Valve Employee Handbook (2012) [pdf]
cdn.cloudflare.steamstatic.com
cdn.cloudflare.steamstatic.com
- https://news.ycombinator.com/item?id=3871463 (April 21, 2012, 168 comments)
- https://news.ycombinator.com/item?id=8818893 (Dec 31, 2014, 17 comments)
- https://news.ycombinator.com/item?id=9250527 (March 23, 2015, 14 comments)
- https://news.ycombinator.com/item?id=12157993 (July 25, 2016, 198 comments)
- https://news.ycombinator.com/item?id=17935030 (Sept 7, 2018, 31 comments)
For instance:
- Ex-Valve employee describes internal politics at 'self-organizing' companies (pcgamer.com) https://news.ycombinator.com/item?id=17584563
Apparently flat structure and bonuses tied to amount of money you bring, combined with apathy towards their software allowed some weasels to takeover and slowly destroy their (few) games in the name of short term gains.
And money is not even an issue for valve, its private company. It all looks like whole leadership doesnt give a fuck anymore, not even to pass on maintenance to 3rd party.
Gabe's "Actually money is how the community steers work." comment during the paid mods fiasco really shows the mindest that Valve works with.
The launch and handling of Artifact is a clear sign of valve's hubris or complete out of touch state. Or more realistically pure greed of people who would benefit of it. By people I dont mean valve as company, they make more money of steam in a day that they could have ever make of artifact in a year.
Launch a paid card game at high of Hearthstone, that requires you to buy cards with competitive decks going over >$100 and micro-transactions to play it. Everyone knew it was crazy. And game died pretty much at the launch.
Why the fuck would you even attempt making it a paid game?? Valve has money to keep a game like that free, develop community and get it to stable state before attempting at monetization of some kind.
Though I would love to see a game come out of there again sometime. :)
Australia — A place that’s either very near or is New Zealand where more than half of Valve’s employees were born.
My point, in response to this specific quote from the person I was replying to, is that good people don't bounce when a project gets hard, they in fact do the opposite.
If you disagree with the statement I quoted from the person I was replying to that's fine, I have no bone to pick.
Then you have their VR and steam deck hardware that is setting the bar for new gaming platforms. The fact that they also found time to ship a game is just icing on top IMO.
Counter - Strike
Steam (and numerous innovations within it)
Dota 2
Portal 1 and 2
The entire VR industry as it exists today comes from Valve showing Oculus their far more advanced prototype and giving them the tech, which later led to FB acquisition
CastAR which was initially developed at valve - so far not successful but I think it has potential.
The best VR game currently available - half life Alyx
Gaming on Linux through Proton
Steam Deck
> Dota
> Portal 1
> Gaming on Linux through Wine
These were all invented outside of Valve.
Does wine even compare to what Valve has accomplished with Proton?
My guess is no and there is still a hierarchy, however obscure.
This is my suspicion as well, and my experience with "flat" organizations. Hierarchy and power dynamics are phenomena that occur in organizations. If you do not create the power structure on purpose, it will arise spontaneously, as your organization grows. If you try to suppress power structures, the power structures will hide themselves.
This isn't some conspiracy theory stuff. It just means that in any organization, some people will have some authority--because they are liked, maybe because they get things done, maybe because they know things, maybe because they've just been around a long time. These people end up getting kept in the loop for decisions, not because anyone is directed to keep them in the loop, but simply because it's the way things are done. If you try to do something outside this power structure, you may encounter organizational resistance--if for no other reason than because you're doing things differently from everyone else. It is more difficult to navigate these kinds of power structures because there's still someone in charge, more or less--but it's not written down.
Sometimes you encounter a truly flat organization, but they tend to be very small. I don't think it scales well.
The gist is that in officially flat organizations, there always exists an unofficial hierarchy that arises organically, and that because it's unofficial it's unaccountable, unlike a formal hierarchy one could theoretically dissent with or challenge.
[1] https://en.m.wikipedia.org/wiki/The_Tyranny_of_Structureless...
Why would I want an org chart to be fun or interesting? Usually, my goal is for anything "fun" and "interesting" to be either a direct responsibility at work, or something I do outside of work for pleasure.
Don't work somewhere interesting. Walk the maze of the traditional establishment, and enjoy your quiet workday. There are so many opportunities for you to do that.
