That is not true. Not even close. This is exactly what everyone said in 2020-2021 and suggest no one will ever catch up to Apple because they are so far ahead. From software industry / programming industry / web framework prominent figures, yes all the software guys all apparently have ZERO understanding about hardware. And then people blindly follows them as source of truth.
If you actually follow the path of ARM Cortex-X design and using GeekBench as performance indicator ( Please dont argue about using Geekbench as reference ) They have seen performance increase with every generation edging towards Apple's Single Thread Performance. You have to also remember the lead time from design to market is roughly 3-4 years. So a lot of these designs started before the insane internet uproars.
What was at one point 2-2.5x the difference in Single Thread Performance, Cortex X-4, on a N4 node, coming in later this year in Snapdragon Gen 3, should be within 8% difference on equal clock speed, compared to Apple's A16 on N5.
I cant remember the ISO die size difference on X4 and A16, but what used to be the case ARM were only using 60% of die size compared to Apple in X2 era. I would imagine the X3 and X4 has gotten larger. But generally speaking ARM's design are much stricter in Die Size simply because that is their potential customer's margin.
Of course, not every SoC vendor will use the maximum X4 design due to all sort of constrain. Especially on Android which tends to prioritise Multi-Thread scenario. That is not a CPU design issue, but a vendor issue.
So unless Apple A17 has another 10-15% performance jump without increasing clock speed. The ARM Cortex X design is surprisingly close in terms of pref per watts and yet no one is talking about it.
And the good thing about Cortex X design is that they are commercially available for license to anyone. Sony, LG, Samsung, Microsoft, Google, Amazon, Facebook, Oracle, or anyone with interest to fab their own SoC could buy the design from ARM. Heck even Intel or AMD.