AI winters may come back, I agree, but regardless of how you define 'a few years', it's undeniable that the gaps between breakthroughs have been shrinking. Isn't this what exponential growth means, start slow and get fast very quickly?