TeX is highly parallelizable

One thing that I think is missed here is that if you look at what the benefits of Moore's law
have gone to in the past, they've been split between making the *computer* more capable by
exploiting the increased power for more sophisticated tasks (like real-time video decoding)
and making *programming* the computer easier, by permitting programmers to write less
efficient programs in higher level languages.

Now hardware manufacturers are saying "we're going to produce chips where the only way to
exploit the benefits of the increased power is to adopt significantly more complex and
difficult programming models, and we're going to expect programmers to recode the applications
to do that." 

Well, so basically what they're saying is that they're not going to (be able to) continue to
make computers easier to program. That sounds like Moore's Law hitting a wall to me. Sure by
going massively parallel we could perhaps continue to reduce time to compute in those rare
situations where it's worth it to do so, but that's only half of the benefit we've
traditionally gotten from Moore's Law.

Instead of saying "rewrite your programs to take advantage of multiple cores" they could *not*
give us more cores and say "rewrite your programs to be more efficient with the hardware you
have," but that would be blatantly admitting that Moore's Law is dead.