The only issue I have with these announced beta prototype specs is the Core i3 option, because it is dual-core. Should've gone with a cheap i5 instead. I think all final retail Steam Machines should be at least quad-core, excluding the small low-end boxes that are mainly meant for streaming and light indie games.
Valve makes money by selling people games on Steam.
Steam is getting more and more popular, but PCs are getting squeezed more and more by other markets and by strategic incoherence on the part of Windows.
Valve's goal is to make sure Steam is able to keep getting more popular even as PCs (in a vacuum) decline.
One way to do this is to make Steam compatible with other computer OSes that already exist; that's why Valve created Steamplay and released Mac/Linux clients.
Another way to do this is to create their own gaming-oriented OS and encourage more and more ports to it; that's why Valve created SteamOS and is pushing for Linux ports from major developers.
Another way to do this is to make it easier to play games on the couch/TV so people don't feel compelled to sit at their desk when they play PC games; that's why Valve created the Steam Controller and Big Picture mode.
Yet another way to do this is to make it easier to buy PC gaming hardware at a reasonable price without having to know much about computers; that's why Valve created Steam Machines and is working with manufacturers to make them efficient and effective.
The common denominator of all these efforts is that they answer the questions of people who don't already use Steam. If someone says "well I'd get into PC gaming but it's too expensive/obnoxious to build a box," you tell them to buy a Steam Machine. If someone says "lol comfy couch," you hand them a Steam Controller. If someone wants to play indie games on a miniature HTPC streaming box, you point them to SteamOS. And so on.
Great post, very concise. This should have been obvious to anyone who actually read the Steam Universe announcements and paid attention to some of Gabe's past interviews over the past year or so about the future of Steam and the desktop PC market. Steam is where Valve makes much of their money, and Valve wants to diversify and give more entry points to people who are interested in Steam and PC gaming in general. They are attempting to future-proof their main revenue stream.
Just because one aspect of their overall vision isn't for <specific person> doesn't mean that it's pointless or a waste of time. Sure, you can build your own box, and probably at lower cost than a pre-built Steam Machine at similar specs, and throw Windows 8 on it...and Valve would be just fine with that. But the fact is that some people out there who are interested in mid/high-end PC gaming simply don't want to build their own, period. Even if they understand that it's likely cheaper to do so. They simply perceive the act of learning and researching parts, buying them and putting the thing together as time-consuming or tedious or busywork. They want to get to the actual "entertainment" (playing the game at good performance) without jumping through a number of hoops to get there. A mid-to-high spec Steam Machine w/ SteamOS meets their needs IF sold at a reasonable price. Even better if a number of well-known devs get on board and support the OS.
That said, the high amount of confusion...not only here on GAF and other forums, but also among a lot of people in the press, is a little concerning. It reminds me a little bit of the Wii U's initial reveal. Valve may want to get out there in the near future and give some good interviews or something to really reiterate and clarify why they are doing all of this, what the point of it is, whom they are targeting, and so on. And I expect them to do that.