Abrash on Valve: How I Got Here, What It's Like, and What I'm Doing

http://blogs.valvesoftware.com/abrash/

by psykotic • 14 years ago

288 86 14 years ago

blogs.valvesoftware.com

dmishe• 14 years ago

Interesting that subdomain is "blogs" and still it redirects to /abrash/

bvdbijl• 14 years ago

This one also exists: http://blogs.valvesoftware.com/design/ but it is empty

judgardner• 14 years ago

Learned a lot about optimization from the black book, might be worth a scan if you like asm.

agumonkey• 14 years ago

This book almost reads itself. Will never forget it.

sp332• 14 years ago

I wonder if this is the secret hardware that everyone's been speculating about recently?

aaronbrethorst• 14 years ago

From the article: "To be clear, this is R&D – it doesn’t in any way involve a product at this point, and won’t for a long while, if ever – so please, no rumors about Steam glasses being announced at E3."

ChuckMcM3• 14 years ago

Interesting read, a long way to go for a recruiting pitch but I digress.

The description Michael gives of Valve is exactly the description of Google's work environment / culture in the early days. I found that that part fascinating. Google's ended up blowing up, but its not clear the mechanism. Old timers espoused the 'too many bozos' theory, I think it was deeper than that.

As a trend / theme / disruption wearable compute has been a staple of talk and research at the MIT Media lab, NASA, and elsewhere. The question I have is whether it is any more than the display equivalent of a 'hands free' kit for your phone. (you wear the speaker & microphone on your ear, you wear the display on your glasses or forehead?).

I've tried in the past unsuccessfully to create a 'fixed' heads up display, which is to say a display which places pixels in the environment space rather than display space. This requires understanding precisely where your head is pointing, and creating a view frustum which is aligned with your physical orientation. The challenge is that it has to respond quickly enough that your brain doesn't see the pixels move from their 'relative' position in the environment. That degree of frequency response has been hard to come by, but there are lots of new inertial units which might address that these days. I'm guessing that you need at least 250 updates / second for 6 degrees (X, Y, Z orientation, X, Y, Z angular velocity) but its an interesting problem none the less. And fun to work on since it combines old school 3D graphics with today's internet of things.

ericd3• 14 years ago

It may be similar to a hands free phone kit in the abstract, but computers are much more broadly useful than phones, so I think that something which makes it feasible to use them into other situations in life will have much, much greater effects on the state of things. I'm almost certain that it will be a huge, life changing innovation once someone gets it right.

ChuckMcM3• 14 years ago

I am right with you there, I see two ways it could go, one is that the computer adds information into your visual field about things that are not necessarily around you (calendar appts, tweets, etc) vs adding info and annotating things you are near.

A hands free lets you talk to people who are not close to you but it does nothing for the people near you (there was a submission about smart hearing aids that would be different) and video assistance has the same potential about bringing personally relevant data to you that can even be situationally relevant but only as an artifact (your calendar said you were meeting 'Bob' so you have information about Bob in your eyeset not because something 'knew' you were looking at Bob and thus brought up information about him.)

Personally I'll be thrilled if I can read email on a decent sized 'screen' while sitting what passes for 'nominal seating space' in a modern jet aircraft.

shuzchen2• 14 years ago

I'd garner to say the difference between Valve and Google would be their size. Valve (according to wiki) has ~260 employees. I'm sure Google has more janitors than that. As I'm assuming very few people leave a position at Valve, that must mean they're very very very*9000 particular about hiring, much more so than google is.

raphman2• 14 years ago

The parent compared Valve to Google's work environment / culture in the early days - not as of today.

10098• 14 years ago

It's a great thing he started this blog. I hope he posts regularly, I have always wanted to take a peek at how things work inside Valve.

jdrols• 14 years ago

I agree. Everything I've read about them just seems so weird and fascinating, especially since it seems like they could introduce something completely unexpected with the ad-hoc way the teams are organized. Valve doing R&D on wearable computing just because someone felt like doing it is mind-blowing to me.

justjimmy• 14 years ago

Awesome read! Valve have really caught my attention ever since I learned that they had wheels on desks, allowing their employees to freely move about and join up with others as they move from project to project. I can definitely relate to that part – I love working closely with our Android and iOS developers, but our company has them in like separate pods so I have to keep getting up and walking over to them whenever we discuss our projects. Tried asking if it's cool for me to move next to the developer I work with till conclusion of project but I guess the management likes to keep the teams 'neat and tidy'.

