There were fun, bug free games before steam, so it's not a prerequisite. Do you think a steam-Civ5 game will be funner/bug freer than a non-steam civ5 game (factoring in the $ that goes to Valve being instead spent on development/etc.)?
Working on the assumption that they have the same manpower, budget, and time in both scenarios? Yes. There is no replacement for Steamworks that you can simply drop in, they'd have to use something inferior (GFWL or Gamespy) or roll their own solution. Creating their own solution is time spent building, testing, debugging, testing, implementing, testing, rolling out servers... and testing. Locklear93's post on the last page talked about that, from the perspective of QA.
Worrying about money going to Valve here is like saying that FPS games would be better if they built their own graphics engine instead of licensing a solution. Does anybody really think that money spent on the Unreal/Crytek/whatever engine isn't worth it compared to every FPS developer building their own? How about Havok for physics (for an Elemental example)? Havok's not free, but Brad licensed it anyway. Developers do that for a reason: it's cheaper to buy an already built and largely proven solution over building your own, especially when you don't have the knowledge to do so in house already.
What Steamworks provides is no different, and it's a business decision. Valve can provide that code more cheaply then each developer can implement it themselves due to scale. Developers can focus on their game and not on the little details of how to implement a friends list and provide server infrastructure for it. (Blizzard can do that, and is rolling out their Real ID solution today for WoW/Starcraft 2 in fact, but Blizzard has their own little empire and has more money then some European countries.)
Of course it has a budget, and gains in one area might be available for other areas (like CEO bounses?), and no one is denigrating civ5 production values.
Your response quotes but doesn't address my question. Perhaps I was unclear so I'll try again... First the points I'm considering, then putting it together:
Pretty sure I did answer it, but alright. The budget doesn't change because of Steamworks. If the cost of Steamworks is less then the cost of implementing the same feature set and supporting infrastructure, Firaxis comes out ahead by using it. Since the budget doesn't change, less money spent on those features/infrastructure means more money spent on gameplay and artwork. Since the budget is fixed, the only way a CEO bonus goes up from it is if the game sells more due to being more polished due to having more developer time spent on gameplay (and less on achievements). If the CEO gets more money because they put out a better game? I'm cool with that, the better game part is what I care about as a customer.
-Valve, 2k, and Take2 are taking a cut from civ5.
If it's sold on Steam at all, Valve would already take a cut from those sales. Take Two owns Firaxis, so strictly speaking they earn whatever Firaxis does anyway. 2k gets a cut as the publisher...
-Today the net has changed (or can change) the old model for distribution/etc. -- the stuff publishers typically do for studios like Firaxis.
According to Brad, the overwhelming majority of game sales are still retail. I believe he's in a position to know the numbers, as the owner of a company in the business. 
That being true, publishers still play a role in the logistics of actually getting a game boxed and out to retail, as well as marketing.
-Firaxis is a proven success, and doesn't have to shill for $ like an unknown, unproven startup.
So, could Firaxis take the $ that's going to Valve, 2k, and Take2, and self-publish/etc., and implement themselves all the useful things steam does (without the stuff like sharing personally identifiable info with unspecified third parties), and still make as good a profit and be as successful? The bonus is that they'd be masters of their fate -- no outside party dictating how some things should be done.
They *could* do it, but it means hiring networking guys, buying servers, server admins, and so on. They'd then need to contract someone to handle the logistics of shipping retail copies, or hire staff to do it. Same thing with marketing. Of course they won't need those marketing/shipping people until they're ready to ship an expansion, so those would be temporary contract workers or put out to third party companies to do... which is pretty much what a publisher does.
On the networking side, they need to pay for those servers as long as people are playing Civ 5, even if no particular revenue is coming in from it. Firaxis isn't a big company and they don't exactly put out a ton of games. I'm not seeing any particular gain for them in paying money to do things they have no internal skill in doing, when they can contract out to companies that are already providing those services and can do it with the benefit of economies of scale.
So no, I don't see how they'd do better by going on their own. In terms of having control themselves... they seem to have control over Civ 5 for the most part anyway. Gameplay wise there are a lot of radical changes happening, and corporate suits tend to dislike radical changes in proven franchises when sequels that don't change things much sell reliably. Does anybody really think someone from 2k came in and said "hey, completely redo the entire combat model"?
The only decision likely made by 2k here was a release date and Steamworks... and really, there isn't a better option then Steamworks right now if you want the functionality it provides.
