Since here we're in the business of getting to the bottom of things rather than basking in the haze of the reality distortion field surrounding Cupertino, let's be sure we understand why Mac developers are now "eating crow" and fussing over endian issues on Intel chips.
The first reason is that Steve Jobs is going to do everything he did at NeXT over again, at Apple. It's like Bill Murray in Groundhog Day. Do it over and over until you get it right. We had FAT binaries in OpenStep (040,PARISC,Sparc,x86), we have FAT binaries (ppc,ppc64,x86,x86-64) in OSX. We even had FAT binaries in Windows NT (Alpha,x86,MIPS,PowerPC).
The second is that Motorola and IBM turned out to be lousy partners in the race for high performance. And once Intel realized that AMD was eating it's lunch using chips that used about half the power, they essentially kicked Motorola in the nuts. Inadvertently crushing the G4 in the process. Intel didn't mean to crush Motorola, they were just gunning for AMD.
IBM, on the other hand, (it is obvious now) was looking for POWER partners who would ship large volumes of chips, but wait just a minute - as long as those chips were not going to endanger IBM's high-end workstation and mainframe business.
Virginia Tech's G5 supercomputer. Great for Apple. Great for POWER. Really, really bad for IBM's mainframe and supercomputer business. Think about it. Instead, IBM went out and courted Microsoft's XBox business while sitting on their hands in the G5 arena. This is a sound business strategy for IBM and I wish them success with it.
Apple, on the other hand, can now plunder and endanger IBM's supercomputer and workstation business at will, using chips from AMD, Intel, Motorola and IBM. They are sitting pretty.
I don't believe the hype surrounding CoreDuo etc. These are just chips with big on-chip cache. A big on-chip cache makes everything seem to run faster. But the CoreDuo is a lousy multi-tasker. Get too many tasks running at once and you can't even play a song in iTunes without big gaps. Not so on G5 or G4 - both these chips are svelte multi-taskers. I can load up my G4 laptop with tasks and iTunes never, ever drops a frame of audio. It works the way it is supposed to. On MacIntel, just starting an XCode compile is enough to choke the machine to a stuttering
K-K-K-Ken.
I also suspect that Apple has tweaked the tasking/thread priorities for MacIntel to make it seem like the boxes are faster than they are. Later, after everybody has calmed down, Tiger on x86 can be retuned to make it run smoothly, if not more slowly.
Sooner or later, everyone will realize, RISC and unix are the future of everything. Naturally, since you're a Machead and owned a G5 for a couple of years, you think that you understand IBM. Listen, if you haven't read
Fortress Rochester, you have no idea what IBM is up to, what their strengths are, and why they are anything but a dinosaur.
IBM is the reason we have at least one 64-bit ABI on OSX, not Apple. I mean, the MacProXeon boxes are here but there is no x86-64 ABI. I guess the scientific community doesn't get the "just a recompile" treatment.