1,655 comments

[ 5.1 ms ] story [ 394 ms ] thread
Very cool. Seems like a steam-powered Switch. Wonder how the other console manufacturers will react.
I wonder what this will mean for Linux support on the Steam store in general? Since this is SteamOS[0], then manufacturers will have to support Linux / Proton in order to have their games on this machine.

[0] https://www.steamdeck.com/en/software

It is! A new version, in fact. It includes Proton.

(I'm not sure if SteamOS v2 and below were already Arch-based, but it's interesting to know that v3 is, given that the Steam Linux Runtime is mostly Ubuntu-tested)

It is indeed SteamOS, though they're claiming that Proton will be improved for this:

> For Deck, we're vastly improving Proton's game compatibility and support for anti-cheat solutions by working directly with the vendors.

I noticed one of the screen shots was using KDE on the connection to another screen. Honestly that got me to realize it's another PC and made me consider getting one. Worst case scenario, I just have another PC that's marginally better than those small box office PCs.

Edit: I read further, it is a fully fledged PC. You can install or do whatever you want to it.

Glad to see two sticks in addition to the trackpads.

The Steam Controller with its one pad/one stick approach was...interesting but I just couldn't adapt to it for most games.

Storage seems like it'll be an issue unless you only want to play indie games. It says it's got the hardware to play AAA games, but even last gen AAA games like Doom Eternal won't fit on the base model, and the next step up with more sufficient storage is more expensive than the next gen consoles.
microSD cards are pretty cheap though.
Does anyone play AAA titles off of SD cards? Wouldn't it take forever to load stuff?
On switch it isn’t an issue for 12gb games. But I couldn’t imagine trying to load an 80gb game. Especially if you needed to load stuff while playing.
Yeah - I imagine that the games are designed thinking most players probably have an SSD or - at worst - an HDD. An SD Card - especially a not-good one - can be 100x+ slower than an HDD for small reads. Any games that depend on this would effectively not be playable.
Micro SD cards are up to about ~290 MB/s for read times these days, IIRC. That's the same as low-end SSDs.
That's throughput.

IOPS is important.

For random 4kb blocks / second

SD Card => 2.1 [1]

SSD => ~200 [2]

[1] https://www.cameramemoryspeed.com/reviews/micro-sd-cards/

[2] https://en.wikipedia.org/wiki/IOPS

And that's comparing good SD cards to good SSDs.

As I was saying, bad SD cards can be REALLY bad. The difference can be over 1000x. Anything that depends on being able to read ANY small piece of information regularly in a reasonable amount of time from disk would simply not be playable.

That's assuming the games are writing to the SD cards and not the SSD though, right?

If this is Debian/Arch OS, I'd assume they are writing to a system folder.

No - those are random READS - not writes.

If the game is on the SD Card - it will be read from the SD card.

Sure, if the game needs to randomly write data to disk (I assume this is much rarer) - then it would write to the system SSD as usual.

Got it. So SD cards are optimized for large, continuous file reads, not varied file reads.
I'm not sure that's what they're optimized for. I think it's more of just that's how the technology works.

It's the same way I wouldn't say that film is optimized for storing lots of data with horrible access cheaply. It just happens to be what it is.

Used to play them fine on hard drives. Micro SD cards became as fast as them quite a while ago.
I think it's about as fast as a regular hard drive. So I guess use that as a point of reference.
All the next gen consoles have SSDs though, so games are going to be designed around that. Current AAA games might be playable on a microSD, but I doubt any of the AAA games currently in development will.
You're saying 100+ MB/s won't be enough anymore? Why would that be the case?
Because the PS5 SSD does 5GB/s, and that's the kind of speeds new AAA games are going to be built around.
Well, those are PC games, optimized for all sorts of hardware, like spinning rust, where other stuff is happening in the background as well.
A lot of AAA games are built for PC and consoles, that's why you don't normally see much progress in graphics until a new console generation comes out. Current AAA games are also already recommending SSDs, so if people think they're going to have a good experience playing AAA games on an SD card, they're going to be dissapointed.
Oh right, NVMe. I have yet to experience it. Yeah, I suppose if the software is optimized for it, it is a much better experience.

SD cards also have the drawback of having pretty limited write cycles compared to any SSD.

NVMe does next to nothing to PC games (compared to an SSD). Direct Storage (Windows 11) requires a minimum of 1TB.
We still have yet to see if that will actually become a standard going forward. Sure there's probably gonna be some fancy games without loading screens doing crazy content-streaming tricks to show of the new console hardware, but in the long-term, studios may decide it's not worth the time and effort compared to just having a loading screen that loads the whole level in a big chunk.

Plus you're gonna be targeting a much wider userbase for the time being by not targeting SSDs.

I think there's the argument that most studios will do whatever's easiest. Only a few will really try to be fancy and use the new hardware to the utmost, right?

So then yeah, sticking with a load screen _might_ be easiest. But if the tooling supports it, it might be even easier to just not worry about loading and having to make a loading screen, and let the engine handle that stuff for you. UE5 at least seems to be going in that direction.

Being not NVMe is enough to call it slow
Aren't read speeds a little too slow for heavy game loading?
Well for heavy gaming I doubt you'll be able to put any more power in such a small package. I don't think the target audience is heavy gamers, but mid-tier games.
You'll probably be able to hack it and add your own NVME drive. Everything is standard and won't be locked down like Apple's hardware.
They explicitly say the storage is not upgradable, and it's absolutely not surprising given the form factor.
Valve doesn't try to prevent it. But if it's soldered, it won't be easy.
> Deck

What an uninspiring name! (with unfortunate mispronunciations). Looks very interesting though, price point (starting at £349) seems very competitive.

The closest competitor to this is called the switch. Other competing products are called a station and a box.
Steam Deck just doesn't seem to roll off the tongue, maybe it will grow on me.
Sticking with the ship metaphors, it was better than Poop Deck.
I'd like to believe they named it that simply so kids could pull a "My Deck is bigger than yours!".
> "My Deck is bigger than yours!"

Ah, the rallying cry of EDH/Commander players everywhere.

Not a William Gibson fan I take it - I heartily recommend you read Neuromancer if you enjoyed the Matrix / any cyberpunk at all.
I really hope that this market picks up steam (heh) and we get to see even more bigger names coming up with devices. The Aya Neo [1], OneXPlayer [2], and GPD Win3 [3] have all come out relatively close to each other with different ideas of what people want in a handheld gaming PC.

Valve seems to want to get in on a price that is more competitive with the Nintendo Switch, so its hardware specs are a bit worse it seems than the others in the market. The plus side to this is that the base model comes in at just $399, though that is with eMMC storage, while the next bump up uses NVME. The other specs seem to be identical across the 3 SKUs, though.

The trackpads a la the steam controller seem appropriate for the types of games that require a mouse on PC, which is unique in this product market. Gyroscopes are also built in, which I'd presume would work similar to the Wii U and Nintendo Switch for aiming. I don't think it'll be as good as a mouse, but I've found that aiming with gyroscope on the Wii U was much easier than using just the analog stick, so hopefully it makes FPS games with a controller more pleasant for me.

I also think it's interesting that it is using SteamOS, which had been kind of abandoned by Valve for quite some time. This also means that it is depending on Proton for game compatibility, which in itself is a huge statement on their confidence in the maturity of it. Without it, this would be a complete failure like "Steam Machines" were. If it works out well, maybe we'll finally see a real product instead of a tech demo from Dell and other PC manufacturers. Exciting.

[1] https://www.ayaneo.com/aya-neo

[2] https://www.indiegogo.com/projects/onexplayer-best-performin...

[3] https://www.indiegogo.com/projects/gpd-win3-the-world-s-1st-...

I'm surprised they went with a custom AMD part. Seems expensive for what is likely to be a low volume product. Why not just do a laptop in a handheld form factor.
I don't think you come at the Nintendo Switch expecting low volume. Whether another company can be successful at it is a good question, but if you build expecting low volume, that's what you're going to get.
Nintendo has some rabid fans as well as selling mostly exclusives. Valve's market mostly already own gaming PC's and won't buy a handheld. It's just too expensive for the few time's you want to play games and aren't at home. For most people.
(comment deleted)
Well I guess their time developing the Steam controller isn't all for naught. Looks similar in concept, except with dual joysticks and thumbpads (instead of just 1 each) and extreme deemphasis the traditional X/Y/A/B buttons, and 4 (instead of 2) back paddle triggers. Which is great, I loved the paddles on my Steam controller.
I wonder if the triggers are also two-stage, like the Steam controller's. I loved that feature, found some niche uses in some games.
The IGN video seemed to suggest that.
Finally a linux gaming console

Now valve, for the love of god, PUSH NATIVE LINUX GAMING, proton only as a way to supplement the library..

Reduce your steam tax for the companies that provide a native linux build

I will not buy this if your main selling point is "proton"

As for the price, 419€.. very bad marketing, 399€ would have been perfect... you got greedy for 20€, that'll hurt sales

It looks like it is $399 USD, so the price point may mostly be a conversion thing.

Edit: Just looked up the conversion rate, right now 399 USD is 288 GBP? Weird. The mid tier is 529 USD, so something like 382 GBP. Upper tier is 649 USD, or 468 GBP.

1€ ~= 1.18$ US So, probably not just a conversion thing
$399US is ~338€ though. Even adding in 19% VAT we'd come out at a nice round 400€, so I don't really understand the 419€, which is just a weird number from a marketing perspective as well.

That we're getting overcharged on this side of the Atlantic isn't new though, but happens with basically everything.

And if you add 5% for import duty, you get a number very close to 419 EUR.

I don't however know how much the import duty is, it seems to be a number that is very difficult to find :).

With 24-27% VAT in use in some countries it comes to exactly 419€, which might explain the price. As they really don't charge different price per country in Europe.
Isn't that a tax thing? Prices in Europe are displayed VAT-included, while prices in the US are VAT-free, are they not?
> Reduce your steam tax for the companies that provide a native linux build

The idea that this would move the needle seems like wishcasting. Ten percent less of 0.89% of the Steam hardware survey is a couch-cushions rounding error. Meanwhile, Proton is a really good way to get hold of that rounding error as-is and works with surprisingly few problems across most games I've tried; For the amount of noise that 0.89% of the audience makes, Valve's spent quite a lot of time-and-effort to come up with something that works quite well.

Where have you been? Valve has been pushing native Linux games for about 10 years.

https://www.phoronix.com/scan.php?page=news_item&px=MTcxMTY

yeah i can't play MMOs, yay

i can't play most multiplayer games, yay

i can't play most AAA titles because of DRMs, yay

proton is a waste of time

most of the time you have to fiddle with weird settings, portable console appeal more to casual people, majority of them are tech illiterates, if something bug or doesn't work properly and requires tweaks, it'll backfire at both the devs (they bought the game after all), and valve (they bought the console)

if they don't push native builds, it'll flop due to lack of proper NATIVE ecosystem

"daddy, why this game doesn't work, this console sucks"

this only asks for a proper windows based machine, if XBOX releases their XBOX portable, it's game over for valve

> yeah i can't play MMOs, yay

> i can't play most multiplayer games, yay

> i can't play most AAA titles because of DRMs, yay

That's on the developers and publishers.

> proton is a waste of time

It works well enough for a lot of people.

> most of the time you have to fiddle with weird settings, portable console appeal more to casual people, majority of them are tech illiterates, if something bug or doesn't work properly and requires tweaks, it'll backfire at both the devs (they bought the game after all), and valve (they bought the console)

That sounds like PC gaming in general. If you're talking about Proton messing up, there's lots of warnings that things might not work right.

> if they don't push native builds, it'll flop due to lack of proper NATIVE ecosystem

Way to miss my point. There's lots of games with native Linux versions: https://store.steampowered.com/search/?&os=linux <-- AFAIK, Steam's "os=linux" means without Proton.

but not the best sellers, because devs have no incentive to release native linux builds

with that new console however it can change

valve has to make it compelling for them, by lowering their tax for example

(comment deleted)
Why on earth would you think trying to incentivise developers to release native Linux builds would be a more worthwhile investment than simply improving Proton? Linux desktop probably occupies about 2% of the steam market and that's being generous and assuming that their actual steam hardware surveys are underreporting adoption.

Even assuming wild success on this products part, that would only drive that number up a few percent since I don't see production for these running particularly high.

It might make more sense to encourage developers to provide proton support for anti cheat (like what Valve are actually doing leading up to this device's launch) rather than asking AAA developers to invest additional time on support that is arguably a waste of resources and certainly something of low priority.

Aside from that, my own experiences with Proton are far more positive than yours. I've hardly ever had to tweak things for it to work. Occasionally, I need to use an older proton version for a game but that's hardly a chore.

because that's the difference between:

- supporting a platform

and

- hopping everything is compatible

will they hook and reimplement EVERY WINDOWS APIs?, what about Windows 11? what about DX 12 and DX 12.1?

Oh, now imagine that windows 11 now requires X security features that are impossible to emulate with proton, how will that work?

native vs non-native

that's the difference

i'm not against Proton, i think it is a very nice idea to SUPPLEMENT my library

nothing can beat native games, built for the device and its constrains

and and an other WORSE thing

CONTROLLER SUPPORT in steam

valve never encouraged people to support controllers, WORSE, 99% of games expect you to use XBOX controller, as a result XBOX controller are displayed even if you use a DUALSHOCK one

valve could have provided APIs so they can feed games with proper icons etc etc, but nothing

so many wrong things, including the focus on proton as main selling point

that's part of the reason why you'll see majority of people install Windows instead, and "the year of linux on the desktop" will yet again be killed due to poor decisions

i'm kind of tired of trying to provide arguments, i reply to that kind of comment on a daily basis, it's painful to see linux people stuck with their WINE, it's WRONG, it's BAD, it's COUNTER PRODUCTIVE

microsoft helped a lot to drive linux adoption up with their crappy Vista/Win8/10 releases, but yeah, let's focus on emulating a bloated OS instead!

you don't come up with such a device over night, they had years to think about that, they focused on emulation

we will see who was right in the end, i think i am

"CONTROLLER SUPPORT in steam

valve never encouraged people to support controllers, WORSE, 99% of games expect you to use XBOX controller, as a result XBOX controller are displayed even if you use a DUALSHOCK one

valve could have provided APIs so they can feed games with proper icons etc etc, but nothin"

Out of all of your rather dubious points against Valve's plans, this is by far the most absurd. Have you heard of a little thing called Steam Input? Valve in fact have provided an API that does everything you've listed including feeding the "proper icons". It's up to the developers to use it. That would not change even if the game was 100% native because xinput is still the industry standard. Controller support on Steam is second to none.

"you don't come up with such a device over night, they had years to think about that, they focused on emulation

we will see who was right in the end, i think i am"

No, I think it's safe to conclude that you are wrong even from this date. You're asking Valve to compete with a platform that has all the games available. You're also assuming that native is always better than non-native but as I've already seen, a crappy native port is often beaten in both performance and feature set by Proton. Instead of asking developers to gamble away valuable time on the OS of a device that could be a flop, it makes a lot more sense to release the device with the games currently available in the market in mind and then give developers an incentive to add special support for a device that is actually out in the wild in the hands of paying customers.

You also really do not need to implement every Windows API, you only need to translate draw calls to Vulkan and then a small subset of API's that actually relate to games such as sound. Your example doesn't make a lot of sense, the only way a security API would matter for Proton is if the game is from the Windows store but most people do not in fact use the Windows Store even on Windows so games tend to get released on Steam as well - including even many Microsoft titles.

At the end of the day though, Valve's approach has made it so that I already do not miss Windows for gaming even with the current non-ideal situation with anti-cheat. If they adopted your approach, we would have likely had a small increase in the number of native games but most of the industry would have probably ignored such a call. It would also be very unlikely to magically make desktop Linux a real competitor. I think we need to stop chasing that dream because frankly, it just doesn't make a lot of sense to me. The only way you would be able to compete with Windows and Mac is by having a similar marketing and UX budget and by removing or downplaying many of the attributes of desktop Linux that likely drew us towards it in the first place - package management, hyper-extensibility, the terminal, GPL licensing, etc.

I do not think it's worth it frankly. There is nothing wrong with being "niche".

The pricing even doesn't seem unreasonable. If the device is of a good quality, this could sell pretty well.
I don't disagree, but I think the $399 version might be their undoing. 64GB is about 1 AAA game, and 0 in some cases. So people are going to get this one for their kids for Christmas, the kids are going to say it's shit, and it's going to get a bad reputation, even if the high end models are good.

Maybe they have solutions lined up for the problems that will come with low/slow storage, but we'll see!

The Switch has had the same problem, which causes me to basically never keep AAA games installed for long, but they are still quite successful.

The Switch also tops out at 64GB with the new OLED model, whereas that's the starting point for the Steam Deck. There is also a history of consoles being offered with a wide variety of different storage sizes (Xbox 360), and I don't remember that causing any serious problems for them (if you ignore the humongous size of the Xbox dashboard at the end of its lifetime, which is an avoidable problem).

One caveat, at least for me, is that Switch games tend to be designed around cartridge limits, and the transfer speed there isn't too much faster than the Switch's UHS-I SD card reader (although I can't find any tests or specifications on the cartridge port). PC games nowadays tend to be designed around either a SATA-III SSD or a PCI-E SSD (over NVMe), which will be much, much faster than a UHS-I port, so PC games over an SD card won't be anywhere near as performant as a Switch game.
>64GB is about 1 AAA game, and 0 in some cases.

They will just use an SD card. Yeah it's not as fast as SSD but it's the low end device.

It will be easy enough to hack it to add your own storage. Probably just a screwdriver is needed and you insert your own NVME drive. Everything is standard hw.
I'm guessing the NVME is soldered into place. But with a full usb-c port you should be able to add fast enough storage.
Valve, you created a lot of products that were great but then you dropped them for some or another reason.

I would like for once for you to keep working on a singular product, release multiple iterations, make it better and make it the market leader.

In many ways this is clearly their second-take on Steam Machines. They learned the first time around that licensing the name to third party hardware vendors didn't work (a lesson they could have learned from the 3DO), and that people didn't really like streaming games from PC to TV.
It was pretty clear why they dropped them, which is most people hated it (while a few very much loved them). I loved my Steam Controller but right now I'm using mostly my Xbox controller for PC.

The Wii U was a failure, but I feel (and I guess the sales can confirm that) that Nintendo learned from their mistakes.

Steam Link and Steam Controller were dropped because they weren't selling very well, but the steam link application continues to be updated, and the Steam Controller continued to get firmware updates until 2018.

If this Steam Deck thing sells well, they'll probably make another one. If it sells poorly they'll probably keep the software updated and use put discounts on it to clear out stock.

To me this seems like a decent combination of all of their past hardware experiments. It’s a steam machine, but in a portable form factor. It has the streaming capabilities of the link, and the touch controls of the steam controller. If this takes off, I’d expect to see them continue to release accessories and iterations of it. Maybe we’ll eventually new steam controllers for use in the docked mode.
I'm looking forward to the 2nd generation of Steam Deck. There won't be a 3rd as we all know.
Valve has a pretty bad track record with hardware, do they not? I'm not very aware of this industry, but the vague impress I have is that they've tried this a couple of different ways and it hasn't gone well for them. I'm curious why they would make this play again.
The people who have used their VR hardware say it's excellent.
I have a Valve Index VR headset. It was clearly the best on the market when I got it last year. The controllers and headset tracking are great and the video quality was a notch above the competition (though recent newer headsets have closed the gap). However it did have some hardware issues such as crackling audio in one of the speakers, so I ended up removing the headset speakers and wearing my own headphones. Anyway, overall I'm satisfied.
They have a tendency to drop their hardware quite early - I had a Steam Link so I could play computer games on my television, it works really great.
Steam link and the controllers were great. Too bad they dropped support for them so quickly. Went looking to replace one that finally broke down and found you couldn't buy them anymore.
Do they? My Steam link still works, and its great. As is the controller.
I think the issue is that both of those are dead products now. Dropping support/sales for hardware so quickly is a bit of a turnoff.
The Valve Index is one of, if not the best, VR headsets on the market. Quest 2 is a very close competitor and arguably better when you consider price + wireless ability. But the Index is no slouch.
The Steam Link was awesome. Their VR headsets are the industry standard. The controller was niche but I heard good things. Steamboxes weren't really a hit, but I don't consider them a failure either (users still got fully functional gaming rigs).

So no, I would disagree with that assessment.

They've got a good record with making hardware, but not making it mainstream.
I think their hardware has been great. The Index VR headset is high quality, and so was their controller although it didn't seem to find market fit.
I'd argue the opposite.

The Steam Controller was a work of technical art (albeit without the centuries of experience in ergonomics its competitors had).

The Steam Link was a few years ahead of its competitors.

And the Steam Vive's Lighthouses were a literal light-speed ahead of the Oculus for a while.

Fully agree however that you're right that they're too small to support a hardware long past its release when the market share is too small.

Happy index/steam controller customer here, had to RMA one of the index controllers for a weird resetting issue but they handled that really well so I feel I’ve got my money’s worth. I may be biased though because valve is actually the company I have been a customer of the longest in my life
Valve's arguably biggest most obvious failure are the Steam Machines [0], but that doesn't feel like a track record failure in the way that, say, a PS Vita was. Weren't Steam Machines basically just branded (and reasonably priced) PCs pre-loaded with SteamOS/Linux? Not even the Steam Controller feels proprietary. I mean, in the sense that I can still use it handily on macOS despite its ending production 2 years ago.

[0] https://arstechnica.com/gaming/2016/06/its-time-to-declare-v...

They have a great track record with hardware. They have a poor track record of hardware adoption.

I have the Steam Controller: it's fantastic, so versatile, and everything it claimed it would be. I have the Steam Link: works great, brings my desktop to any TV in the house. I haven't used their VR equipment but from those that have, seems like basically the reason get Oculus instead of Valve is because of Oculus exclusives.

I've been dying for this form-factor for years, since long before the Switch came out. Valve has been losing tons of money from me in the form of purchases, because games in their library haven't been in the right form factor, so I've bought them for mobile and for Switch, when really I'd rather have bought them in Steam with a mobile-play option.

I've explored the mobile-phone-with-controllers-streaming-from-a-desktop option, and it's just not what I'm looking for. I've been looking for a feasible option in this space since the Razer Edge came out in 2013, and there just hasn't been a good one...not with phones-with-controllers-attached (have to stream), not with mini-laptops (bad ergonomics and performance), etc. I've looked into a number of homebrew efforts to try and get Steam games onto my Switch (streamed or otherwise) but it's just not worth the commitment of time and effort when I could spend the same $10 for that indie game I want on Switch instead of on Steam.

I reluctantly got a Switch two years ago. I was reluctant because although I've been wanting this form factor for years, I also wanted a device that was compatible with my Steam library. Since buying the Switch, I've only bought Switch games (and not a single Steam game) mostly because being able to play on that form-factor is worth the price-premium to me.

It's Valve's excellent record of delivery on hardware that makes this a pre-order for me. Super excited about this.

Typo in the title. It’s “Steam Deck”, not “Steam Desk”.

Interesting that Valve is diving deeper into hardware. Remember when they still made video games? Remember when there was still a hope they would some day release Half-Life 3?

My understanding is that their Index VR hardware is quite good too.
I can confirm this. The index is a marvelous piece of tech in so many ways. It's a premium bit of kit (with the associated premium price) and I have zero complaints about it.
Does the Index VR work with the Steam Deck?
Yes. It’s a PC running Linux and can do anything a regular PC can. It’s probably a little lacking performance wise for VR though. It’s optimised for the native 1280x800 resolution.
Don't forget Alyx exists.
Alyx is a VR game, not a video game in my book.
(comment deleted)
"Half-Life is a PC game, not a video game in my book."

Seriously, Alyx is pretty obviously the future of video games. Have you tried it? It's incredible.

I doubt it's the future. There's a mini documentary called "Half-Life Alyx: Final Hours" and the majority of the developers at Valve said they don't want the next game they work on to be VR.
“Video” is a thing you can project on a flat screen. VR games is a separate category, imo.

And no, I don’t own a VR setup.

You can project VR content on to a flat screen. You don't want to look at it that way, but you can do it. The part that makes it 3D is entirely down to the lenses and optics used. The display inside is just like any other.

Because it is video.

That’s an extremely pedantic interpretation of the word video. The game isn’t designed to be playable on a screen with regular control input. It’s a video game the same way a full flight simulator is. Technically true, but it requires specialised and expensive hardware to “play”.
Hardly specialized and expensive, it works fine with a $300 headset aimed largely at serving as Facebook's next social platform hooked up to your PC. It's like saying a Wii game isn't a video game, or that games you can't play without stereo headphones aren't video games.
VR headsets are just a flat screen with a pair of lenses on them.
Cmon, don't gatekeep games. There are inputs, outputs and challenge. Fits the classical definition.
Not saying it’s not a game. It’s just not a video game. “Video” is a thing you can project on a flat screen. VR games is a separate category, imo.
Honestly I'd much rather have Hardware Innovation and Linux Gaming Enthusiast Valve than Made Yet Another AAA Shooter Valve.
Yet another successful game vs Yet another failing hardware device.
The index is fantastic, minus the c-stick of all things to get wrong. Still play it often, lobbies for Pavlov fill in seconds.
The Valve Index is working just fine. If you follow VR devices, it's actually now the only one with good support and no huge privacy issues.
I wonder about the privacy issues here to be honest. The "FAQ" linked to by the site says "we can go to ign.com on it" but what happens if we do? Does Steam see that we went there? Do they log that? If we log into our bank accounts on the device does Steam see that too? What telemetry are they collecting? Does changing the OS prevent all of it? Will there be ads?

It's cool tech, but I have to ask how will it be used against me? Will it only be pushing me to spend more of my money on Steam games or is there something more?

They've basically configured an out-of-the-box Arch system to boot into Steam Big-Screen. That's as privacy-respectful as it gets.
I thought Half-Life: Alyx was basically our HL3? It certainly got critical reviews/scores suggesting it deserves its moniker of Half-Life.
Finally, the year of Linux on the Desk.
https://www.steamdeck.com/en/tech

> Operating System: SteamOS 3.0 (Arch-based)

> Desktop: KDE Plasma

This is sure to excite the Linux community.

And they switched from Debian to Arch. Can't say I don't understand them, but I would have thought there would have been better bases than a rolling distribution (Gentoo, NixOS, Fedora).
They'll surely host their own repository rather than use Arch's directly. There's no reason to question stability when they're in control of both ends, and Arch is light with excellent tooling.

It's a good fit.

I find that odd too, but in my recent experience, Manjaro has been the best Linux experience I've had since Ubuntu 14.04. Just install and everything works. Although Gnome was a bit of a hassle, xfce and Cinnamon run beautifully.
(comment deleted)
Both Debian and Arch are very fine distributions. Your other three options aren't really good for what Valve is going for.

Gentoo is source based, and while distribution of binary packages is possible with it, why would they do that?

NixOs ignores FHS, and while it is possible to install a steam compatible runtime, why would they do that?

And Fedora belongs to RedHat, and why would they do that, when they actually try to switch away from a OS distribution controlled by one company?

I am not saying those distributions aren't good, just that it would not make sense for Valve to use any of those.

Debian (via Ubuntu) and Arch are probably the most well tested Linux distributions by gamers, that do not belong to a single company.

Gentoo, like Debian, has always been a good distribution for derivatives. It is quite flexible and you can definitely distribute binaries in the case of a commercial derivative. Gentoo has also been used more widely in the embedded world (because it has the ability to lower its footprint). Known derivatives include Chromium OS and CoreOS/Flatcar Linux. Both of them do not use the source model.

Fedora is also very friendly with derivatives and features the ability to upgrade painlessly through atomic upgrades/rollbacks, thanks to rpm-ostree. Something that is shared with Nix. RedHat is mostly a sponsor and Fedora is community-maintained. Unlike CentOS, RedHat didn't get any bad press on how it tries to influence Fedora. Fedora is pretty consistent and quite polished on the desktop side. You also get a distribution who has access to the most influential upstream distributors (including for the graphical stack for example).

Debian/Arch do not offer such things. Arch does not even offer a stable base to start from. Taking a random snapshot of a rolling distribution needs a lot of work. Ubuntu does that with Debian. All Arch derivatives are desktop distributions. It would be interesting to know why Steam did not continue with Debian. Is it bad experience on the previous SteamOS or inability for Debian to move forward?

I think you're overestimating how much Valve wanted to customize this. I'm pretty sure that a large part of the appeal here is that it's "just regular ol' Linux".
I expect the same. They probably will not provide much support for the operating system and just redirect users to the communities.
Yeah I love NixOS, but for something like this it'd probably be a mistake.
It's based on Arch? That really is exciting. I never expected something like this.
Tech specs – https://www.steamdeck.com/en/tech

800p resolution is a bit disappointing, but I guess that's the only way to get a good frame rate without a massive power-hungry GPU.

7", bigger resolution would be useless, and a waste of ressources tbh, it's a portable console, i'd rather have better FPS and better battery life than higher resolution with lower FPS
You can plug it to a bigger screen, so a bigger resolution could be useful
but you'll have to make a bigger "portable" console as a result, and it would cost way more than $399

it's a trade off

the main selling point here is it's portable, let's not forget that

Maybe the 800p thing only refers to the console's display, as it's in the "Display" section, and it could output more resolution when plugged to an external display?

If it can't do at least 1080p on a bigger screen, it would be a deal breaker for me, to be honest. I'm considering reserving one but it would see plenty of big screen use.

It does have a bigger resolution when connected to an external display, 800p is only for the internal one. Tech specs say "up to 8K @60Hz or 4K @120Hz" for USB-C display.
It's about the same size as a large smartphone. Having a screen with half the ppi of an iPhone would definitely be noticeable.
Looks more like the size of an 11" tablet to me, but with a smaller screen. The physical interface uses most of that real estate - which is good, because otherwise a tablet would be a superior device.
What, no 5G? Just WiFi? That seems surprising.
I was more disappointed that it didn't support 802.11ax.
The site Crashed my browser (chrome) as soon as I scrolled down.
All the more reason to buy Steam Deck, I guess.
(comment deleted)
This looks great. A Nintendo switch with more than a handful of AAA titles and shovel ware and power. I’ll be buying I think since the Nintendo DS is dead, switch isn’t for me, Sony gave up. This looks like a winner.
What os does it run? And the more important follow up: what games can it therefore run?

I _assume_ it's linux, but that must limit the games to protondb and linux compatible games which, speaking from a lot of experience, misses quite a few.

(comment deleted)
If you click to the site and go to the Software tab it talks about it being SteamOS with Proton etc.
First impressions:

Really good hardware for the price.

They fixed the worst thing about SteamOS: It's now got an Arch base.

They're definitely overcharging for storage, but I imagine it'll be easy enough to modify one, so it probably doesn't matter much.

Why is any company having more USB-2 slots than USB-3 these days?

Going with a relatively low resolution compared to modern ones is actually a surprisingly smart move, given how small the display is.

USB 2 on a mobile device might be for limiting power consumption?
Nope, it's on the dock. One USB-3, two USB-2. There's no power consumption reason for the dock.
Internally, it might be a single USB 3 interface routed to a hub chip that breaks out one USB 3 and dual USB 2, or maybe even three with one of them going to a GPIO controller or some other random peripheral.
USB 3/C has really fine grain power control - using "PD" (https://en.wikipedia.org/wiki/USB_hardware#USB_Power_Deliver...), it's possible to negotiate the rates pretty finely - so I don't think it's that.
I meant that there’s probably a usb 2 controller with lower power draw reqs compared to a usb 3 controller, but I now see that the ports are on the dock like others said.
In a mobile device, it's not necessarily about the power delivered over the Vbus wire, but about the energy required for the transceiver and serdes logic in the controller. USB 1.0 "low speed" had a clock of 1.5 MHz, and USB 1.1 "full speed" is clocked at 12 MHz. You can communicate at that frequency with a microcontroller on an ancient process node on a coin cell. USB 2.0 "high speed" runs at 480 MHz. Running a processor at that frequency requires significantly more power, but is not too egregious.

USB 3 is clocked at 5 GHz, which requires more power still. Just having a transceiver capable of that frequency enabled will draw a significant load from a battery, regardless of whether you're using the bus for power delivery.

There are a surprising number of USB devices that don't work right when plugged into USB 3. I figure it's a compatibility-driven move to support random cheap janky controllers/keyboards/mice that customers will inevitably try to use.
The best part is that you don't need to buy games if you already have a Steam library full of games.
One negative: With all that hardware it’s almost 3 times heavier than the Nintendo Switch, which is already near the upper limit for holding comfortable for longer play sessions.

edit: On closer look that's comparing the Switch without controllers (297g) to the Steam Deck with controllers (669g). The Switch with controllers is 398g, so the Steam Deck is only ~1.7x times as heavy - though that's still a lot of extra weight to hold in your hands.

That’s almost a deal breaker. My hands get tingly after playing in hand held mode for too long. Not sure if it’s the weight or ergonomics.
Sounds like ergonomics. If it was weight based I’d expect you to experience pain or cramps near your shoulders or upper back, not your hands.

Tingling in your hands sounds worryingly like RSI, I’d see a doctor about that.

I wonder if they have lanyard loops so you can hang it around your neck like those heavy RC controllers xD
Tingly feelings in your body are almost always nerve related. Sounds like its pressure on a specific nerve thats causing this issue for you. Hopefully nothing to be alarmed about but that you are putting a strange pressure on a specific nerve.
Had this issue as well. I now use it as a display only and always play with a controller (tabletop mode). If I can't use a controller, I play slower games (like Civ).
I bought a "switch grip" that completely fixes the tingling and cramping for me. I have the "satisfye pro" I think. It's pretty big, but there are smaller ones out there if portability is important to you
(comment deleted)
Do you play lying down, by any chance? I sometimes get that if I do, but not seated.
Yes I think it's usually playing while pretty much laying down on my couch.
Your problem is probably not the weight of the Switch, but that you're resting the weight of it and your forearms on your ulnar nerve (the "funny bone"). Next time it happens, try lifting your elbows off the couch. You'll be supporting more weight with your muscles, but I bet the tingling would go away pretty quickly.
Ergonomics for me. The buttons are far too close to the edge for my big hands, and it's so thin and flat I have to grip it more with my outer-most fingers than a random comfortable controller.
Definitely second the recommendation for the Satisfye Pro in terms of ergonomics
It's much less than 3x the weight of a Switch. They list it at 670g and a Switch with Joycons attached is about 400g.
(comment deleted)
As someone with with incredibly poor wrists, the OG switch is already too heavy.
Unless you have an actual disability or are like 7 years old, this is trivial to fix in about one month.
Really interested in how to fix this. Strength training?
Yes! Start with low reps and weights and just do it regurarily. There are plenty of sites that show what to do if you want to strengthen certain muscle groups.

I've had success with daily low reps, but some longer intervals may work too. Once they've been primed, muscles grow when you rest, so rest is crucial.

On the other hand I find the switch pretty light.
The tech page says that it weighs "Approx. 120 grams" [1] Is this information incorrect?

1. https://www.steamdeck.com/en/tech

edit: crtl-f mistake, that's the docking station

That's the docking station.
Yes, I had scrolled too far before I used find. I was trying to figure out how it was lighter than my phone until I realized that it was the dock.
120 grams is the weight of the dock. You scrolled a little too far. From the same page, "Approx. 669 grams" for the system.
I've been thoroughly unimpressed with the build quality of the Switch, I'd probably have paid more for a sturdier 'pro' version, to be honest.
That’s what I keep saying. The switch is amazing but the quality is truly bad. Apple launching a switch-like console would be amazing.
The is zero chance of Apple ever releasing anything gaming-related
Apple arcade?
In what sense are the iPhone and iPad not gaming related? Games are 90% of the App Store’s revenue, every new i-device boasts a more powerful GPU that they advertise on stage with top gaming companies. I agree they’d probably never be so gauche as to launch pure gaming hardware but it’s clearly a core pillar of their business at this point.
Hopefully I can save up the downvotes for a third pair of replacement joycons.
I actually got the pro controller over the joycon issues. I'll be damned if I'm going to buy new joycons for the drift to happen again.

That issue alone is bad/wide-spread enough they should replace joycons for free.

The build quality is the main I never considered getting a Switch. I simply hated playing with the joycons, they felt like things you win in puff snack bags.
This is such a short-term knee-jerk reaction. We should be thanking Steam for their wonderful push to get gamers into shape. Toned arms might seem like a bit of a joke but when you consider you get your biggest exercise wins doing literally anything at all this will literally save lives. How can you not be for it? Is it heavy enough is the only question to be asked here. There is nothing valve won't do for us!
I can't tell if you're posting copypasta... Sure reads like one.
I wrote every word of it. Probably it isn't really funny at all, even the greatest comedians have misses and I don't claim to be that. But probably not less funny than your response..? Best!
What was wrong with SteamOS being based on Debian? I’ve not used SteamOS myself, but I wasn’t aware there were problems in that regard.
It's ancient and Valve had to do tons of backports to support it properly. I imagine they didn't like that very much.
Could just have switched to Debian Sid, which is basically the rolling release variant of Debian.

That could have saved some time on tooling/packaging adaptions.

Not that I dislike Arch Linux, I run both Arch Linux and Debian (Stable on servers Sid on laptop).

Sid is still far behind most rolling release distributions, and anecdotally I haven't found it more stable than any of the ones I usually use.
This. Only rolling releases are fit for end user systems; point release is a server meme.

Not only is having new stuff good from a feature and hardware-support standpoint, but, paradoxically, new software seems to have WAY less bugs than old software. (I assume this is because devs are mainly on the new versions and don't care about old versions that much.) Try using CentOS 7, with ancient, should-be-rock-solid versions of apps. Dolphin crashes every time you load into a large directory. Kate's keybinds are broken. Tmux refuses to support 256 colors. I get none of these issues on Arch or Manjaro.

My own personal experience with Arch & Debian Sid is that Sid is noticeably more stable than Arch, long term. Maybe Valve thinks it's easier to address the stability problems on an Arch-based distribution rather than deal getting the latest versions of stuff on something Debian-based?

I'm also sure that Valve's system is going to have a smaller set of software than I'd use on an Arch desktop anyway, so there's less surface area for stuff to break.

It seems to me that most people finding Arch unstable find it so because they rampantly abuse the AUR. I'm not implying this is something that you, personally, would do, but it seems to be the case in most "Arch is unstable" scenarios.
I might have one or two packages from AUR, but mostly I don't touch it. I've had a number of discussions in meetups about stability and Arch over the years and have encountered plenty of people who switched from Arch to Debian or Ubuntu because of stability problems with Arch. I have a hard time imagining that it's mostly down to AUR.

Maybe the people using lots of AUR are complaining a lot, and the people who don't just quietly switch distros?

Quite possible. The number of packages in the main arch repos is a lot smaller than the number in Debian/Ubuntu. I would imagine the kind of people you're talking about are leaning on the AUR or packages found by default (and thus more heavily tested) in Debian/Ubuntu.
Has there ever been an analysis to see what is most likely to break on a rolling distribution? I have been using Arch for years, and have only had a couple of hardware related issues. If that is typical, rolling distributions may not even be an issue when the vendor has tight control over the product and can isolate hardware related packages in their own repository.
In my own experience, the hardware support that breaks is the GPU. For me, the GPU has broken occasionally (once every couple years) regardless of distribution, unless I'm using an Intel GPU or an older GPU with mainline support.

I'm sure that Valve is going to install a much smaller set of core software than a typical desktop, though.

Looks like the Deck is using an AMD GPU, which tend to have good mainline support.
> My own personal experience with Arch & Debian Sid is that Sid is noticeably more stable than Arch, long term.

I have the exact opposite experience, I used to use sid and had to do full-blown reinstall every couple months because e.g. dpkg would break too much and I wouldn't be able to install anything, or once, even boot and ended up migrating to arch despite the warnings - never had to reinstall the distro once since then

Do you remember when that happened? I vaguely remember that I used to experience problems installing packages in Aptitude and it would make me "fix" them without a way forward that I liked, but that was years ago.
eh, I switched to arch around.... 2013/2014 ? after an especially bad crash with sid. Never used aptitude, only apt-get. Since then I'm carrying the same "distro" from computer to computer.
Aptitude is mostly a front-end to apt-get, but if you try to install some impossible combination of packages or get your packages in a weird state, aptitude offers solutions to fix it.
I've been running a mixed testing/sid install since 2005 and not once did I have to reinstall. Even managed to cross-grade from i386 to amd64 a couple years ago.
I'd been using Debian sid for almost 10 years and now use Arch for more than 5 years. Sid is definitely much more stable, as in "not requiring manual interventions", than Arch is. In fact it was actually a bit "too stable" for my taste because it effectively becomes frozen together with 'testing' right before a new 'stable' gets released.

That said, never had to reinstall either of them. Had my system broken by pacman in ways dpkg would handle much better, but it was fixable anyway.

Have you tried openSUSE Tumbleweed? It is a well tested rolling release (https://openbuildservice.org/), with snapshots on update via btrfs.

With the opi package you can install: chrome, codecs, dotnet, msedge, msteams, plex, skype, signal, slack, teamviewer, vivaldi, vscode, vscodium, zoom and more.

Brave is app i am missing so far, but having the newest version KDE and other software that required a PPA on Ubuntu is pretty nice.

And it feels much snappier installed on a SATA SSD than Mint from a NVME SSD on the same machine!

Yes, actually I ran that for quite a while successfully before I switched to Arch Linux and Debian.

The auto-snapshot feature on updates was really cool, and direct integration then (almost 10 years ago for me) was ahead of most other distros IMO.

I then setup Arch to be more confronted with the things that actually go on under the hood, learned a lot and had generally a good experience with it. A bit later I got a job where Debian plays a key role, that sealed the deal on using Debian and Arch only for me (remembering the usage of two package managers is enough for me ^^).

But I have good memories of openSUSE, and it was the distro that I first used "full time", even introduced me to Linux to some degree (I dabbled a bit around with Ubuntu and Knoppix before).

Debian GNU/Linux: operating system so conservative that “unstable” means production ready rolling release

(My go-to though)

Not really, it slows only down around the time when Debian wants to make a stable release, i.e., every two years for 2-3 months or so.

Debian is in general really rigorous with ABI/Symbol versioning, dependencies and packaging in general - there are some outliers, but most of the packages in Debian is just high quality work.

On the other side is Arch Linux, where they do not even care about Kernel ABI bumps, which are trivial to detect and track, so if you pull in a kernel update you cannot load any new module, as they just overwrite the modules of the running kernel - I used it, but I find that still lunatic!

Nothing like having some complex workspace setup and then needing to connect to a VPN/WireGuard network but also forgetting to preload that module (I just load all possible relevant on start nowadays), forced a reboot, nothing lost but tedious.

On Debian this and other things where you're in limbo while some library version mismatches, or some AUR package is depending on another one, won't ever happen on Debian. It has its disadvantages, but they certainly dot their i's.

Pedantic comment: it’s not that Debian itself is ancient, it’s that they prioritize stability and thus older packages. Both Arch and Debian get frequent updates in their packages, but Arch prefers cutting edge cool and Debian prefers “don’t break any users existing setups!”

Arch is still a better choice because gaming is generally using cutting edge software.

And yet, Debian based desktop systems are the ones that have always broken after updates for me. It was only when I started using an Arch based system that I was able to go past 6 months without any issues. I’ve been running Manjaro now on three different systems for almost 3 years. No other Linux desktop has ever lasted this long for me and I’m not even a newbie. I’ve been using Linux since Red Hat 4. I’ve actually had less problems with it than I have had with any other desktop of us including Windows and Mac.
If you're combining external packaging and debian packaging, or installing things manually, this is typical. Oftentimes if you're doing those, you'll have broken dependencies because Debian lags behind. I had issues keeping Blender and some other creative software up to date, because of this.

It works great if you mostly stick to official Debian packages through and through.

Yeah, I almost always need something from an external repository with Debian based systems.

I need external repositories (albeit rarely) with Arch too though and that has never caused a problem.

My solution on Debian has been to use Flatpak.
I was same, but now in Docker era, I no longer need to install newer programs directly on base system OS.
So Debian is fine if you want to use your desktop computer like an iPhone?
Or you can put in enough time and effort to actually read error messages and do a little work

If you want iPhone ease Linux use Ubuntu based distros. But traditionally Linux hadn’t been single click easy. Much like smartphones weren’t originally iPhone level easy.

Polite edit: if you’re a Linux noob start with a vm or live disk Ubuntu image and play around. If you like computers and understanding them, you’ll find the lessons you need as you need them by searching the web. Then you’ll install a bunch of distros and understand what I mean.

What does Valve have to backport? Surely games aren’t targeting super new libraries? (I’d expect Valve themselves to be the ones deciding what versions are targeted.)
3d graphics libraries, drivers, input libraries, etc. It is a nightmare getting the latest versions for those on Debian/Ubuntu. Valve is pushing Linux gaming forward a lot and recent package versions are required in order to run games well.
>They fixed the worst thing about SteamOS: It's now got an Arch base.

I'm a little surprised they didn't go with something like Fedora. The kernel / drivers are kept just as up-to-date as Arch but the rest of the system is a little more stable.

Aren't up-to-date libraries fairly important for steam?
I said "a little" - Fedora is still more aggressive with updates than Ubuntu or Debian. Just slightly less than Arch.

The most important libraries for Steam are things like Mesa, which Fedora updates on a rolling basis along with the kernel.

fedora will also patch stuff from upstream to include in their distro. for better or worse arch will try to upstream the patch and if they don't like it just keep an older version around if they don't accept it
Both Fedora and Arch share that philosophy, though. There are very very few patches in Fedora compared to the Debian ecosystem - and often the only exceptions are support for features that in the process of being upstreamed, like Firefox' hardware acceleration support that was added by a Red Hat engineer.
I don't know a ton about this, but I thought their major issue was games needing new versions of some random libraries. Being more up to date than Debian would still put you (presumably) years behind.

Mesa and the kernel would be less important, as all the hardware needs to be supported by them on release.

> I don't know a ton about this, but I thought their major issue was games needing new versions of some random libraries. Being more up to date than Debian would still put you (presumably) years behind.

That's not the case w/ the Fedora release cycle. Like I said, it's only slightly behind Arch.

From my personal experience not really. I’ve steam running on an about 2yo old (at this point) Gentoo system that’s still working perfectly fine. I really should update it but I don’t do much on that system except play a single game.
My point is that new games can need new libraries, so they're not really factoring in your use case.
dnf adds a layer of complexity that's really unnecessary; pacman keeps things much simpler.
(comment deleted)
Admittedly I've never used pacman, but how is the workflow more simple than "dnf update", "dnf install", "dnf remove"?

I guess I don't understand what complexity there is to remove in the first place.

dnf is a large and complex tool to manage a format that has miles of backwards compatibility (and, for extra inconsistent fun, random cases of incompatibility, since they forcibly broke spec at some point, without changing the extension of rpm, so while there's the expectation of compatibility, many old packages just won't work at all and you'll have very little clue as to why unless you're familiar with the history of the format), and has a lot more to it than pacman.

I'm not talking about the workflow for an end-user: Valve is definitely not going to force people to run terminal commands to update their systems. While pacman's update workflow is way simpler than dnf's, it's just not relevant here.

I was primarily talking about the complexity of the tool itself, of updates, and also the complexity of packaging. Arch is a packager's distro, and much of the foundation Arch is built on is the Crux-style "as simple as possible" mindset. There's much less that can go wrong when you're doing much less with much smaller tools.

I'm not hating on dnf or rpm, here, for the record. They're fine tools for their use cases.

It's all about how it works underhood. The speed of pacman is on another level compare to even dbf
I don't think either of those points are correct.

1. Fedora is not a rolling release, so you don't get every single kernel point release like you do in Arch.

2. Steam does not want to do a full Fedora 30 to 31 upgrade every six months. They want to roll. It's a better model, especially when kernel/drivers are improving gaming quickly.

3. Arch has a stable base and gets small updates every day. Regular software services tell us this is a better model than infrequent large updates. Things like RHEL are the exception, but take years of QA, and leave you with old drivers in between releases.

> 1. Fedora is not a rolling release, so you don't get every single kernel point release like you do in Arch.

Wrong. The kernel and mesa stack rolls and is rebased regularly throughout the life of a Fedora release. Fedora is prepping for a rebase to Linux 5.13 for both Fedora 33 and Fedora 34 now: https://fedoramagazine.org/contribute-to-fedora-linux-kernel...

Bingo, the kernel updates through the life of the fedora release. Then it stops. The life of a console is 8+ years. This needs to be supported for that long. Valve don't want to do multiple dist-upgrades over that time. They just want roll
I suspect they're not going to do that with SteamOS. They'll likely freeze the tree and do effectively the dist-upgrade style process themselves.
> 1. Fedora is not a rolling release, so you don't get every single kernel point release like you do in Arch.

Fedora rolls the kernel. I'm on Fedora 33 (a full release behind), and yet I still have kernel 5.12.15, the exact same as Arch currently has. I don't have any custom repos configured.

On rare occasions Fedora even gets new kernels faster than Arch, but usually it's less than a week behind.

>2. Steam does not want to do a full Fedora 30 to 31 upgrade every six months. They want to roll. It's a better model, especially when kernel/drivers are improving gaming quickly.

Fedora gets constant Mesa updates as well, it's not pinned. Admittedly it's 2 months behind Arch (21.0 vs 21.1)

Fedora updates everything that isn't a breaking change in between releases. Stuff like major gnome versions or replacing core components with similar but incompatible components gets held back for a major release.
A distro where only the kernel is rolling and cutting edge while the rest is slower paced sounds like the worst of both worlds to me.

My favorite part of Arch is the rolling packages and my least favorite is the rolling kernel. I dislike having to reboot so often.

You can add the 2 kernel packages to the ignored packages list of pacman and update them manually if needed. Pacman will warn you that it ignored a package update including the currently installed and new package version.

You will not get security and bug fixes but that's what's bothering you, I think.

Regarding that storage thing: in the hardware section, they specified that the storage could be expanded with a microSD [1] card which should already improve storage capacity by a lot if you wish.

[1]: https://www.steamdeck.com/en/hardware

(comment deleted)
Micro SD is an order of magnitude slower than the NVMe option, though.
I am not sure I can tell the difference between games that load from my SSD versus my NVMe. Is there real benefit?
So, this is actually something that is in flux right now, and the tech is shifting towards loading from storage directly into memory, like the cartridge days back in the 1990s (Nintendo 64, etc). This is in flux to the point that it's really the new games from the past year or two you'd want to test.

What caused this shift is that the current generation of consoles shipped with SSDs. The PS5 and Xbox Series X/S both shipped with NVMe storage standard. Games going forward are designed with this baseline in mind.

The reason this affects game technology is because it changes the tradeoffs. You always wanted assets to load quickly, since forever. The amount of time it takes to load an asset has two parts that contribute: I/O time and the CPU time spent decompressing.

In the past, storage was more precious and slow, so your game would load much faster by compressing the data. NVMe users would never see worse loading times than SSD or HD users, so you'd just compress everything, and the NVMe users would see a slight benefit.

Now that NVMe is a reasonable minimum requirement, you can ship a game that uses uncompressed data on disk. People with NVMe see faster load times, and people without them see slower loading times, possibly MUCH slower.

When I say "uncompressed", I just mean "uncompressed, relative to the runtime representation" which may still use compression, like ASTC.

The situation with the Nintendo 64 is remarkably similar to what's going on now... because the Nintendo 64 had such a fast storage system, games could "easily" load assets on the fly. The system had only 4MB of RAM, which seemed like a very severe limitation, but you could make very expansive areas in games by streaming assets from the cartridge while playing. It was fast enough that you could load some assets from the cartridge into RAM, and they would be available to render during the same frame! These are the kind of technological changes that are going to happen inside engines, now that NVMe storage is more common.

I put "easily" in quotes because nothing about Nintendo 64 programming is easy.

SD class UHS-I is about as fast as a SATA HDD, so it will be quite a bit slower than your SATA SSD.
> SD class UHS-I is about as fast as a SATA HDD

As fast for sequential data, but at least an order of magnitude more IOPS.

Because SATA and NVMe are both extremely fast compared to SD cards.

Just take a look at the painfully long loading times on the Switch to see the difference.

Is SATA fast compared to an SD card these days? Recently, when I've worked with newer SD cards (UHS-I and faster) I've been quite impressed with the speed in practice... and I've been hitting fsync at the end, so it's not just the speed of the cache. It no longer feels like I'm working with removable storage.

UHS-III works with Micro SD cards and it's supposed to give you 624 MB/s. SATA 3.0 is supposed to give 6.0 Gbit/s, or 750 MB/s.

That's only 20% faster. I'm sure the real number is different due to overhead, but how different can it be?

I'm not going to be shocked by low Switch performance, since the Switch is basically an obsolete mid-range phone, in terms of specs.

No, there is not. Loading times from SATA SSD and NVMe SSD is basically the same. Look for example at https://www.techspot.com/review/2116-storage-speed-game-load... - In all loading time benchmarks there is a huge spike for the HDD, and then a minimal difference of ~1 second or less between the slowest SATA SSD and the the fastest NVMe SSD. Note that this includes PCI-E 4.0 SSDs.

Sure, might change in the future, but no game right now uses the future APIs (direct storage) that might lead to bigger differences. Which is no surprise, as Windows 10 does not support it.

> All models of Steam Deck support expanding your storage via microSD cards. Games stored on a microSD card will appear in your library instantly.

This makes buying the model with limited storage a bit of a no-brainer. I'm glad Valve supports external storage.

(comment deleted)
Loading times will likely be a lot better on the high storage versions due to NVMe. And PC loading times can be brutal so I think the upgrade will be very noticeable.
Eh, UHS-1 is pretty slow compared to an SSD.

I think for the types of games I’d want to play on it (emulation) that’d be just fine, but for some it could be a bit annoying.

I think they should have gone with a faster interface though.

Sort of agree. I just purchased a UHS-1 microSD card, UHS-I U3 certified. And the advertisement only claims up to 100 MB/s speed, which is terrible in comparison to the SSDs you mention.

But then, it cost 60€ for 512 GB, and that's way more price effective for most. At least on the Switch, loading times are not that noticeable imo.

Raw speed really isn't a major factor in loading times, faster SSDs have almost no impact, almost all of the advantage is gained from better random read preformance
Looking forward, that may change. The two new consoles have shipped a pipeline that allows the GPU to read directly from storage, and the same pipeline will be shipping to Windows 11. An SD card won't be able to run games that are designed around that quick pipeline, even if they have comparably good random seek times.
You could do a small USB-c storage dongle as well.
In my opinion, unless you only want to play extremely lightweight games (like less than 50 megabytes to load) such as older retro games, the SSD is a requirement.
Arch is a surprising choice because of their rolling release schedule. How will software updates for Steam Deck work? Will Valve just snapshot Arch at a random time, and then stabilize it?
my guess is pretty much what manjaro does. i was surprised to see its running kde. i thought they just woulda launched a steam frontend and that's that.
They do also have a customized Steam frontend on top of KDE.
Yeah, it's not that hard to hold updates back a bit.
> Why is any company having more USB-2 slots than USB-3 these days?