Myself, I don't want to be driveling stupidly many hours a day over whatever apparati I might be charged with the operation of. I want stimulation and novelty. I want to be curious and satisfy it. I want to be confused and amazed. The many, many hours I'll spend at work, I want them to be edifying. Not some sacrifice to a faceless quorum of buzzards that I might live another day.
This whole wage labor regime is a fairly new thing, and it hasn't been rightly tested in all its degrees of freedom. Something Hobbesian about the direction and shape it has taken, it's sad and cynical and skeptical but without faith.
What you are describing sounds like the subjective experience of what you feel when you work in the traditional establishment. You feel unstimulated, but that's a product of your personality and your immediate environment.
To put it another way... whether an environment is "interesting" depends on whether someone is interested.
Choosing to leave a large organization gives you some freedoms and takes away others. It sounds like you value the freedoms that you get outside large organizations, and you have only harsh words for people who feel otherwise.
Just to put this in concrete terms... let's say you are curious about compilers, or memory allocators, or network flow control algorithms. It tends to be only large organizations that will pay you to work on those things, and in order to my curiosity about those things, I would want to rub elbows with experts in those subjects. Those opportunities exist at the largest tech companies and in academia (and my experience is that academia is no less bureaucratic).
For me, sometimes, navigating broken bureaucracies is a price that I pay to work on some really fascinating problems. The broken bureaucracy itself is a fascinating problem.
I've worked in several "flat" companies that tried various democratic things like this.
It always turns into the ultimate game of politics. In this case, it would instantly become "I'll vote to raise your salary if you vote to raise my salary".
https://www.pcgamer.com/ex-valve-employee-describes-ruthless...
At every big tech company, there’s always a few employees who assert nothing has changed and their employer is the same plucky upstart they joined 5 or 10 years ago.
Who sets the salaries? With a flat structure and no "managers", who picks your salary for next year? Who picks how much a new person gets an offer for? How are raises and equity comp determined?
That all seems to be missing from (at least this version of) the handbook. The word salary appears once and compensation eight times, but it just says "we try to be the best" but never discusses the mechanics of it.
Thanks!
the interview process is also kind of interesting -- you'll know if you got an offer or not before you leave the onsite interview. also interesting was that they sent a driver to the airport instead of offering the normal taxi/uber reimbursement, then had the same driver pick me up after the interview to head back to the office.
That makes sense, having the people who interview attempt to pick a comp that will get you in the door.
> then is adjusted up/down as your peers estimate your impact over time.
It is the mechanics of this that I am most interested in and no one will answer for me. Do people vote on their peer's salary? How does that not become super political? What happens if they all vote for each other to get 1000% raises, who stops them?
> also interesting was that they sent a driver to the airport instead of offering the normal taxi/uber reimbursement, then had the same driver pick me up after the interview to head back to the office.
If I were them the driver would be the beginning and end of the interview process. Were you kind to the driver on the way there? Did you say nasty things on the way back?
Or maybe they just got a good deal with a company that only had a few drivers. :)
That said, true co-ops of all sizes also struggle with compensation. Much like startups, you want to reward early risk acceptance and loyalty while preserving fairness based on skill and contribution (otherwise you won't be able to recruit new members). It's really, really hard.
(I know who you are but I figured it was worth commenting since others may not know as much on the subject).
(This is also why I highly suspect Valve is not truly horizontal: you will never get 100 people to agree on anything important.)
I'm still more interested in WHAT I'm doing than what the compensation is. The patreon arc of Airwindows is sort of peaking out and subsiding very gradually and I'm looking at what else I can do: if I'm compensated much less than minumum wage it begins to affect what projects I can take on as it'll erode my resources too much. Such is the lot of an open source dev who's used inherited money as capital :)
I'd work at Valve. Sounds fascinating, and a bit like the production work I've done: working with a band and getting the most out of their album, except it's an ad hoc group of devs and the 'album' is a game or software infrastructure. There will be people in such a situation who are maximizing their own wealth. Such people are only minor obstacles to getting the THING you're making, optimized. Basically if you're getting enough money anyway and they will let you have the thing needed on the grounds that they take too much credit, but the thing can be made, then let 'em have the credit as long as they're not screwing up anything important.
It's sufficiently obvious who is contributing and for what reasons. There will always be outside perspectives, including outside of me and whoever is credit-seeking for wealth maximization.
> How pay is determined
>"This is a haphazard process. The payment mechanism is to a very large extent bonus-based. So the contracts usually have a minimum pay segment in it, which is more or less established by tradition. And then the interesting part in this contract is how much is left to the peer review process, which is very complicated. It involves various layers of mutual assessment. "In companies like Microsoft or elsewhere, usually the bonus is something between 8, 15, 20 percent of the basic salary. In Valve, I'm told, there's no upper limit to bonuses. Bonuses can end up being 5, 6, 10 times the level of the basic wage." "Gabe [Newell] had this view from the beginning. He wanted a community of partners, he didn't want to be the boss of anyone or to be bossed around by anyone."
https://www.gamedeveloper.com/business/how-valve-hires-how-i...
This part about why these teams are temporary and they prefer a flat structure makes a lot of sense and resonates with me (and probably many here who've worked in larger organisations):
> Valve is not averse to all organizational structure—it crops up in many forms all the time, temporarily. But problems show up when hierarchy or codified divisions of labor either haven’t been created by the group’s members or when those structures persist for long periods of time. We believe those structures inevitably begin to serve their own needs rather than those of Valve’s customers. The hierarchy will begin to reinforce its own structure by hiring people who fit its shape, adding people to fill subordinate support roles. Its members are also incented to engage in rent-seeking behaviors that take advantage of the power structure rather than focusing on simply delivering value to customers.
Humans are good at organizing, if the goals are clear and everybody has the motivation.
And rhis is the actual problem. If I feel like the company is fucking me over in some way or viewing me as a cow they have to milk as much as possible, that motivation will be gone.
Say, for example, I am working on a task and I need something done which is not normal. For example, I may need to make a change to some service that is running, or get permissions or ownership of some object assigned to my account. Because of the legibility that bureaucracy brings, I can find the relevant team with ability/authority to solve my problem, or I can discover that no such team exists.
If the team exists, then that's great. I can talk directly to people with experience with whatever problem I'm solving, who know the relevant policies / risks / recommendations. If the team doesn't exist, I can still just make the change myself. If the team exists but won't help me, I can hand the problem off to my manager. If the relevant people are on vacation, I can find out who has power in their stead. (There are some cases where this doesn't work--my experience is that this usually involves some inflexible legal requirement, or some essential technical expertise, and the relevant person is on vacation.)
This legibility brings a sort of "mindlessness" to the bureaucracy. I have high confidence that soon, I will either solve the problem, or the problem will belong to somebody else. I don't have to think too hard about what the right thing to do is when it's outside my domain, except in the exceptional cases.
Yet it's interesting that despite Valve having what is essentially a monopoly on PC storefronts and the implicit corporate hierarchy with alleged double-crossing, there is little complaint about Steam as a service from most users.
In hierarchical organisations you have to deal with the whims of the sociopath above you as well.
In many ways these power structures are a feature of any large people operation. You just need to get used to them.
Things will be chaotic, but modern software development does some good ideas to the table to maintain order. Things like Agile, OKRs etc might not be perfect but they do help.
And in contrast to, "centrally consolidated power attempts to shift and contort people and cliques to meet external, largely disinterested expectations and save face" -- the appeal is pretty obvious.
Holy shit. I have to know what other groupware voodoo they're running internally, now...
gaming companies don't generally do their primary development on wifi. not the rule, some people will, but most people aren't making halflife alyx on a macbook.
The one I'm thinking of is called Cisco Prime and you can see it here https://youtu.be/xrtGXg5ICa4?t=2235 (when I saw it before, they had 802.11x auth info integrated so it could pull Active Directory/LDAP info, too)
For some reason, this manual gave me the same feeling I had when looking over college brochures as a young man. There was a sense of potential and excitement that was hard to put into words. Of course, the reality turned out very differently but I still remember that passing mixture of floating forms. Whether or not the document accurately depicts what it's like working at Valve, I am still taken aback that it managed to reproduce so specific and personal an emotion.
It says a lot about how large valves market share is that they can take such a large cut and publishers will still sell there. Plenty have tried to release alternative clients without mixed success
Why then install an alternative client?
