Grue_Slinky comments on What are concrete examples of potential “lock-in” in AI research?