[jongjong]

My perspective is that AI is a multiplier if existing skill (both positive and negative aspects). A bad developer can now produce bad code at 10x the rate and a good developer can produce good code at 10x the rate.

There is a net positive gain on the automated testing side of things but I think a bad developer, even with AI will not be able to out-compete a good 10x developer without AI. The costs of incorrect abstractions is just too high and LLMs don't really help avoid those.

You have to ask the right questions. It's a bit like in Hitchhiker's Guide to the Galaxy... Where a super-intelligent computer took millions of years to calculate the meaning of life as the number 42. The wrong question will always waste computation.