ZiadHilal• 14 years ago

heh, read all about this dude back when I read Masters of Doom.

roc2• 14 years ago

As perhaps an aside, I think the first-mover advantage is quite debatable.

I'm pretty sure the benefits of Unix have not primarily accrued to the firm that generated the original creative act. Neither did the benefits primarily accrue to the firm that generated the original creation of the Windowed GUI. Nor the web browser, web server, search engine, blog platform, flash-based personal media player, social network, smart phone, tablet, etc.

I believe a strong argument can even be made that followers-on have an advantage in seeing the landscape laid out before them and planning for the future, while the first-mover is mired in legacy concerns.

Granted, much of this rests upon the definition of a discrete invention and thus the original creation of it. But it seems things are quite a bit muddier than Abrash claims.

Edit: Indeed, his citing of any Zynga games as examples of first-mover advantage seriously undercuts his argument.

InclinedPlane• 14 years ago

Look at Apple as an example. In their early phase they were often a first mover. Today they've moved into a different phase where they tend to be a strongly innovative later mover. The iPod wasn't the first mp3 player, but it also wasn't just a clone or even a slightly evolved clone of earlier designs. The same applies to the iPhone, the iPad, Macbooks, iTunes, etc.

courage• 14 years ago

I don't think Id got most of the money, or any sustainable advantage from making the first networked first person shooters, but that's not the point. First mover gets to change the world. That was true for Id, Xerox, AT&T, etc.

roc2• 14 years ago

I don't dispute that. But the way Abrash phrased things, it certainly seemed to be an economic argument.

crystalis• 14 years ago

It seems to fit better with "winner take all" instead of "first-mover advantage"

The_Sponge• 14 years ago

I like the concept of people being allowed to work freely on things like wearable computers. The downside is that it might go nowhere, the upside is that you might end up with everyone in the world using steam and paying for groceries through steam wallet.

corford2• 14 years ago

A little off topic (and the last game I played obsessively was quake so please excuse me if what I'm about to ask is stupid, I'm a little out of touch!), but... why don't Valve have much of a footprint in the mobile game space? Considering what they have done with steam in terms of a distribution platform, it strikes me as odd that they haven't gotten in to the mobile game world e.g. something like a hybrid app store/steam platform but exclusively for mobile game content. Or have I missed something and they are actually engaging with the mobile/smartphone platform?

Splines2• 14 years ago

They could do something for Android, but iOS and WP7 would never allow a 3rd party "store" application.

They do have a Steam iOS application, but it's an adjunct to the PC Steam application.

I wouldn't mind for them to create cool stuff for iOS (since I own one), but Valve's games don't lend themselves well to portable versions. I haven't seen a portable first-person game done well on any platform (even NDS and PSP). They'd be better off creating new IP or a different genre for mobile (like ID's mobile games - the on-rails shooters aren't bad, and the RPG games are great fun).

corford2• 14 years ago

Interesting that you say no one has managed to crack first person on a portable platform. Maybe Michael Abrash can bring some of that game changing "Quake" style magic to mobile (although I guess Carmack is a pretty big missing ingredient).

lloeki• 14 years ago

> I haven't seen a portable first-person game done well on any platform

Not 1st person but close enough: try Dead Space on iOS. Mass Effect Infiltrator is close behind (and recently got a nice update regarding free aiming). Both nailed touch input for this class of games, avoiding the "shoehorning an on-screen gamepad emulation" kludge entirely (like e.g Max Payne and GTA games, which I did not buy, because the show-me-a-gamepad-on-screen is a dead giveaway of a milking-the-cow release).

