AMD's problem is pricing. And AMD will fix that problem by copying Intel's idea of using two different type of cores, like what Intel is doing. And AMD can do this easily. They can build a CCD with more and smaller, or just older cores and put those as a second chiplet in their CPUs. Mendocino is an example of combining old tech (Zen 2), made at a newer node (6nm) with a newer architecture (RDNA2). Intel and AMD win, we lose. Why? Because until recently 16 cores meant 16 equal P cores. We where paying 16 P cores and we where getting 16 P cores. In the future 16 cores will mean "X number of P cores needed to win most benchmarks, plus 16-X number of slower/older cores to advertise a high core count on the box and the marketing material". The final price of course will be that of 16 P cores.
The 2 P cores example is already happening today. The mobile Intel i7 1260U is a 10 CORE and 12 Threads CPU. It's an i7 with 2 P cores. 10 years ago a mobile i7 with only 2 cores/4 threads was considered a scam. It shouldn't be called an i7. Today it is not a scam, because it also integrates 8 Efficiency cores. So it is OK, because it is somewhat fast at multicore benchmarks, fast in situations where only 2 cores are used and comes with 10 cores. Let's focus on the number of cores. 10 cores, many cores, so let's call it i7.
The thing is that, what we see today in laptops, we might see it tomorrow in desktops. Well, we definitely will see in desktops. And in a way we already see it, but at high end, that's why no one is complaining. The 24 Cores 13700K comes with 16 Efficiency cores, more E cores than P cores. But should we applause because it beats 7950X, or just worry about what will come next? And what comes next is here already. We see Intel cutting 2 P cores in it's next design. So instead of offering all the performance to the consumer with Meteor Lake, it chooses to cut 2 cores so it can make Arrow Lake look more of a valid upgrade when that series becomes available. Meteor Lake will obviously be faster in everything, but still this is the wrong direction in my opinion.