As to the where in the product stack.. both comparisons make sense, as long as people are applying whatever their test is to both "sides" so to speak.
Yes, it's valid to say "this is how much the product in the same position in the stack has improved, but for this much more money" (ie RX 6600 vs RX 7600). A "how much has it progressed, and is it worth it?" scenario.
But, it's also quite fair to compare prices for cards that are roughly equal in absolute performance, or compare performance of cards of roughly equal pricing, even if they're in different positions in the stack of their generations. The "Yeah, I can get that, but I can get the same performance for less with the older card / I can get better performance with the same money for an older card."
In the former, you'd compare the 6600 to the 7600, and you'd then also compare the 3060Ti to the 4060Ti. In the latter, you'd compare the 6650XT or 6700 to the 7600, and you'd also compare the 3070 to the 4060Ti.
They both make sense, but for differing reasons. The only real problem arises if you try to use one of those comparisons for one brand, and the other comparison for the other brand. At that point, it becomes picking-and-choosing in the pursuit of fanboyism.