actually AMD GPU is the one with best price/performance ratio, but CPU wise it's a tie with Intel. Intel still slightly edges out in lower tier ones.
I went Intel because 7950X and 7950X3D were more expensive than 13900K...and all X670 mobos costed both arms and a left nut

for now. microsoft is currently betting on snapdragon x elite for their Windows on ARM. snapdragon x elite does look promising.

however, x86-64 may not be going anywhere soon despite the rising popularity of RISC-based instruction set architectures (ISAs).

the reason why windows still requiring x86-64 (for now) is because they have a lot of legacy codes that are dependent on x86 ISA. many wizards we use in windows today can be dated back to win95. this is why x86-64 hangs around for so long -- backwards compatibility for the past 20+ years. hell even intel 8086-based software can run on modern x86 CPUs with little tweaking. people may attribute better efficiency to ARM, but it really matter not on the difference in ISAs, but how CPU vendors are pushing at which direction while designing their CPUs. for so long both AMD and Intel been pushing towards high performance (both of them compete in high performance computing while having a lower priority on battery life), while Qualcomm and Apple have been pushing for efficiency (both of them compete in mobile space and therefore having handful of experience on low power operations).

a good read of RISC vs CISC if you're into a bit of technicality: https://chipsandcheese.com/2021/07/13/arm-o...-doesnt-matter/
if u don't already know, x86-64 is CISC-based while ARM and RISC-V are RISC-based

most people in here give intel special treatment only
out of this circle u will see a lot of criticisms on how intel is handling the situation and the criticisms are warranted.

they were the one who let mobo vendors loose in an attempt to gain upper hand on fps regardless of it consuming 300+W...then kpkb about mobo vendors when chips started degrading. a few years ago they said unlocking power limits was in-spec and wasn't counted as overclocking during an interview with anandtech. the unlimited PLs have been here for years now.

don't forget this new "intel baseline profile" has a legitimate performance hit across all workloads.

conversely, if this incident happens on AMD, I predict the hoo haas will still be lesser than intended simply because many people still regard them as underdogs. whataboutism will be brought into the conversation. all in all both sides also gt fanboys. better just buy what suits you best.

and that is what apple is doing extremely right, and at times their low power operation performance is rivaling AMD/Intel high end CPUs which are using 4-5x more power for the same operation. their M4 has just been released recently, with MT performance closing in a 13700K, and ST performance obliterating many, if not all, modern x86 CPUs. a BASE M4 is doing that? at 5x less power? node advantage + SME aside, cannot dismiss what Apple has been doing and they're definitely putting more pressure on AMD/Intel. ESPECIALLY INTEL.

M2/M3 era has already seen the chip performing faster than x86 counterparts in a number of applications such as in Adobe apps, DaVinci Resolve and Handbrake. new M4 era will be another eye opener similar to M2.

with M4 being released this early, Qualcomm is shitting themselves too. I mentioned I had faith previously on X Elite but things do change fast within a month. X Elite is due for >1 year now. after their shoddy X Plus reveal a few weeks ago, rough rumors are saying they're in a very messy situation rn. we'll see how Qualcomm handles this.

hence why x86 is living for backwards compatibility to cater for relic systems. been so long since 8086 era.

however, it really isn't x86 fault for the "slowness" simply because it's just an ISA. a good uarch (microarchitecture, or simply CPU design) will yield great results. Apple has greatly improved their uarch to gain higher frequency, as well as shoving in ARM SME into it, even with small IPC gain (and still yield ~25% improvement over M3).