What Valve got right and wrong with the Steam Machine | Hacker News Reader

What Valve got right and wrong with the Steam Machine

polygon.com

63 pointsbpierre68 comments

Overall I love the Steam Machines idea. I have many games in my library and I can't wait to play them in the living room with my friends. I think that what Valve got wrong was just two topics:

- There is no game compatibility information. When I buy a console game, I know that the game will be compatible with my system. There is no such guaranty with Steam games. I would like to see a simplified benchmark tailored to pre-made Steam Machines so that consumers can understand how well they'll play their AAA games.

- They lost momentum with the SteamOS platform, and many developers are now skeptical of porting their games to Linux. The Steam Machines were supposed to be launched one year ago and with them their Linux platform. Some developers took a big investment to port their games to Linux and fight around subpar OpenGL drivers. When they were done there was no Steam Machines and they were met with a Linux userbase of just 1%. Now, I believe, we'll only see more Linux launches once the developers trust the Steam Machines as an established gaming platform.

Edit: formatting

Does the second point matter any more? Most of the big engines I'm aware of make portability to all three (or even five) platforms effortless, so there's really no excuse to not having a version for each platform any more...

Many good games do not use unity/unreal, as whilst they are packed with features, they are at their core flawed. Many people build their engines tailored to their game, in order to get the best performance. In those cases it is 'effort' to port to Linux, as with any other platform. If steam machines don't flop then I expect Linux porting will become more mainstream and the process easier.

Increase in continuous integration infrastructure, additional testing scenarios, new deployment configurations, new driver issues to deal with,...

...whole new market...

It already exists, there are plenty of companies specialized in such porting action.

However, the question is the ROI in supporting the platform vs sales, even if the code doesn't need any major change.

If you already have a PC why not just use Steam's built in home streaming? Low powered box in living room, desktop pc accessed remotely

How much does the low powered box in the living room cost? The Steam Link is supposed to address this ($50) as most don't want to, or can't afford (or justify the expense) another PC for the living room.

Well, for me, $50, reused spare parts. Didn't know about the Link. So the Steam Machine is supposed to be for people without a desktop, to compete with consoles?

Or for people without a gaming desktop, but yeah, more or less. I imagine there are a lot of folks who have a computer running Windows already, but it's a 4 year old Dell with an Intel HD 3000. So they'll play a lot of the Indie games, but not the AAA graphical showpieces that advertise better. And IMO a lot of those indie games (Bastion, FTL, etc.) are better on iPads than computers/consoles anyway.

Outside of the tech crowd, actual GPUs that can compare to current-gen consoles aren't very common. Steam Machines are an attempt by Valve to widen their market and include more customers who want an off-the-shelf "it plays games" experience.

Steam Machine = console replacement

Steam Link - Stream your PC games from another Steam PC on the network

The former is several hundred dollars and competes directly with Xbox, Playstation - the other is only $50.

You still have to deal with trying to play games that were designed around a desk setting, with a keyboard and mouse, in a living room setting.

You can fix this with a decently sized dinner tray, like the kind you get in hospitals.

> a Linux userbase of just 1%

My impression when it first came out was that Linux game sales quickly surpassed Mac ones. What actually is the case with Linux as gaming platform in terms of sales revenue rather than percentage of users.

For the game I have on Steam[1], a logical puzzle game, this just isn't the case. Mac sales are at about 10%, and Linux sales are at ~3%. However, the most outspoken fans of the game, and the ones from whom I get the most nice emails from, are mostly Linux users.

[1] - http://store.steampowered.com/app/386900

I look at the console world and feel like I'm taking crazy pills.

The massive reduction in cost of hardware has enabled every device to get smart. And, surprise surprise, the first things to get smart were the things with screens - phones and TVs.

Video game consoles (and all set-top boxes) hearken back to an era where the average house only had one TV, and it was dumb.

These days, there's a dozen screens in the home, and all of them have computing power many times greater than even the original XBox - more than enough juice to power a huge catalogue of existing games.

So why do big game companies (Sony, Microsoft, Nintendo and now Valve) still feel they need to make hardware?

And I'm not suggesting that these companies give up their 1st party platform advantage - just that their platform doesn't need to be hardware anymore. A cross-platform Nintendo App that runs the Nintendo VM and stores all of your games in the cloud makes a lot more sense than building a second rate piece of hardware and managing the supply chain to get it to the living room.

