Re Fstoppers, my reading of that review was that someone buys a device, then next year buys another one, then argues with himself in front of an online audience about whether he's got his moneys worth on either of them in exchange for ad share... He can't lose with circular reasoning regarding "value".
Per year over year benchmark increases and marketing increments like M1, M2, M3 / 12th, 13th, 14th...this is a hype cycle built around a pipeline of progress.
I prefer to keep in mind the long-term trends. To repeat myself from a previous post, a 2023 iPhone can do what 2013 top-of-the-line Mac Pro was designed to do. At it does it at 1/5th the price while fitting in your pocket, running on a battery. And BTW it's got a 4K camera with a range of optics built in, just in case you care. And you can pipe your vid around the world from wherever you are when you get done. Oh and it's a phone.
Correspondingly, if there's an app that doesn't realize obvious generational perf gains but does respond to pure compute scaling, such as what seems to be portrayed for Logic Pro, then questions emerge about the details of that app: the design, the toolchain, and the underlying machine. This makes that app interesting. I mean, isn't it good news that Logic Pro users continue to get leading performance even on older designs? If something is done as well as it can be, is that anything to complain about? Or should the Logic Pro customer feel dismayed that the developers are incapable of realizing performance gains on evolving HW? There's no obvious conclusion about the devices that run the app: it has to be studied.
I read a recent post at Macrumours by a developer who lamented that his mainline work is with Docker containers and everything seems limited to E-cores so maybe he should get a Mac with more E-cores. But his lament included no words for the key trait of running containers, which is to limit scope of allocation of underlying HW to the task. Maybe it's never occurred to him to read the documentation on controls for HW allocations. Idk. But I do know that the whole point of virtualization is to abstract away the underlying details, so why isn't he overjoyed?!
These strange loops of thinking are everywhere. And everywhere appearances are becoming proxies for reality.
There's a recent trend in cinematography to shoot at super high resolution, and relax details of composition on set, then frame later in post. Why such displacement of work valuable is open to consideration, but the natural evolution of this will be to render the entire world inside a machine then let an AI create the movie, maybe dump some actors in there to cross the uncanny valley. Big steps have already been taken in this approach: