The crackling issue for voices and SFX was the main issue, but it had major slowdowns as well. For the Steam Deck requiring a Steam account, it's a temporary solution to combat the parasitic scalpers.

As far as small form-factor PCs go definitely. I just wonder how bad that fan noise will be under load. *cringe*

Like any of these things, it’s not going to bother scalpers, just honest customers. I expect they already have thousands of acccounts to scrape up keys and giveaways and such like.

Possibly, but to get a SteamDeck your account needs to have spent money and be in good standing.

Personally, I'd prefer the device to be fanless rather than powerful.

Good, another "progressive" toy which I will not buy.

You needed to have a Steam account before June with at least 1 purchase. Unless scalpers could have seen this coming a month in advance, I doubt they could have obtained more than 1 unit. Also, considering that most people on Steam who wanted one seem to have been able to pre-order, I guess that was a successful strategy.

GPD Win 2:
Does AMD run cooler (user much less power)?

That is generally the problem I see with x86 (Intel or AMD, doesn't matter) based mobile/portable devices that are supposed to be used primarily or only with battery power. There's a reason Apple M1 laptops, or any Android tablets, have so much longer battery life than even the lowest end x86 based (Windows) laptops...

So I personally would prefer my battery-based mobile devices to have more power-efficient CPUs (like based on ARM), but then of course they aren't hardware compatible with my PC games, like GPD Win 2 and Steam Deck are...

With my (gaming) laptops, I don't care as much about the power usage (other than by the heat production, ie. does the laptop run cool enough) because I mostly use them only when they are connected to power, ESPECIALLY if I intend to play games on them. I use them on battery only in exceptional cases, like I am somewhere outside my home or office and get a call from work that I need to check up on something on some server... then I naturally use battery power, and I don't do power hungry stuff anyway as I need only a remote desktop or ssh connection to somewhere from that laptop, and that's it. As long as the battery lasts one or two hours with such light use, it is fine (for my laptop).

The main thing I want to know with battery-based devices is: can I replace the battery myself, with third-party batteries?

Unfortunately many device manufacturers try to make that harder or even make it practically impossible, both by making opening the device (and hence access the battery) quite hard, and/or adding electronics/firmware that will not accept third-party battery cells changed by the user (e.g. the new battery refuses to recharge due to that). That way they make sure you at least have to send it to their official repair for a battery replacement, or scrap the device and buy a new one.

I recently had to research on that when considering replacing the battery on e.g. my Bose Soundlink Mini 2 bluetooth speaker, or some older Dell laptop.

I like how effortless that is on e.g. the wireless 8bitdo Pro+ gamepads. The battery packs are easily reachable and replaceable by the user, and in fact the user can even use a couple of AA batteries instead of the battery pack (naturally that is not possible on more power-hungry devices like laptops etc., but just goes to show that 8bitdo was not designed to artificially force the user to buy and use only official 8bitdo batteries, and/or send their gamepads to official 8bitdo repair centers to get their battery replaced).

We need more honest and customer-friendly companies like 8bitdo!

So I guess my main question was, for both GPD Win 2/3 and Steam Deck: can the user change the battery themselves, even with non-official (= cheaper) third-party battery cells? You have to remember that the batteries WILL die at some point, and need to be replaced.

You can never know for sure. My wife's Honor 10 (in my eyes an expensive high-end phone) battery just got broken and swollen after a couple of years. It got swollen during recharge so it basically broke the whole phone as the swollen battery "expanded" the whole phone out of its frames. I didn't feel it made sense to try to fix the phone anymore, as the swollen battery probably physically broke something else in the phone, besides the battery itself.

It helps if the battery can be easily replaced by the user, also with third-party battery cells. With some mobile devices it is not an issue, with many it is, making it seem that the manufacturer is trying to prevent the user from doing just that.

Recent Ryzen mobiles, very much so.

My 4700u based laptop (convertible) doesn't start the fan under "office usage". Only if I have it compiling for a while or play somewhat demanding games.
It took a couple of minutes of Shadow Tactics for the fan to start, and then it still stayed at an acceptable noise level.

A fanless Macbook Air M1 is even better in this regard, of course. But some things (e.g. Shadow Tactics) didn't run when I tested.

Have you found actual competent reviews for the M1 CPU, with a good analysis of the battery consumption? I haven't, and I can't seem to find any. I keep running into reviews where they show you that the M1 lasts longer in idle, but who cares about that because you know I actually use my computers. Reviews where the M1 allows for longer video playback, but they never tell you exactly what the intel/amd based laptops they use for their comparison.

Has anybody been able to find a proper review of the M1 chip when compared to intel/AMD, where there is a good analysis of battery life?

I didn't do any measurements, but actually using a Macbook Air M1 for a few months for browsing, C++ development, occasional games (mostly Windows games running in Crossover Wine), the battery life was truly amazing if you stayed with native M1 apps. Less so when emulating Intel arch stuff, but still good enough to blow the pants of x86 CPUs.

My private AMD 4700u based HP Envy x360 can't keep up with the M1's battery life. But instead, it runs everything Wintel without issue. The M1's Intel emulation is incredible, but some games didn't work.

And the battery life of my work laptop with an 11th gen i7 doesn't stand a chance against the 4700u, let alone an M1 Macbook.

And that's what I'm talking about. I have no idea what the specs are on your HP Envy x360. Is it the 15 or 13 inch panel? What resolution is that panel? Does it have a discrete GPU? What capacity battery does it have(Wh)? That's why I'm hoping to come across a review that actually considers these variables and does a proper comparison.

EDIT: I almost forgot, hopefuly a comparison with AMD's 5000 series mobile CPUs...

My big laptop, with the 3500u, also is pretty quiet; if I play a game, I don't hear the fan until after I stop playing, as the in-game music easily drowns out the music. (Can't say that of my desktop, whose fans run all the time when it's awake.)

My small laptop, with an Intel Celeron and integrated graphics, is even quieter, as it has no fan. It gets a bit warm if I'm doing a lot of compiling with the power cord connected, but not to the point of being painful to touch. As an added bonus, it has no moving parts at all. (I wouldn't mind a handheld with specs similar to my small laptop; won't be able to play moderately demanding games (Hollow Knight would be playable, but not at full framerate), but there'd still be a good selection of playable games on it).

Your HP ENVY x360 has 51Wh vs Macbook Air M1 which has 50Wh. So they are extremely close!

I'd love to see somebody do some testing. Basically disable all other tasks, and run the same game benchmark on loop, or the same stress test on both, until the battery dies out on them. That would be a great comparison between the two.