Good points about how in practice it won't work as smoothly as in theory.
As for justifying the idea itself, it comes down to who pays for development of features. By segmenting, they can have the people that need and can justify the cost of those additional features, pay for them. If they enabled them in every chip, people who don't need them would pay more.
In the abstract, isn't software itself a version of this? "You paid for all that nice hardware that can hold any program in memory, but now you've got to pay us for that program to make use of your hardware."
They are two different things. One is useless without the other. And you don't necessarily have to buy software when a free alternative like Linux exists. I understand that hardware companies need to make money, but what basically amounts to a subscription service to use it is not the way forward.
(post is archived)