From a purely hardware perspective, I think it's always been in that "limbo" for almost every use case anyway.
For a hobbyist even in Ye Olden Days (before there were other single board computers) you were always able to pull up Craiglist or walk to the thrift store and buy a $50 PC that vastly outperformed a Pi anyway. Underclock that PC and it could be a fanless power miser too, albeit not quite down to Pi levels. (Pretty sure I remember people clocking 900mhz P3's down to 500mhz fanless, that kind of thing. Still faster than a O.G. RPi...)
So the Pi's hardware was really only a revelation if you needed that semi-embeddable form factor, or if you wanted to do something like manage a classroom of cheap low-performance PCs without the nightmare of managing a fleet of heterogeneous junk PCs.
It was always all about the objectively awesome ecosystem of support and peripherals, and deciding whether that ecosystem mattered for you and your use case. Sometimes that ecosystem is a massive boon and sometimes it doesn't matter at all.
A raspberry pi 5 8gb CanaKit. So pi, fan, microsd, cords, power. Comes in at $160 on Amazon https://www.amazon.com/CanaKit-Raspberry-Starter-Kit-PRO/dp/...
You can get any number of Intel N150 MiniPC's with RAM, SSD and power cord for $140? https://www.amazon.com/GMKtec-mini-pc-computer-n150/dp/B0CH8... Maybe less on sale. And an Intel N150 is going to score 2-3x on standard benchmarks what the RPi does. You can find older N100 based mini PC's for even less.
The pi still has its place, again, for those ultra low power use cases. But for "I just want a small computer", it's not the option. Mostly because of economies of scale. Everything in the MiniPC's is made at thousands of times more scale than the RPi. Also.... in Chinese factories for significantly less cost to start with.
Sure I could use an arduino in many cases, but the Pi running linux can be very convenient, with bash or regular python or even small unix C utilities (which I can tweak and re-compile right on the Pi!)
The Pi 5 and the RK3588 boards is when I'd get an Intel N100 miniPC instead.
It looks like the RP2040 on the X4 is connected via USB. The RP2040's bootloader mode that enables flashing emulates a mass storage device over USB to transfer firmware, so it would be confusing to not connect it via USB.
This means that your x86 program doesn’t directly control the GPIOs, it must communicate with the RP2040 over USB serial. The RP2040 must be flashed with a firmware that knows how to handle these serial commands.
It works, but it’s an extra step and timing may be an issue.
On the other hand, a Raspberry Pi Linux program can toggle GPIO pins natively simply by using the official libraries.
I mentioned the RP1 in another comment.
The only question, for me anyway, is how much value I put on that experience and whether the increased cost is worth it.
I would reconsider the choice when I have a commercial product that I have some scale of production, where a few bucks off the BOM is important.