The only area where you can argue hardware is still necessary is the controller - afterall, our hands haven't changed that much and immersive gaming experiences haven't adapted well to touch screens. But surely it'd be easier to manage a supply chain for bluetooth controllers than it would a console.

If you add cloud graphics processing to the mix to overcome the horsepower issues of lesser hardware, something NVidia and Sony have both demonstrated at this point, then there's absolutely no reason to continue to push hardware other than inertia.

I recoil from this idea as it pushes further in the direction of free play homogenized games. Cloud graphics seems like a subscriptions based nightmare, or a world where the game no longer works after 5 years as the cloud services no longer exist.

Maybe it will go this way, but I don't have to like it.

I recoil from this idea as it pushes further in the direction of free play homogenized games.

This may be true, but we won't know for sure until we figure out how to make the AAA games as ubiquitous as Candy Crush and Game of War.

Hopefully, that will shame the market into demanding better gaming experiences from every developer.

"Cloud gaming" has plenty of issues with performance and latency (see the now-defunct OnLive) and is too reliant on a user's internet connection. TVs may have better hardware than 10 year old consoles but certainly aren't anywhere near current generation consoles.

It might work for low end or previous generation games, but for cutting edge high-end AAA games, you absolutely need cutting edge high-end hardware.

Not to mention, TV software is an absolute mess of unstandardised differing software between every manufacturer, so it would be very difficult for game makers to get their software/games to work on every TV.

It might work for low end or previous generation games

You may be right that the cloud tech needed to support the best XBox One and PS4 offerings isn't there yet, but Nintendo has been pushing under-powered hardware for generations now. And I don't know about you, but most of my Steam games would work fine on state of the art hardware from the early 2000s. A pseudo-incumbent like Nintendo and a potential disruptor like Valve should absolutely be thinking this way.

And I don't think TV software should be an issue. Virtualization should be a trivial problem for tech companies of this size.

I'd wager that your steam library is atypical, but I don't have any evidence for that beyond my own and those of people I know.

Also, tech companies that big can't even get basic navigable menus right, so I'm not sure I would agree that virtualization would be easy for them.

> And I don't know about you, but most of my Steam games would work fine on state of the art hardware from the early 2000s.

Most of my games probably would, but the games I care most about right now would definitely not.

> Not to mention, TV software is an absolute mess of unstandardised differing software between every manufacturer

That's where a console platform comes in. Instead of being a random smartTV with second rate software built by a hardware manufacturer, it would become a random smartTV with the "Nintendo living room suite" or whatever.

Maybe I'm alone but I personally prefer my screens to be dumb. Do one thing and one thing well applies here. I like being in charge of the computer side of the equation.

1. By having a single platform to target, developers can focus on the actual gameplay and a better game. Supporting all PC configurations is a nightmare.

2. Piracy - consoles are locked down so fairness ensues and devs get paid when someone buys their game. Piracy is rampant on the PC because of it's openness, with even indie game studios suffering from ruthless pirating.

A 500Mhz ARM chip in your TV and some embedded firmware that can run an outdated Netflix app does not an X-box with multicore processor, GPU and gobs of RAM and storage make.

Not to mention the market is so splintered with Smart Devices right now. How do you develop on one that makes up 5-15% of the market and turn a profit?

The one area you're right though is that you don't necessarily need the console right there at the screen, it can be somewhere in the house, and stream the content over the LAN to your screen. Steam works this way now.

Sony, Microsoft and Valve need to push for their hardware because otherwise there will be no need for their platforms. Why would Bathesda create games for the PlayStation VM or Xbox VM if in the end the VMs are running on iOS or Android or whatever else OS? They'll directly create games for the App Store or Play Store instead. Nintendo makes enough first party games to consider doing what you're suggesting. Just what Sega did in the past. Valve had to work on their own hardware due to the threat of the Windows App Store.

iOS or Android or whatever else OS

You just answered your own question.

Remember the screen paradigm. In the 80s, Nintendo made it possible for developers to make games that worked on Zenith TVs, Magnavox TVs and RCA TVs, but they only had to develop it once, for the NES.

The Playstation VM means the developer only needs to make the game once, and Sony is doing the work to make the VM work on iOS phone, iOS tablet, Android phone, Android tablet, Windows, Windows phone, Visio TVs, Samsung TVs, etc etc etc.

