All this it's just clickbait, there's no ARM Mac, or at least won't be an transition to ARM exclusive future.
Few things non programmer and non system architect knows about ARM vs AMD64 (aka x86-64):
IPC myth : not easy to compare, as with ARM you need more instructions to do the same as an AMD64 instructions, many more, while current IPC seems close (at half clock speed) to leading AMD64, things are very different when you run code (real world applications) where you realize you need almost twice cycles in arm than AMD64.
Optimized vector processing, while ARM introduced vector specific instructions in ARM, the truth is are far behind to be as productive as AVX, even maybe even slower as some ARM vector processing sometimes is done by microcode instead dedicated ASICS.
AMD64 future, there are nothing new and undeserved about x86 stagnation and some emerging x86 killers, to date x86 survived: itanium (with AMD64), PoweraPC, N-emulator/virtual x86, and won't be different now that fight among Intel and AMD reignited is not the same to try catch someone sleeping than running to stay alive, this is what's enabled ARM to get close x86-64 but doesn't undermine x86 development, as is not an dead end yet.
Intel and AMD foresee deep architecture changes which I feel ARM won't catch, mostly due development atomization (not the same having 2 big r&d team's than a dozen customizer teams or teams just doing close development (as Apple).
The Cisc/WISC/risc debate this is long, on the paper every strategy has its strength and weakens, but I bet on WISC as the future for Von Neuman architecture, while cisc is almost dead (most x86 CPU actually are risc with Cisc decoders), risc is stagnant beyond vectoring/smt/branch prediction, there's nothing More to improve performance, and is where AMD64 and ARM are doomed to become the same thing, while AMD64 decodes Cisc into risc (a instruction decoder it's an asic moreless the same for every CPU family) then intrude it into the execution pipeline where smt strategies are executed (an smt pipeline just put to work the asic on another thread when a thread is sleeping or holding for I/o) smt may rise tdp about 60%-200% while increase throughput by 1.2x to 2x, vector processing it's 99% the same among x86-64 and ARM, but x86-64 use to have a more extensive set of vector instructions (indeed needing more gates), then comes where more gates are needed in either platform: prediction pipelines, most CPU includes a miriad of ASICS to run certain frecuent code and deliver a result in fewer clock cycles, this is what differentiate a high performance CPU (or s faster core) from a high efficiency CPU (or efficiency cores), you can't have both, as for faster CPUs you need a deeper (or broad) prediction pipeline requiring more gates than for efficiency cores where it's used to avoid this and smt execution.
Why I believe in WISC, as you see either x86 or ARM to be faster needs more gates (almost the same, x86 add the instruction decoder), which indeed make the CPU more power hungry, at the end execution pipeline are just an arbitrary implementation of WISC, but a modern well planned WISC CPU may have the best of both worlds: efficiency and speed, but it's deeply corelate to the code compiler, both have to be planned by the same team, instead to compile a high level code as C or Rust into assembly or relatively simple llvm objects,. The compiler translate it into a "miriad" of specific purpose instructions which don't need to decofe neither prediction branch, certainly the CPU will require a lot of gates for that miriad of instructions but at the end less cycles and more efficient smt, and much better I/o synchronization. The best is WISC CPU can emulate arm/x86 for non WISC inherited binaries, what I've read is the approach both intel and AMD are studying for x86-64 successor.
