Plus, she didn't keep a job more than a year and half through her entire career (thank you linkedin).

Some people need things to move and aren't fit for corporate life. The fact she's very skilled and possibly a genius only makes this point stronger.

It takes the right people, and a self-made self-taught hardware developer hardly accounts for mainstream people...

I can't imagine a large percentage of people do actual productive work at Valve. Sure, I get how if you get enough people from within the company that think your idea is cool enough, most likely it will be fueled, released, and probably succeed in the marketplace. I would think the problem would be that *everyone* probably thinks their idea is infinitely cooler than everyone else's, so you've got lots of half-finished products and hundreds of valuable man-hours wasted because you've got no real authority (other than gravitas) to say "let's all focus on this". R&D teams are needed to keep things fresh and they shouldn't consist of the same people all the time, but the company manual makes me think the culture there is "R&D all the time" because they say their personal project time is 100%. Any company will tell you that they need some workers to focus on the core products people buy in order to fuel future development. Working to refine a product that's already "complete" isn't the most fun thing ever, but it's needed to allow the company to take risks without collapsing. I'm sure the older, more experienced Valve employees see that, but maybe their fringe employees don't. Makes me wonder what will happen to the company when the "old guard" start to exit.

I can't imagine a large percentage of people do actual productive work at Valve. Sure, I get how if you get enough people from within the company that think your idea is cool enough, most likely it will be fueled, released, and probably succeed in the marketplace. I would think the problem would be that *everyone* probably thinks their idea is infinitely cooler than everyone else's, so you've got lots of half-finished products and hundreds of valuable man-hours wasted because you've got no real authority (other than gravitas) to say "let's all focus on this".

Assuming respect and non-corporate bullshit, they might actually have their own sub-system to trade amongst peers: "If you dedicate some time to my project, I'll lend a hand with yours" similar to what civilized people tend to do out of the industry (and even sometimes "in" the industry).

R&D teams are needed to keep things fresh and they shouldn't consist of the same people all the time, but the company manual makes me think the culture there is "R&D all the time" because they say their personal project time is 100%.

I don't know that. Steam was a pretty cool R&D piece. It is true they don't release often, but there's something to be said of companies that make a lot of money and can afford to trash multi-million ideas (WOW fueled Blizzard so much they could actually trash Titan, and kill SC2 2 times before releasing something). What's to say we're not better, as customers, for not witnessing half-assed ideas? I like there's some form of filtering that can happen when a company makes piles of cash because, I for one, hate shovelware.

Of course, this could mean a bunch of good ideas don't get through because the majority of valve's employees don't believe in this, but I don't think this should be perceived as some form of oligarchic censorship.

Any company will tell you that they need some workers to focus on the core products people buy in order to fuel future development. Working to refine a product that's already "complete" isn't the most fun thing ever, but it's needed to allow the company to take risks without collapsing. I'm sure the older, more experienced Valve employees see that, but maybe their fringe employees don't. Makes me wonder what will happen to the company when the "old guard" start to exit.

So-called maintenance teams? I think their products probably are maintained so long as people consider they have value. Steam must be supported to a degree, given the amount of success they get from that...

Assuming respect and non-corporate bullshit, they might actually have their own sub-system to trade amongst peers: "If you dedicate some time to my project, I'll lend a hand with yours" similar to what civilized people tend to do out of the industry (and even sometimes "in" the industry).

If you're dealing with people that have respect, I agree. However, most of the brilliant and talented coders I've worked are astoundingly arrogant. Now that may be just my experience (and it can be said arrogance is arguably a necessary trait for brilliance in coding), but I can't imagine trying to get a couple of them to work on your project without them intimating that the trade isn't equal in their view.

I don't know that. Steam was a pretty cool R&D piece. It is true they don't release often, but there's something to be said of companies that make a lot of money and can afford to trash multi-million ideas (WOW fueled Blizzard so much they could actually trash Titan, and kill SC2 2 times before releasing something). What's to say we're not better, as customers, for not witnessing half-assed ideas? I like there's some form of filtering that can happen when a company makes piles of cash because, I for one, hate shovelware.
Of course, this could mean a bunch of good ideas don't get through because the majority of valve's employees don't believe in this, but I don't think this should be perceived as some form of oligarchic censorship.

There's no question that Steam is a wonderful example of R&D work. However, they'll need maintenance teams to really get all the bang out of it. Maintenance, IMHO, is not glamorous. Most coders I know (including myself) wouldn't do it unless they "have" to. I'd much rather choose to start a new, fresh project rather than maintain the old project's code. Valve purposefully allows their employees to choose what they do, so why wouldn't they decide to do something else? I'm sure like all coders they've come up with a few ideas during development they've been wanting to work on. I think a few people that are really interested in Steam might choose to maintain the code and some might do it to make their peer review ratings higher or possibly even garner favor with the "decision-making" employees, but I don't see it being done in an altruistic way like Valve would have you believe based on their employee manual.

Don't get me wrong. I think Valve's culture is admirable. They're trying to take the open-source coding model and build a radical business management concept from it. Props to them. However, I think it gets more complicated and eventually unstable when you introduce money into the equation.

I like refactoring.

Perhaps you can challenge yourself into refactoring to clean-up the mess?

I think you should read this: http://www.gamespot.com/articles/ex-valve-employee-blasts-the-company-for-feeling-like-high-school/1100-6411126/

Specially the part about the "hidden layer of management".

EDIT: I'm not saying that Valve's flat hierarchy is a lie or something like that, but I'm just referencing a different point of view about the whole thing, which is a nice reference.

Yeah, from time to time we get hints from inside Valve that it's actually not 100% flat.

One example of this that stuck out in my mind was when Portal 2 was released, one of the lead developers was giving a talk/playthrough of the first few minutes of Portal 2, on day of its release, before he rushes off to vacation. He started playing, talking about various design decisions, and came to a point where one of the level designers and artists had designed a certain room, and he got visibly annoyed, saying something like, "I told him not to add that! I'll have to talk to him later about that.", or something to that effect. So much for flat structure!

Can't find the link to the video, unfortunately.

Yeah, from time to time we get hints from inside Valve that it's actually not 100% flat.

One example of this that stuck out in my mind was when Portal 2 was released...

Of course it isn't. There are still people with various amounts of power. People with power over budgets, power over schedules, power over information, power over talent, and so on.

The key is that one commonly misused chunk of power, organizational power, has been removed. Unfortunately it is also one of the most USEFUL forms of power for communication and authority. Without it you need to build your own little web of power, carve out your own little domain, and do more work than would otherwise be necessary. Yes, it avoids one of the problems where sometimes people in structural power will horde the power or refuse to relinquish it frequently, which is when the power is most beneficial. A good hierarchical system works best when used to fling information around rapidly, when used to distribute power, when used to enable people to do work.

I don't think I would like working without that. It would require either getting finding a guide to all the different power hubs or struggling for a while and discovering them on your own. SOMEBODY controls the builds. SOMEBODY controls when things get launched. SOMEBODY controls what features are included. SOMEBODY controls the design. SOMEBODY distributes tasks. SOMEBODY coordinates with QA and localization and ESRB and other groups. If you don't happen to know you are left groping in the dark.

It likely worked well in the beginning, but relying on that kind of hidden structure does not scale well, nor does it tend to last.