Plus it can't be understated that there still needs to be a controller. Touch screens just don't offer the immersion that controllers do.

Both Apple and Google take 30% cut. There's no way a game developer will want to use a PlayStation or Xbox VM and pay them another 20% (that's what console makers get today I think). They'll rather use an engine that already supports multiple platforms. All app developers are already building apps for Android and iOS without using any VM in most cases anyway.

Microsoft and Sony make billions of $s in revenue from their console platforms. It's not easy to switch to a VM play where you're a "nice to have," not a "must have" and continue making billions of $s.

> Video game consoles (and all set-top boxes) hearken back to an era where the average house only had one TV, and it was dumb.

I'm pretty sure the average TV is still dumb.

> These days, there's a dozen screens in the home, and all of them have computing power many times greater than even the original XBox - more than enough juice to power a huge catalogue of existing games.

The TVs and so forth tend to be based on ARM SOCs that have miserable I/O systems; they're pretty good at decoding compressed video, but they're awful at locally generated frames. And they are not platforms in the sense that they have storage in the tens or hundreds of gigabytes range, or decent UIs, or a stable and common set of APIs that developers can successfully target.

So the TV hardware isn't sufficient, and the TV software definitely isn't. And you're not going to get (say) Fallout 4 or anything even remotely recent running on your TV without putting a decent GPU in there because now you're talking 100 watts (honestly, more than that) with corresponding amounts of cooling, plus boatloads of flash at $0.50/GB. TVs are a low-margin business. Better to keep the specialty hardware separate so you can stay in business.

[Don't get me started on cloud rendering. Capex and latency are . . . I won't even get started]

I definitely hadn't considered the low-margin nature of the TV business.

But surely there's an avenue to starting a platform like this on phones and tablets, with an initial focus on older games appealing to a niche audience.

Once you have a niche, you can expand the service and wait for broadband/video compression to catch up and disrupt the big boys later.

You are looking at this from the perspective of someone living in a developed country with a fast broadband connection, but the global gaming market is much more diverse and complicated than that. Sony has been making attempts to push PlayStation Now on to consumers in North America with very mixed reviews. The idea of streaming games directly from the cloud to your Samsung smart TV sounds really appealing on paper, but broadband speeds just aren't there yet to push this out to a global audience. One must also take into account that most people in developing countries don't have access to a credit card required to pay a monthly subscription fee. That is why the "big game companies" will keep on shipping hardware until the remaining pieces of the puzzle are in place.

I have yet to see a TV that is actually smart. What passes for Smart TVs today are a joke. The menus suck and the remote controls are even more unwieldy than 'dumb' TV remotes.

Until we get a TV that boots to an Android home screen and comes with a tablet that mirrors the TV screen, TVs will remain dumb.

Is this some kind of hipster humble-bragging along the lines of "I play old games, so old, they could even run on the TVs build-in SoC and honestly, HONESTLY, do not understand why would anybody bother with anything newer!"?

The reason big game companies make hardware is exactly the same why the big non-game companies (Apple, Tesla, Boeing etc) make hardware - it fucking sells.

padobson the problem is that a lot of gamers do not want this.

Last year during the E3 press conference, Microsoft tried to "modernize" the whole living room gaming paradigm. They mentioned things like - 1. Having a whole slew of servers to cloud compute 2. Buying and sharing games via online distribution 3. Always on connectivity to the cloud

Gamers hated this idea (as you can see with the sales of the One vs the PS4).

The issue is,and I can't blame them, that we still do not have the infrastructure to do this yet. Yes, me and you may have a good enough internet line to pull this off, but even now our internet access, latency, connectivity, and security is a joke. Remember, a lot of these guys do play online but have issues with latency by playing just a multiplayer game. We need this fixed first before a company goes all in. It may take for the next cycle (or two or three) for dedicated game machines to fully embrace this.

I'm excited for the game Crackdown 3. Their online play will feature fully rendered buildings which stands on their own dedicated server. I think this is awesome and is a step in the right direction. Can you imagine if every NPC can have its own server controlling it? Crazy-power. It will take time though.

And I'm all for it. Take the time to do things right and we'll get a much better product than just a gimmick.

Especially Microsoft, if they've learned anything in the last 5 - 10 years it's the cost of letting OEMs be the interface between your product and your customers.

