Sounds like software is going to drop drastically in quantity, but jump in quantity. Codegen/AGI are useful but not paradigm shifting. Still some big computer science problems to solve.