I can't find it either. The article in the second link contains a claim that they found the specific defect(s) that caused unintended acceleration, but I can't see anywhere where they actually mention what the defect is. In fact, in the court transcript itself the guy says he wasn't able to reproduce anything:
Q. Now, you have not reproduced in vehicle testing your theory that there's a software bug that opens the throttle and then the task dies, have you?
Q. And you have not reproduced in vehicle testing your theory where there's task death and then the throttle is opened farther by a software bug or corruption, correct?
A. Right. So the second corruption that I talked about yesterday has not been demonstrated in a vehicle. We've not attempted to.
It really sounds to me like this "embedded software expert" came up with a whole bunch of possible things that could've gone wrong, but didn't actually find a bug.