This has no relevance to the real world and even describing it as "already faster than the fastest CPU...Intel provides" is highly misleading. In fact no ARM or Apple CPU is remotely faster than Intel's latest desktop CPU. They are not faster. They may have equaled Intel's IPC but that has nothing whatsoever to do with which one is faster from a real-world performance standpoint.
Stating performance at different clock rates than the CPUs were designed to run is no different than taking a nitrogen-cooled i7-6700K at 7Ghz and benchmarking it against an ARM/Apple CPU at that same speed. The ARM/Apple CPU won't run at that speed, even cryogenically cooled? Does that mean the Intel CPU is *infinitely* faster? No. It's not a valid test.
There has long been a CPU architectural debate about "Brainiac" vs "Speed Demon" designs. In general Brainiac designs have good IPC but can't be clocked as fast. Speed Demon designs have less IPC but can be clocked faster. Just because a given design achieves a certain IPC doesn't mean it can be up-clocked to match a competing design. You can read about the history here: http://www.lighterra.com/papers/modernmicroprocessors/
Intel spends *double* the R&D of Apple -- every year. After spending a quarter of a trillion dollars on R&D the past 20 years, Intel has figured out how to make a high IPC CPU that can be clocked at over 4Ghz in stock form. Just because ARM or Apple has finally achieved a good IPC metric does not mean they will be able to run those at competitive clock rates. A test running CPUs at half the normal clock rate is not predictive and it's misleading to imply otherwise.
