The official explanation seems to be that the key management defect which lead to thermal management problems did not occur in the development machines used during the engineering process. This seems possible as presumably every machine used internally goes through a different provisioning process than those released into customer hands ... Thus the claimed cause of the problem
seems to indicate that the flaw did not affect the internal benchmarks they ran. It would be _very_ concerning if the benchmarks they ran for engineering/marketing production were insufficient to catch this issue.
I think the official explanation seems pretty likely. If the flaw had been present in the machines used for benchmarking than either their benchmarks didn’t catch the issue (unlikely) or the benchmarking numbers were fabricated which also seems unlikely as it would invite an infeasible amount of legal risk for an organization of apples size and mindshare ...
Engineering, especially at this level, is not supposed to work like that. You certainly not do dev proto, then switch to prod both in one step and without even rechecking the tests you did on the proto. So their official explanation might be kind of right, but then it shows major deficiencies in their quality process.
Which given the track record of quality issues they have had, is possible.
> You certainly not do dev proto, then switch to prod both in one step and without even rechecking the tests you did on the proto.
Exactly. I work in software engineering for cars and even for things that are not mission critical or life-threatening, the engineering and quality process is insane because so much money is on the line. I wouldn't be surprised if macbook pros are worth more in profit, or at least revenue, than a car line and should therefore see an equal engineering process.
I can't understand how the firmware was released without the proper key, unless there was a bad merge or they dropped a pilot firmware that was testing without thermal management... maybe a clue there is a desktop in the works and the desktop firmware made it onto the macbook pros? Either way, like you said, their quality process has a major deficiency.
> It would be _very_ concerning if the benchmarks they ran for engineering/marketing production were insufficient to catch this issue.
Clarification -- I meant to say -- 'It would be _very_ concerning if the benchmarks they ran for engineering/marketing production were insufficient to catch this issue on a system configuration in which the issue manifests.'
I think the official explanation seems pretty likely. If the flaw had been present in the machines used for benchmarking than either their benchmarks didn’t catch the issue (unlikely) or the benchmarking numbers were fabricated which also seems unlikely as it would invite an infeasible amount of legal risk for an organization of apples size and mindshare ...