Source: I tried to compete with steam once. Tech was better. Nobody cared.
Remote Play is fun on games that support it.
And it runs on osx, linux and windows.
And then theres Broadcasting, chat. Its a heavily featured launcher, and offers way more than some of the alternatives out there.
I had all the same gripes you do, believe me. I face similar issues with Steam. Nonetheless, people don't care. It works well enough.
There are all kinds of factors at play here, none of which I have any concrete data on.
First up, there are probably way more console users than pc users. However, a percentage (not a majority, but certainly significant) of those would be kids whose parents buy them one or two games a year.
OTOH, PC gamers tend to have more disposable income (based on the fact that they have at least a pretty decent PC). Plus, steam sales come up all the time and lots of PC gamers have massive libraries of games they bought cheap and never played.
All of these (sometimes conflicting) factors make it hard to tell who spends more.
Especially once we take into account that the "1060" is actually a merging of multiple 1060 SKUs (including mobile) with pretty different performance
I'm not talking about the people buying i7s and 4090s. 100% agree they are a tiny minority. I'm talking about the same modest setups you are.
Even a 1060 (or it's modern equivalent) with an i3 and a 24 inch 1080p panel is more expensive than a PS5/Xbox. And I think it's becoming increasing rare for a family to have a desktop pc unless it's for gaming.
> OTOH, PC gamers tend to have more disposable income (based on the fact that they have at least a pretty decent PC).
If we're talking about on average it's probably the exact reverse of this. I'd be surprised if more than 25% of Steam users were in North America. Not because Steam isn't huge in NA, but because getting a console and games is way harder in most other countries than running what you can on the family PC.
For example (at least before the sanctions) Russia makes up the largest portion of CS:GO and DOTA2 players. At least for DOTA2 it's more then double the number of American players.
i bet without cs and dota steams revenue per user has been in decline over the past 5 years as money moves from paid libraries to paid addons in free games
From the link I posted above.
Of Valve's 8~ billion revenue they gross only 15-30% of that while Epic has far higher margins on their ~6 billion revenue.
So four games like Fortnite make more. There are a lot of games on Steam.
Because investors/VC can't buy in on it, so no trying to inflate the valuation.
Given the size of the PC gaming market and the ubiquity of Steam you would think they have >>$1B in revenue though, which makes $7.7B a pretty conservative multiple.
This source says PC gaming is ~$30B, so $1-5B in revenue seems reasonable for Steam: https://www.globenewswire.com/en/news-release/2022/05/23/244...
I'm not much of a video game player, but Valve seems to have millions of loyal, paying customers. Given that and their history of direct financial success (i.e., not just popularity with developers), $7.7B really doesn't seem so bad.
That's pretty much the point. It doesn't matter whether it's a GOOD plan. In a sense it matters much more that it's an evil plan, because that would hurt everybody else more and lead to more capital accumulation and return on investment.
Sometimes these things aren't that complicated. Whether the company's beneficial in other ways is a much more difficult question. Valuation isn't: valuation is more or less a gamble on how much the company can crush everybody else.
Activision was just bought for nearly 10x as much, IIRC.
I remember reading this years ago and thinking it was just so cool (the contents and structure). But I’m not sure how well it really works in practice. Is it better, or just different?
I think the ideals are noble and clever, but the results may not bear out as well as they did at the peak of success.
This is essentially no hierarchy, and everyone interacts with a few levels below them, and everyone is expected to deliver on product (basically, program).
This works because it is actually really hard to get anything done if you're managing more than 1 person.
What happens if you hire someone and then they say “nah working on steam is boring, I’m gonna go work on something completely different”
How do you financially allocate head count if people can just leave the project whenever?
I wonder where Valve stands on WFH.
I wonder if that policy has changed much in the past decade.
Valve, with all its freedom and liberty for employees, lagged behind crunch mills this treated employees like cattle. The market said CandyCrush > Half-Life. So be it. God is dead, he remains dead, and we have killed him.
I require my engineers in the office and doing what I want them to do. Liberalism must be curtailed by the winds, but my god, it’s no fun.
Thanks for sharing, I guess this won't last long, really interesting to see how they operate though!
Timer set for y-minus 2yrs.