While not the same as the Steam Machine, I got one of the Steam Link streaming boxes for $50. It's pretty amazing and can't believe how well it works.

I've only played Rocket League, Pay Day 2, Monanco, and Super Meat boy on my television.

For me, it's the perfect compromise for my casual gaming. I get to have a custom built PC, hidden away in my office and a tiny box on the living room.

I look forward to seeing the platform evolve.

Glad to hear it works well as I'm considering it.

Curious: Is either your gaming PC, and/or the Steam Link box wired?

Not the OP but I did have both sides wireless for a while and it couldn't keep the framerate up (playing Portal). I have an i7 920 @3.6Ghz, 680GTX, 12GB RAM on the main machine. I was streaming to my late-2013 MBP 15". I have an AC1750 802.11ac router.

Once I wired my desktop in, I could play wirelessly via my MBP with 60fps constantly.

I suspect if both were wired you'd have excellent performance. However, one caveat is that low-light games really seem to show a lot of artifacts (for example, Amnesia: The Dark Descent was close to unplayable because of compression artifacts, despite being on 'highest quality' stream settings).

I hope this is helpful.

Yes, they are both wired

Received mine yesterday and loaded up Tales of Zesteria. It works pretty much flawlessly for me and I'm super happy with it.

I have no illusions that I'm going to be playing the latest Fallout 4 with it, but I have a strong tendency towards non-twitch based games, and those will work just fine.

When I do play Fallout 4, et al, it will be on the PC, but being able to stream so many of my games to my television is great.

I'm hoping to try Fallout 4 this weekend, if it's raining out.

I'd love to hear your experience on it. I have someone in the house who very much wants to purchase fallout 4 to play it over the link. I'm just not confident it will work, even with a good, wired connection, but I'd love to be wrong :)

A Steam Machine is a PC running Linux. You can do whatever you want. We're not going to do anything to get in your way or lock it down. A lot of people are going to buy it and plug it in and only run Steam. That's cool. We also think, given history, there's a pretty good chance that users are going to find something new that we have not anticipated that's going to be super awesome, that's going to make a bunch of other customers really happy. To the extent that we can add value and make that more likely to happen, that's what we're going to do.

I shed a tear of joy. A company that isn't run by soulless MBAs, about to curb-stomp a bunch of companies that are.

Marketing I think was the key here. The Steam Machine needs to appeal to people who are not previously exposed to PC gaming. People who already play PC games have PCs to play them and hooking up a computer to a television is trivial. Most of them probably know what the Steam Machine is, but if I asked any of my 'XBox' or 'Playstation' friends they wouldn't have the slightest clue what Steam is, let alone the Steam Machine. Why would they? I haven't seen any advertising anywhere. In combination with the general negative stereotypes about PC gaming (not the least of which comes from the name PC Master Race) I don't see any what that this could catch on.

Who are these people who are exclusively PC or console or mobile gamers? Everyone I know, rich and poor, are all three.

Most guys I know play Xbox/PS4 and some mobile games. I know maybe a few people who play computer games on top of that, and plenty of people who just play mobile. PC games are generally seen as 1. extremely expensive since the cost of computers is high especially if you dan't build one yourself and 2. only for hardcore gamers. Most console gamers I know don't want to compete with the PC gamers, its like when I play Call of Duty and get slaughtered every single round but worse. The atmosphere is one in which you win at any expense, even that of the game being fun to play.

These perceptions may be wholly inaccurate but they will remain until demonstrated otherwise.

Computers are not expensive. Many people use their work computers to play low-end games, and not too long ago, flash games were very popular and not just among hardcore gamers.

The price factor has always struck me as a case of not having enough information on PC hardware and game requirements (which is understandable if you only know console gaming).

You don't need to go all-out on top-end graphics cards, processors, cases and watercooling to play PC games at a similar or higher standard (in terms of graphical quality and framerate) to consoles.

A mid-range PC may or may cost a little more than a console but the fantastic savings that come from things like Steam sales mean you can for example, pick up 5-10 games over Christmas for a fraction of the price of doing the same with the XBox/Playstation versions of those games can mean it ends up being a lot cheaper in the long run.

Of course, knowing about the prices of PC hardware and how to put together the best bang-for buck system is something you would have to consider. That knowledge is unnecessary with consoles.