As for PSP, it lacked a second stick. Their decision of a single stick design was braindead, especially given the PS/PS2 experience Sony had.

Roritharr2• 14 years ago

This is it. Never has anyone more perfectly described the way in which I want to work. I love to work in teams on projects where I feel I can make an impact to the extent of my capabilities and starting projects myself is deeply entrenched in my nature.

I've been thinking about wearable computing for the past 5 years and even went on to study Computer Vision because of it, but after 3 semesters i switched the field to general computer science, mostly because i didn't exactly wanted to reimplement AR-Engines and learn about all the algorithms behind object recognition but just use them for the things I had imagined. I've been working in the past year with Metaio's AR Framework called Junaio and Qualcomms AR SDK for Unity 3D and will write about some of my ideas and findings in the next weeks. (You can create impromptu Project Glass glasses by taping a HD Screen Phone like the Galaxy Nexus in 10-15cm distance from your eyes onto some fake glasses and add counter weights on the opposite side, it's enough for testing all the ideas you have in that field until the real Glass comes along)

Right now i feel very bad about my decision to switch fields and focus on working more while studying to get more experience, simply because I think that if i would've finished my computer vision studies i would feel more adequate to send them a job application now...

wallflower• 14 years ago

> I think that if i would've finished my computer vision studies i would feel more adequate to send them a job application now...

Well... I'd like to suggest that you do not pre-reject yourself. Let them decide. At the least, it will allow you to get on their radar. Remember they seem to have a long term view.

read_wharf• 14 years ago

"Well... I'd like to suggest that you do not pre-reject yourself. Let them decide."

Everyone post that on your wall above your monitor.

tomfakes• 14 years ago

I would guess that 99.999% of programmers have no understanding (right now) of anything to do with computer vision. Combined with your general purpose CS studies, this puts you way ahead of the game in this space. As far as I can tell, Mr Abrash doesn't have any background in this area at all - you're already more qualified!

Scott Adams blogged about having expertise in 2 different areas allows you to fill niches that no-one else can. His were Corporate Politics and Drawing Stuff - and then came Dilbert. And, frankly, his drawing ain't great...

You can't change the future if you wait 20 years until you have the experience you think you need. The future will be set in the next 3-10 years in this space.

If you want to work there, send the application now!

Roritharr2• 14 years ago

Don't forget that Valve is a 300 people company from america who has a fame of being an awesome workplace like no other game company I know of. They are already drowning in job applications from people who actively try to shape their CV to be a good applicant to them. They have the luxury to hire the 0,0001%.

I'm German, haven't finished my studies at non-famous universities and probably won't for a while since i'm working 75% of my time as a freelancer to support myself. I haven't worked professionally on any gaming project, my 3D Math skills are basic and most of my professional experience is in web projects working with Java, Javascript and PHP, although I've worked on 3D Stuff in C++ (Good old Irrlicht Engine!) when i was younger.

As much as i like the idea of not pre-rejecting yourself, there is a line to be drawn at the point where an application in itself bothers someone else simply because it's unrealistic and the rejection produces unnecessary work.

I don't want to be bothered to write offers for customers who can't afford my work either.

benihana• 14 years ago

>To be clear, this is R&D – it doesn’t in any way involve a product at this point, and won’t for a long while, if ever – so please, no rumors about Steam glasses being announced at E3.

http://www.theverge.com/2012/4/13/2947088/valve-reveals-secr...

Sigh...

haasted• 14 years ago

Also, http://www.rockpapershotgun.com/2012/04/14/that-was-easy-val...

guynamedloren• 14 years ago

This is amazing. I love the fact that there are no hierarchies and organizational induced inefficiencies, and I especially love the culture of experimentation. This leaves a LOT of responsibility up to the employees - they must be very creative and very driven for an ecosystem like this to work. I imagine one of the biggest challenges for Valve is finding employees that fit this description. There are loads of smart developers/engineers/designers out there, but not many who can be productive with little or no direction.