Almost certainly has to do with cost and PCI-E bandwidth. Mobile chips like this only have so much bandwidth, and USB3 requires more dedicated bandwidth than USB2, not to mention more complicated circuitry that drives up price. And when your expected use case is attaching keyboards and mice, which generally are USB2 only anyways, it reduces the cost and bandwidth needed to support that many ports.

Presumably most people use usb for peripherals that don’t require usb3 spec. Webcams, input devices, etc.
It's not really a mobile chip, it's a full laptop-grade Zen 2 processor.
PCIE lanes are still a function of both the CPU and the motherboard chipset. It's cheaper to build a board with fewer PCIE lanes, and cheaper to have fewer components that need said PCIE lanes.
"laptop-grade" is mobile, in CPU terms. That's why Intel marks their laptop chips with an "M" at the end.
Yes and no. There are desktop computers with laptop chips and laptop chips are much closer to desktop than to a phone for example
In this context, 'mobile' does not mean 'phone', it means 'portable'.
Sure, and "portable" is a useless class of chips. It's too large to be useful.
> full laptop-grade Zen 2 processor.

That's exactly what we call a "Mobile" chip in the x86 CPU market.

The mobile phone grade chipsets you are probably thinking is probably called "Embedded" chip in the x86 market. ("Embedded" chips is wider than a mobile phone though)

(comment deleted)
> Why is any company having more USB-2 slots than USB-3 these days?

Because it's cheaper to produce and easier to design for USB 2. The cpu and chipset are limiting factors as well.

https://www.amd.com/en/products/chipsets-am4

Scroll down to the table at the bottom of the page to see what AMD's chipsets can support.

I really wonder what it means to be based on Arch Linux. Are users supposed to run packman to grab the latest security fixes? Are users going to be installing apps from the AUR or whatever? Why does this thing even need a package manager at all? Or is there more to a Linux distro that I'm not getting? It seems like a distro is mostly defined by it's package manager and repository paradigm or philosophy, with Debian being the slow stodgy stable distro and Arch being bleeding-edge. Everything else they have in common (wayland, systemd, standard components that don't know or care thst they're running on Debian or Arch).
> Are users supposed to run packman to grab the latest security fixes?

there are plenty of graphical frontends for updates on Arch.

Another big thing is more recent glibc and mesa for instance which can be huge for performance. Also, Arch makes it fairly easy to rebuild all packages with optimizations for a given CPU ; for instance someone recently made a test of arch built with x86_64-v3. Also there's the way packages are built (much simpler than with debian), etc.

Why does valve need to use a "distro" at all? LFS exists and is trivial to throw together a custom install. You can grab whatever glibc or mesa you want. Also building a kernel with improvements for a certain CPU is dead simple out of the box with Linux. Distros make it hard, and Arch makes it "easier".

As for packages, why does SteamOS need packages? They can just make one large update package that includes all library updates for the latest SteamOS version. Or are they expecting users to say "Oh I want to update just this one library"? Seems like a huge missed opportunity to make SteamOS a "stable target" that Linux devs can make games for. But if everyone is like "I want this weird glibc!" then that's impossible.

Valve probably does not want to maintain their own distro. Using Arch as a core with their own packages on top is far easier and ensures you always have the newest core components.
What does it mean to use Arch as a "core"? Can't they just pull a list of packages that Arch uses (or the latest Ubuntu for that matter), and use those in their Linux insall?
sure but doing that with pacman is trivial... and if you're using pacman and arch packages, you're pretty much using arch
There's unattended updates and GUIs for everything, bruh.

The adage that you have to use a terminal if you want to run a distro has been out of date for over a decade.

Oh oops, when I said "pacman" I meant "or a GUI on top of it". I figured that would be obvious.
I would have thought Fedora Silverblue would be the perfect OS for this kind of thing. The OS is a read only image you mount and to update you simply mount the newer image. Its the same model mobile OSs use and it means update failures are almost non existent.
My assumption is that the intended use of the USB 2.0 ports are mouse + keyboard.
> Why is any company having more USB-2 slots than USB-3 these days?

I think it's not a case of "having more USB-2 slots than USB-3 slots", I think it's instead a case of "having precisely two USB-2 slots", no matter how many USB-3 slots there are. The reason for exactly two USB-2 slots is obvious: one of them is for the keyboard, the other one is for the mouse. Neither the keyboard nor the mouse need more than USB-2, so there's no reason to have the more complex USB-3 hardware for these two ports.

Other than those two ports, it makes sense to have the rest of the USB ports be USB 3, which seems to be the case here, even though there's only one (the USB-C port seems to be meant to be always plugged into a charger, so it might also be USB-2 only).

I agree on the reasoning, but at the same time, they might as well have just gone with a single USB-2 port and had people buy a <$10 USB hub off Amazon or something (or offered it as a peripheral for $20). Although, you can buy a few port USB-c hub for almost the same price, and if they just had two USB-c ports it would all work just as well with the addition of a hub (since you only need to add a keyboard and mouse when stationary).
> Why is any company having more USB-2 slots than USB-3 these days?

Most of the prospective users don't need more USB-3 ports than they have.

In my experience USB-3 generates interference that messes with wireless gaming device dongles. I use a Logitech wireless headset and gamepad and they work far worse in USB-3 ports.
Sort of an unfortunate name since it's so similar to "Stream Deck." I imagine there might be some confusion between the two if this takes off.
I initially clicked this link thinking it was a Valve-branded Stream Deck. Only then did I realize it was a Valve-branded Switch instead. :P
There’s obviously a lot of people who like playing games in a more mobile format, as evidenced by the huge popularity of the Switch Lite / mobile gaming.

Giving the PC games market access to that form factor seems, on its face, like an extremely good move.

But we’ve seen products like this before — the NVIDIA Shield, for one. Pretty cool piece of tech, but didn’t exactly start a revolution.

I guess the question becomes “what percentage of current PC gamers are motivated to shell out $400 to play on the train?”

Frankly, I feel like it’s probably a decent number of people? Enough for this product to do okay, if not change the whole market. But people have been very confidently incorrect about almost every iteration of mobile gaming in the past. I guess we’ll see.

The Shield was Android with GameStream. This will run games directly on x86 Linux which, with all their Proton work, will result in a much bigger game library out of the box. There have been more niche handheld x86 devices, but they were even more money, and different form factors. Steam being as big as it is, I think this has way more potential than anything done before.
This is different than the NVidia Shield handheld in that hardware has come a loooong way since 2013 and it’s actually possible to run modern AAA games at 1200x800 using what is effectively a mobile chipset. No streaming from a PC required.

If I played more games I would absolutely buy one of these. I already own the NVidia Shield that plugs into your TV because it’s a great streaming box even if you ignore the gaming aspect.

You have to consider that PC hardware is currently spiraling out of control in terms of price, and $400 for the specs it has presently is a fairly decent buy even if you don't care about the form factor.
That’s a really good point, viewing this as an accessibly-priced entry into PC gaming makes it even more attractive.
If current trends around availability of discrete desktop graphics cards hold out, then it'll be even more attractive.
PC hardware prices will affect this as well, since this is based on shortages and high demand... basically these will be almost impossible to get outside of secondary market.
I'm calling it now, people looking back on this thread six months or a year from now will have found your speculation false.
I gain/lose nothing for being right/wrong, it's just the status quo.

I'm waiting to buy a laptop I want for some months now, still not available.

Unless this custom APU for Steam Deck is made in a special foundry with a special process, then they will be in line just like everyone else.

Plus scalpers love this shortage to make easy money.

You don't think that Valve was waiting until the supply chain was ready before making this announcement?
You think Sony Computer Entertainment didn't wait until the supply chain was ready before launching the PS5?

It's not about being ready or not, it's about supply and demand.

An example from Valve themselves from early this year regarding Index, an interview with Gabe:

“We actually have components that are manufactured in Wuhan and when you’re setting up your manufacturing lines it doesn’t occur to you that you’re suddenly going to be dependent on this peculiar transistor that’s sitting on one board that you can’t get,” Newell said.

“Everybody ended up running into the same problem simultaneously — you go from, ‘Oh, we’re in great shape,’ to, ‘What do you mean Apple or Microsoft just bought the next two years’ supply of this just so they could make sure they aren’t going to run out?’ You went from a situation where everything was getting done just in time to people buying up all the available supplies.”

Newell says these constraints are also why the headset still doesn’t ship to some markets like New Zealand or Australia. “The only thing keeping us from shipping in New Zealand at this point is just getting enough of them made — we’re very much manufacturing constrained.”

[1] https://uploadvr.com/gabe-newell-index-supply-shortage/

I can buy a decent gaming laptop right now, but it's much harder to get a desktop GPU. I'm guessing all stock goes to OEMs.
I wonder if they're selling at a loss? Valve can afford it, plus it'd be no different than Sony/Nintendo/etc subsidizing hardware since they make most of their profits from software sales.

Only difference is that, unlike traditional game companies, this Steam Deck thing isn't a walled garden. But of course, Steam has a near monopoly on PC gaming so it's unlikely that'll be a problem for them.

I don't see why they would, they'd be on the hook for service costs as well.
Similar products (look up GPD Win for example) are roughly in the same ballpark cost wise, and that's without much economy of scale.
I doubt it.

If Valve isn't locking the hardware down, but is subsidizing it, then they might run into a situation where cryptominers (or some other group) start buying these things up in bulk, but never playing any games on them.

I could see owning one of these as a non-gaming desktop.

For what coins is PC hardware competitive with ASIC?
Ethereum, because there are not ASICs for it. See the news about recent crackdowns on mining in China, that's where all the GPUs were at, mining Ethereum.
I'm just making up hypothetical my friend. I can't really predict what these might be useful for as an alternative to gaming, but the world is full of bright people who exploit any opportunity.
They actually did a pretty smart thing, where you can only order it if you have a steam account that has spent money on steam before this month. I would guess that they also limit the amount of Steam Decks you can order as a single person so it would be some effort to start buying these things up in bulk.
Genius. Man everything about this excites me for the first time in a long time, about anything
To be honest, I don't think your assumption is that far off. The way that Gabe talks about this device, I feel like the purpose is simply to extend the PC gaming market, which is essential for Steam.

https://www.youtube.com/watch?v=4FXgDAF6QpM

Nintendo sells its consoles at a profit. It's why their hardware is always "behind" the other guys who are, as you say, selling you $4 of power for $3 and hoping to make it up over time on ecosystem purchases.
I would bet that they do sell at a loss. There's an interview with Gabe[1] in which he keeps repeating that the price point was painful. He is very vague about it but he also mentions some calculus like "they'd have to sell 8 games to be profitable". So I have no doubt they're doing it like the consoles do it.

[1]: https://www.youtube.com/watch?v=4FXgDAF6QpM

PS5 and Xbox are using PC hardware too. The only difference is that they might have secured some supply for those customized parts in advance.
I'm willing to shell out so that I don't have to buy all the games I already paid for.
It’s more less the same price as Nintendo switch and the switch is limited to expensive Nintendo games. Now if you are a fan of Nintendo games, great! But if you are a of gamer then this steam deck sound like a better idea!
Had to google what you meant by NVIDIA Shield (wrt portable gaming): https://www.amazon.com/NVIDIA-SHIELD-Portable-pc/dp/B00E3667...

One major difference (besides the form-factor obviously) is that the Deck doesn't require WiFi + desktop to play your PC games. I'd also argue that the trackpads will be a major difference. The Steam Controller's trackpad wasn't perfect, but it made possible to sanely play a ton of designed-for-mouse games

The Nvidia Shield ran Android. So either bad Android games locally, os streaming. The Shield might do better today as a client just for Stadia, Xcloud or PS Now.

However, this is far better. Not only are the specs very impressive for the size/power, but it runs an OS that can plan anything (with Proton). It's like mixing the switch with a ps4. Imagine playing the Witcher 3, or Doom Eternal on this compared to the 480p blurry Switch version. I hope this does well.

>I guess the question becomes “what percentage of current PC gamers are motivated to shell out $400 to play on the train?”

Yup.

Before clicking the "Reserve" link, I put a personal threshold for this thing at $300. If it was somewhere around that, it'd be an instant purchase for me.

Over $500 for a version with barely acceptable (by 2021 standards) storage? No thank you.

The 64GB version is almost a made-for-landfill device. I can't imagine people not running out of space very, very fast (even with microSD card slot, which we may or may not be able to install games onto: I wish they made it clear in the specs). Still, for $300, it'd be an instant buy for me - just to be able to play in bed.

I really hope enough people buy this thing, because I'd really want it to thrive - and the price to drop :)

Most blockbuster games are over 64gb on their own these days. You couldn't install just one of: GTA5, Red Dead Redeption 2, Cyberpunk 2077, Deus Ex: Mankind Divided, Forza Motorsport/Horizon, Fallout 4, the list goes on.

At least Todd Howard can't sell Skyrim yet another time, it's already in my steam library.

Any game devs here willing to share how large the code/compiled code output is for these games?
You mean code without resources?

The required storage is listed on the steam store page, Cyberpunk for instance requires 70 GB free.

Yeah - the actual “meat” of games of this scale, I guess most of that is video and texture images or similar
Yeah 3D models and all their LODs, textures with all their layers, (one surface could have half a dozen layers used to render it), the myriad audio and music files. Games can be really humungous orchestrations these days.
>>"augment your built-in storage with a microSD card and fill it up with even more games."

I think that hints that you can install games... hopefully :).

SD cards are slow. This is important for large size games that need to load large assets. Loading times will be long.
> The 64GB version is almost a made-for-landfill device.

I felt that way about the Quest 2, and opted for the maxxed-out version. As it turns out, I use it a few times per week and have ~40GB of storage in use.

I originally used Virtual Desktop to remote in to my desktop, mostly from across the house but also over 5G. Now that Airlink is available in beta I use it exclusively at home, and mostly play "PCVR" titles wirelessly.

The Steam Deck seems to be taking the same approach. I see myself using it to play games from my desktop library while in bed or sitting outside. Few if any games will actually get installed on the device, because I won't want to use the on-board processing and drain my battery anyhow.

To be fair, many Quest games are very small (1gb-5gb). In contrast, GTA V on Steam is 72gb.
(comment deleted)
(comment deleted)
I don’t get this, if you’re using it for gaming then storage isn’t a huge deal. Even on a dedicated gaming PC, most people don’t have their entire Steam library installed at once.

Not only is it just too much space, but it’d be wasted since there’s no way you’re going to play all of those games.

The 64gb version of this definitely is too small though. That’s less than the average AAA game. And while there are certainly people out there who don’t play AAA games, I don’t think the average Joe will do the math ahead of time when buying one of these, and will be very disappointed/angry when they find they can’t play the game they want (without using external storage/a dock).

But, if this thing does have an SD card slot (idk, I haven’t checked), it should solve the problem. SD cards aren’t super fast, but the Nintendo switch is proof that they’re at least good enough for gaming.

It does have an SD card slot.

However, when the hypothetical average Joe you propose picks up a cheap 256 GB SD card off of Amazon and it takes five minutes every time they reload from a checkpoint in Control, I think they will be at least as frustrated as when they ran out of space on the internal SSD.

I have a single game that clocks in at 374gb. (Ark + DLC and 50gb of mods.) And I don't want to be limited to a single game. I am really hoping the 64gb can have an nvme drive added.
My guess is that to get the entry level price as low as possible, the BOM is cut to the bare minimum. That would probably imply that they preclude including the NVMe header - but who knows?
Since it runs Linux I don’t see any reason you couldn’t put your entire library on an NFS server in the other room. If I get one I’ll certainly be doing that
If you're at home, you might as well just stream from a gaming PC anyway.
Sure, if you have one. I don't (I play games in bootcamp on an old MacBook Pro) but this perfectly fits my needs. I don't have anywhere to put a gaming PC in my apartment and have been considering getting a switch, but this is basically my dream machine.
> (even with microSD card slot, which we may or may not be able to install games onto: I wish they made it clear in the specs)

https://www.steamdeck.com/en/hardware

> All models of Steam Deck support expanding your storage via microSD cards. Games stored on a microSD card will appear in your library instantly.

Anyway, you can more or less do that today already with external disks, etc. You can certainly connect a USB3 SSD. I wonder if the nvme drive is user-replaceable? I guess not?

Though given the open nature of the device, expect mods sooner rather than later. Replacing a chip and telling Linux to override its size (or mod the bios/dtb) should be a breeze.

I honestly can't believe the 64GB one isn't 128GB instead. I looked through my steam library and I have a lot of smaller games that would fit just fine on that.. but 1 AAA title would completely fill it. Plus, you don't get that whole 64GB, the OS has to take some of it. After the formatting and OS I'd be surprised if there was more than 50-55GB free space.
Personally I find a lot of these devices backwards minded. I'd much rather just shell out for a service like Stadia (if the catalogue was larger) and turn my actual mobile device into something capable of playing pc titles. Especially if the implementation is platform agnostic like Stadia. Gaming devices of all stripes have become expensive enough that I'm more than happy to see cloud gaming takeover and leave hardware for the activists and bitcoin miners.
Your phone lacks the controller interfaces these devices have though. I guess it depends on if you can hook up a controller to your phone, or if you're able to play the games you like without to much fuss on your phone screen.
You can attach your phone to a Stadia controller.

Disc: Googler.

Matter of taste. A pc gamer will hate a mobile control scheme but I can play Cyberpunk fine on my S10 with the touch controller and if it had even the level of customization that CoD Mobile or Pubg Mobile offer the player I'd go so far as to say I could be happily comfortable. Imo, people underestimate both how poor the quality offerings are on mobile and how far ui control design has come (in the few top tier offerings that exist). Cloud gaming kills two birds with one stone in that it suddenly adds a plethora of games to mobile that are good in the real way while simultaneously mitigating the mounting costs of being a pc gamer (maybe consoles too but I can't speak to them) by letting me run high-end games on average hardware so long as I have a good internet connection.
I tried pubg mobile and a few other mobile FPS games. There's one major flaw with using a touchscreen for those kinds of games, you need to be able to aim, move and shoot at the same time. In my experience, touchscreens only allow you to do two of those things at the same time.
It's not perfect but this, speaking as a casual observer, is something I think will improve with time. It took far too long to move away from fixed d-pads and now more games are experimenting with on-screen zones you tap to fire. There's still a lot of unwieldyness exactly as you noted but I think this is a surmountable challenge.
I'm not sure how this could really be surmountable. With a controller you have access to at least 6 simultaneous inputs, more if you do some finger claw tricks or have back buttons or something. A keyboard and mouse gives you around the same or more simultaneous inputs. On a touchscreen you can make 2 simultaneous inputs.

Most action games require at least 3 simultaneous inputs to be played effectively.

There's definitely workarounds like the ones you mention, but unless a game is actually made for touchscreen controls, as in optimized for no more than 2 simultaneous inputs, it's not going to be as playable on a touchscreen as it would be with anything else.

Unfortunately, some genres of game just aren't as suited to only 2 simultaneous inputs.

There's lots of fun touchscreen based games, but they're always better with game mechanics specifically suited to touchscreens.

Apple had a solution to this problem years ago but made the very very very stupid decision of removing 3D Touch from recent iPhone models.

On phones with the feature, you can use your left thumb to walk, right thumb to aim, and 3D touch (press down hard on the screen to shoot). I know at least call of duty mobile supports it, and when I tried it out it felt revolutionary.

But rather than add new input methods, they’re removing them at the same time they’re trying to push harder into mobile gaming (with Apple Arcade).

It’s baffling as an outsider to see such obviously bad/stupid decisions like that from such a massive company.

I don't game on mobile, but wouldn't back-side tap be the obvious solution? I've never tried but I'd expect the accelerometer to be able to detect something like a middle finger tapping on the back of the phone with sufficient precision and latency (just the binary event, not the location)
I do remember a game that did something like that many years ago (posted on reddit I think?). It was an endless runner type game with a fox character IIRC, and you had to alternate taps on the back of the device to get it to move. I do remember that I couldn't get it to work reliably though.

Tapping the back of my phone now, I can see it maybe working for some types of games/shooters where you don't have to hold down the fire button, if implemented well (which probably isn't easy to do for all hardware). The tactile feedback of smacking the phone could help provide a good experience if paired with appropriate audio/visuals.

This isn't quite the same thing, but the Playstation Vita has a rear touchpad, and not a ton of games managed to use it well, even with the much higher precision (Tearaway[1] is the only good one I can remember).

Tearaway gameplay: https://www.youtube.com/watch?v=Jw5LD4B-3DI

Rear touchpad is that one feature that amazes me every year by not being introduced by any of those very many manufacturers desperate to have a USP.
Rear touchpad means you can't put it in a protective case. Unless you can figure out a way to make one that protects it from most of the kinds of drops a phone would take in normal daily use and still exposes the rear pad.
3D Touch was the stupidest idea apple had in the last decade and I don’t say this lightly given there’s also the Touch Bar, the trash can Mac Pro and canceling of the apple display.
> I guess it depends on if you can hook up a controller to your phone

ps4 controllers are generic bluetooth gamepads that I've been able to connect to a surprising range of devices.

Why couldn't you use this with Stadia?

The price isn't out of line even if you consider it as a "thin client" gaming device.

It's not that I couldn't it's that I wouldn't want to with an agnostic service like Stadia. That's money I could put towards a nicer smartphone or into nicer non-internal pc components like the monitor instead. But I also say this as someone who doesn't mind touch controls when they're done well so having physical things to press isn't as big a draw for me as I imagine it is to their target audience.
Playing PC games on a cell phone is painful. You can't even see half the game because your thumbs always obscure the screen to operate the touch controls.
It's interesting timing because the Switch Oled edition was just announced with no bump in power. So if you want a capable handheld with access too all the indies which will play games much better than your switch... here you go.

And you can still keep your switch for nintendo exclusives because they've just made it clear that they're not upgrading the switch in a worthwhile way for at least a year.

I would say this is squarely aimed at the existing switch customer base.

The Switch power bump would've probably only made sense in 4K docked gaming anyways, and people are probably not going to buy one of these for docked gameplay.
The biggest disappointment in the new switch is the lack of native bluetooth audio headphone support and becoming less picky about its hdmi travel adapters and chargers
> I would say this is squarely aimed at the existing switch customer base.

PC gamers and steam gamers are often very different markets. Nintendo, for decades now, was about fun, not power. I'm casual gamer, just for fun, and steam deck, while super cool, is certainly not going to replacement my switch.

An awful lot of indie games release on PC and Switch nowadays, really the switch supplanted xbox and ps in terms of where a lot of indie games sell.

So certainly a lot of switch owners are playing indie games that they could be playing on this device with better performance.

That's great, but "Nintendo not being about power" doesn't change the fact that many of my 40$+ Switch games barely run on the device and run at such low resolutions that it feel like looking through an oily glass.

Seems like something is very wrong with either Switch hardware performance or Nintendo quality control.

After playing No Straight Roads on the Switch, I feel the same way.
I started Outer Worlds on PC and moved to switch and holy crap it looked and ran like garbage. Skyrim as a much older game looks great though.

Also many switch ports are abandoned by the devs in a very buggy state. This will probably replace my switch and PC. Very excited.

It will replace my Switch. I mostly play indie games, but the Switch can only play those indie games that get released to the Nintendo store, and then, usually at almost double the price of what you can get them elsewhere. As for experimental games that never even get sold? Not likely to happen on the Switch, there's just not enough incentive to publish.
Its still painfully low spec even for casual gaming. Sitting for 5 minutes waiting for Animal Crossing or Crash team racing to load is not fun for anyone.
Steam does have one huge thing over Nintendo. The PC platform is so backwards compatible that I can play games from the 80s on a brand new computer today.

Nintendo charges you money for new copies of each game you want to play on their latest hardware. You can't just pull out your old copy of Super Mario Bros and install it on your Switch.

Hell, I've heard that there are Switch games that have been out since the very beginning of the console that are still full price.

I'm guessing you could run Breath of the Wild better on the Steam Deck with CEMU emulation than the Nintendo Switch.
But we’ve seen products like this before — the NVIDIA Shield, for one

Atari Lynx. Sony PSP. Sony Vita. Nintendo Switch. Any others I don't remember?

The inimitable Nokia Ngage!
Sega Nomad. Sega GameGear.
Hey, with the right set cargo pants you can carry as many batteries as you want!
>But we’ve seen products like this before — the NVIDIA Shield, for one. Pretty cool piece of tech, but didn’t exactly start a revolution.

People forget that these devices without marketing and consistency they won't go far.

Nintendo, Sony, even MS that came later to the party have hundreds (if not thousands) of millions poured to build their brands, over decades... Sega knew how to play this game and still didn't manage to hang on...

They are already taking the crops of nostalgia, that's for how long they have been around. Leveraging these devices base on hardware it's nothing in the great scheme of things.

> Sega knew how to play this game and still didn't manage to hang on...

I see what you did there :)

https://en.wikipedia.org/wiki/Hang-On

Honestly wasn't with that intent, but thank you for point this out - won't forget it and will use it in the future!
> People forget that these devices without marketing and consistency they won't go far.

But seeing Valve's track record, marketing isn't their strong point too.

> I guess the question becomes “what percentage of current PC gamers are motivated to shell out $400 to play on the train?”

At first glance at least, it seems like it may end up being much more than that. I see appeal here for non-gamers as well - or at least, for people who aren't only or even primarily gamers.

It's a very nice cyberdeck for general use, though granted it doesn't have a hardware keyboard. Given the size and form factor any keyboard they'd tried to graft on the thing would be basically unusable anyhow so I'm glad they didn't try.

I can totally see myself picking one of these up and throwing it in my bag alongside a small Bluetooth mechanical keyboard. For working in coffee shops in a terminal it'd be just fine. Connect it up to a TV via Chromecast or something and you've got a damned nice setup for less than a grand.

Id just want a smaller laptop at that point then?

I find its really hard to beat the convertible laptop form factor in general

It is a laptop with less than 1kg in weight, with embedded controller support and touchscreen.

Now if it can somehow act as wacom tablet I'll instantly buy it, though it may be not the case.

The extra keyboard is going to add a bunch of weight, probably more weight in total than something like the x1 carbon with a bigger screen and battery that weighs 1.13kg.

I call it the ipad effect, where an ipad + keyboard of the same size ends up being heavier than a laptop of the same size.

Convertibles also tend to have a touch screen and pen support too.

I find the 7" form factor only good for personal usage, if you want to play a multiplayer game on it, it's too tiny usually.

The mere fact that NVIDIA Shield Portable didn't take off doesn't tell us much. It had poor design sensibility (i.e. it looked like ass, like some kind of portable Xbox designed in the early 2000s), and it ran Android, which lacks a compelling games library.

Steam Deck at least stands a chance, although I'm worried about the weight and the battery life.

The battery life is a straight up lie. 8 hours on a single battery charge? What game were they testing, Solitaire?
2-8hours.

If it can do low-resource gaming for 8 hours, its not lie. And honestly, thats not something too far-fetch.

The only one lying here is you, since its not even released yet for you to test the claim. Not to mention how you conveniently left out "2-" from the battery life range

What? Did you expect to run Cyberpunk 2027 for 8 hours on single charge this machine?

I wonder how really portable this thing is; it has PC hardware, a screen and a small factor in which to stick batteries: how long will the battery last?
I think a better question is will games traditionally built for PC's hold up on a mobile device?

Mostly I'm thinking about controller vs keyboard+mouse. Typically you have more controls available on PC and when I play the same game on PC I am complete garbage at it because I'm used to controller, maybe I'm too old though.

Valve has been pushing their Steam Controller for quite some time now, it sits squarely in the no man's land between clear success and total failure. The core audience of this new product will know quite well what to expect.
The videos on the site show people playing Doom Eternal (FPS), Crusader Kings III (4x strategy game), Balder's Gate (top down RPG), and Disco Elysium (point-and-click adventure game). It looks playable to me.

Plus it natively supports "mouse" controls via a few input methods: thumb touch screens and tilt-o-whorl. And if neither of those work, you can use a bluetooth mouse/keyboard.

I'm absolutely in the market for this. I'm fixing to move six time zones from my friends, who all play. Instead of buying a gaming laptop to take with me I'd just buy this if it's let me play with the easily.
> I guess the question becomes “what percentage of current PC gamers are motivated to shell out $400 to play on the train?”

Does it even matter? For consumers, they get what they paid for even if Valve shows zero commitment as long as Steam does not actively shut them out. For game companies, support for controller and Linux/Proton is valuable with or without Steam Deck.

And for Valve itself, all the software investment is a long term hedge against the existential threat of Windows failing them, this is just an opportunistic bonus use. The only actual cost is the hardware side of the project. If it tanks in the market they'll simply refrain from doing a v2 and do damage control with long tail sales of v1. While taking comfort in knowing that the Deck hardware will at least have lured a few game companies into the arms of Proton, strengthening their Windows hedge.

> the NVIDIA Shield, for one. Pretty cool piece of tech, but didn’t exactly start a revolution

Other than emulators, Android is a barren land for good controller based games. Specially compared to Steam.

This has so much more potential because its a desktop OS. You could e.g. play WoW using this. Not to mention the huge selection from emulators. I've tried gaming with Android before and the game selection is awful. Realistically, the only good gaming options are from e.g. snes emulators. On top of that, most phone GPU drivers are horrendously broken when doing advanced 3d stuff so even if a good game was released for android, you wouldn't be able to play it on 80% of phones.
I for one love my Shield and would upgrade / buy again - real-time upscaling/upmixing makes YouTube into a competitive "channel" I regularly watch on my 4K TV.

The shield runs all the major streaming networks and there's now Airplay apps that run in the background, closing that gap.

As with all chrome-based devices, the Google voice assistant blows away AppleTV and Roku.

This is my feeling. Money is no problem for me, but I'm not immune to purchasing stuff I don't end up using. So I'm cautious about that and I'm sitting here wondering about the value. Say, the $650 model.

On one hand it would be nice to have my Steam games on a portable without having to double-purchase(a problem ATM with Switch, on top of the higher typical costs of those games on eShop).

But I have the Switch and Switch Lite already which plays Nintendo exclusives. Honestly it's nice to get a break and dip into a catalogue that's not on my PC when I'm on the go. And I'm 100% not going to take two portables with me anywhere.

Then there is the cost. $650 for me would be better spent toward a PS5 if I could get my hands on one. I like the Sony exclusives(GoW, Last of Us, etc) and would really like to play Returnal.. Though it will probably go cross-platform before I ever see a PS5 with my own eyes.

The more I think about it's price point and value to me the more I get the Shield vibes, and the more impressive it is how well Nintendo knows their customers and rides a very fine line with their product placement.

Edit: As an added thought; I like to experience certain games, shows, and movies in a very controlled environment. I won't play/watch them on-the-go.

I view this as more of a tinker device. They’ve stated you can install anything on it. I just know I haven’t regretting buying the steam link or controller, so I may give this a go.
When comparing value with the switch, factor in:

1) Nintendo switch game prices

2) The size of your existing PC-compatible library (including emulators).

That drastically changes the value proposition to me.

I for one own a switch and has been waiting for a better quality product, I've also been looking to get a desktop gaming PC but the price + the space it takes has been putting me off for a long time. So this steam deck is an insta buy for me.
I believe it's more about gaming on sofa than gaming on train.
I intend to buy and use it as a mobile workstation for programming, since it has a x86 chip, 16Gb Ram, Linux-compatible gfx chip, decent amount of storage. Basically this will fill the void left by the Nokia N900 for a lot of us. It being able to run steam/games is secondary for me.

Or, if AMD would please get on with it and release a small form factor like the Intel Nuc, that would be great. But this device does have a screen attached to it and a battery, so it would work fine as a nuc for me.

Super excited!

Whoever named this an idiot, my first thought was reading “Stream Deck” and I thought it was streaming device. Valve makes good hardware but their marketing is terrible leading to abandoned hardware that enthusiasts rave about. The marketing page is white... it is Valve... why is the marketing site white?!?!
I guess that would be a reasonable complaint if it wasn't so that Steam is already Valve's main product, and that this device is intended for playing games from Steam!
This seems more like an opportunistic rant about Valve's marketing than a sincere response to the particular situation. The name makes perfect sense given its a slab that plays steam games and if you'd pay attention to the website beyond the background colour (???), you may notice how it has several pictures, animations and videos illustrating the product in far more detail and clarity than many console launches. It worked for me anyway because I studied the site cover to cover so to speak.

Granted, I agree that Valve did a genuinely poor job with the Steam Controller and Steam machines but in the former case, it was because even Valve themselves didn't quite understand what they had just made and in the latter, things were complicated by having to deal with multiple third parties.

The button placement seems really high, but it might be comfortable to play while resting the bottom on a table.
Huh yeah that is odd. Valve is typically pretty good at ergonomic hardware, maybe the buttons aren't the primary interaction tool for this.
I just hit the site on my desktop. Like the steam controller before, it's got a lot of back buttons that you can likely remap (R1 to R4 and L1 to L4). So yes, the front buttons are backups.
Judging by the image in the "Speeds and feeds" section, there are only four shoulder buttons. Also, looking more closely at the placement, it looks like they're putting more emphasis on touchpads over the joysticks, and certainly over the D-pad, much like the Steam Controller.