I for one am a PC only gamer. I had an Xbox 360 for free that I barely used and prior to that there weren't enough exclusives I was interested in to make me want a console.

I'm just a gamer. I play everything from Metal Gear Solid V on my PC to Final Fantasy Tactics on literally every platform its ever been released for to weird puzzle games [0] in my browser and on my Android phone.

[0] maybe calling some of these things games is an over aggrandizement. They're probably more "graphical amusements."

I used to be PC + Console but the PC was always my preferred platform. I am now PC only though.

There was very little about the last generation of consoles that appealed to me and basically nothing about the current one.

There are a few console exclusive games I would quite like to play, but honestly my Steam backlog alone is like 2+ years at this point and always growing.

I have more than enough stuff to play as it is.

So yeah I see some complaints about Linux, but there is no way Windows or OS X would have been a better choice for a product like this. Steam Machines are supposed to be the Android of the console world. Valve would have had to work closely with MS to create some locked down platform to keep people from screwing it up. You couldn't just slap stock Windows on there... and imagine the licensing fees. Linux was the best choice.

I do believe that when Vulcan comes out and Valve or some other company throws some money at advertising, steam machines could easily come to dominate the console market, just as Android came to dominate the smart phone market.

>steam machines could easily come to dominate the console market,

What extra value would it provide over existing consoles?

If Valve would join forces with Google merging ChromeOS & SteamOS -> AndroidPC then Microsoft will have to watch out.

http://www.theverge.com/2015/10/29/9639950/google-combining-...

Then you would have a device with world class apps+good gaming experiance combined with a good user interface.

I doubt it would ever happen, but I really wouldn't be surprised if Google attempted to acquire Value. They've been making moves to get in to the "casual" gaming space with Google Play and the Nexus Player. If they purchased Value they would effectively own a large segment of the "hardcore" PC gaming market share.

Again, I doubt it would ever happen, but it's interesting to consider what the impacts would be if they did. I don't think Value is for sale, and if it were Microsoft would be much more likely to purchase them.

They would have to pay more for Valve than any other company they've ever acquired with the exception of Motorola, I think. And I'm still not even sure Valve would say yes, even to a $10b offer (I'd hazard a guess at a fair offer being $5b-10b).

I really doubt Gabe would ever sell. He's just not that type of person, and Google would get zero benefit from Steam, its not their realm at all.

Completely agree about the not selling. Not so sure that Google would get zero benefit, though.

Firstly, Valve is extremely profitable, crazily so, and seems (imo) unlikely to stop being so profitable, so short of a crazily high PE ratio it would eventually pay for itself. Secondly, they'd be buying a userbase of 100m+, including presumably lots of interesting data from steam chat, gaming preferances, etc.

The media has been slightly mixed up about the whole Chrome OS claim.

http://chrome.blogspot.com/2015/11/chrome-os-is-here-to-stay...

Just what I needed, ads in my Steam and Steam games!

Honestly I'd prefer a merge with KDE Plasma Active (designed to handle multiple formfactors well), maybe with Android API:s and VM embedded. Don't know what ChromeOS offers specifically that would be an improvement here.

Honestly, the last thing I would want is Google that heavily involved in PC gaming.

I'm hoping someone with more knowledge about Direct3D/DirectX could clarify some things for me. Since we're seeing a lot of progress on .NET running on OSX and Linux, does this mean that at some point you'll be able to run Windows games on both OSX and Linux? If not, what are the major hurdles there?

I don't think there are a huge number of Steam games running on .NET, but the number might be increasing due to the rise of Unity.

Many, many games are written in C++ and require reasonable porting efforts to get to Linux support. Someone like Ryan C Gordon (icculus) is much better qualified to comment on that, though.

I doubt Microsoft would ever port DirectX to Linux, but then again I doubted they'd do that with .NET, and look where we are.

The DirectX->Wine code works pretty well, but it's always going to be behind the state of the art.

There is already a native Direct3D 9 implementation on Linux, called Gallium Nine. It works with any driver exposing the Gallium interface, such as the open source radeon driver.

Direct3D and DirectX have nothing to do with .Net. The port of the CLR will only allow to run .Net(core) applications on OSX or Linux . And even there DirectX seems to be a blocking factor e.g. XAML/WPF which uses DirectX for drawing won't be ported/opensourced ,IMO for this very reason.

Reply on news.ycombinator.com