I was thinking about this yesterday. I personally am a in great favor of differerent types of software testers working toghether on a certain goal.
Pictures like these (from previous blogs):
Then what type out of the 9 will the A.I. be?
It can NOT be all 9 at the same time, right?
So how are we going to solve this "problem"?
If we want A.I. to be objective.
Then we should at least build 9 different types? And all those 9 together produce the 1 output? or....
I think there's going to be a split in the type of A.I.
One type that's going to help us code software (and find bugs) on the technical aspects.
the other A.I. type that's going to help us build and test the software so it's going to be intuitive and people like to work with it because it works great.
In any way...it looks like we have an explosive number of combinations: