money by giving away access to these tools but they sure are useful.