The author stresses that he was pushed to work on the most valuable thing he could possibly work on. For many programmers, just working on anything is okay, as long as they're staying busy and committing code. If Valve is filled with employees who are constantly assessing value and thinking about things other than lines of code, it sounds like quite an amazing place to be.

cmicali• 14 years ago

The book Zen of Graphics Programming was one of the seminal books that got me deeply interested in programming. There's a good argument that my life today is as good as it is in no small part to the hard work Mr. Abrash put into that book. So nice to hear what he is up to 16 years later

ericd3• 14 years ago

The extent to which Scifi can change the world indirectly by inspiring creative people and changing the course of their life is fascinating.

harryh• 14 years ago

I'd be interested in learning more about valve's peer based compensation system. Does anyone know more details?

xal• 14 years ago

Did you hear this? This was the sound of Seattle becoming the gravitational force to the best and brightest.

Valve has been known for all this. They clearly think that now is the time to loot the rest of the truly great in the game industry. Genie is out of the bottle.

gruseom• 14 years ago

Actually, I think that was the sound of Michael Abrash rebooting his writing superpower. Seems to me one of the biggest ways that guy could add value to Valve is by unleashing it (like he did in this post) on a regular basis.

stcredzero• 14 years ago

The success of Doom made it obvious that this was no longer the case. There was now little value in doing the same thing even twice; almost all the value was in performing a valuable creative act for the first time.

Movie studios forget this all the time. Book publishers and TV networks often seem to forget this. A lot of folks looking to do startups forget this as well.

Bamafan• 14 years ago

I don't know. Michael followed that sentence with the following:

"Similarly, if you’re a programmer, you’re probably perfectly capable of writing Facebook or the Google search engine or Twitter or a browser, and you certainly could churn out Tetris or Angry Birds or Words with Friends or Farmville or any of hundreds of enormously successful programs. There’s little value in doing so, though, and that’s the point – in the Internet age, software has close to zero cost of replication and massive network effects, so there’s a positive feedback spiral that means that the first mover dominates."

The problem with this: there are TONS of examples in the software world where the first mover DOESN'T dominate.

That said, I enjoyed reading this job post. It might be the best job post I've ever read.

EDIT: user "roc" below says this same thing, but much better: http://news.ycombinator.com/item?id=3839190

aiscott• 14 years ago

I agree with you about the movie studios, but I think Abrash overstated his point a little.

There is certainly value in taking someone else's idea and making it better. "Everything is a remix" and all that.

Perhaps not as much value as the original idea, but it isn't non-existent. Google certainly didn't invent search, but they did do it better (for a while).

FreakLegion5• 14 years ago

> Valve has no formal management or hierarchy at all.

So that's how IceFrog made it through their hiring process. I've always wondered why they were so interested in a person whose sole accomplishment was inheriting a Warcraft III mod created by someone else (Eul, real name unknown), then appropriated and popularized by a second person (Steve Feak, a.k.a. Guinsoo, now of Riot Games).

I'd certainly like to think that the series of events leading up to Valve's current legal battle with Blizzard over the DOTA trademark was an organizational rather than an ethical lapse. They seem like a great company to work for in other respects.

oconnor0• 14 years ago

This seems like a unnecessarily antagonistic comment. It's not like Eul & Guinsoo are gods & IceFrog's cretin. As I understand it, Guinsoo had some good ideas but couldn't balance them well enough to make the game fair & IceFrog's done a far better job of that.

It's not like Valve or Blizzard should own the DOTA trademark.

FreakLegion5• 14 years ago

Antagonistic toward whom? It's a fact that Valve hired IceFrog, apparently thinking it an acquisition of DotA (like they did with Counter-Strike, Team Fortress, etc.). It's also a fact that IceFrog was neither the original creator of DotA (Eul was) nor the person who elevated it from a merely very popular mod to a worldwide sensation (Steve Feak was[1]). So it's always seemed to me that Valve skipped their due diligence on IceFrog -- otherwise why hire him? He didn't create DotA, was only tangentially responsible for its success, and doesn't own it[2]. On the other hand, he brings with him the good will of the community and a claim to the DOTA trademark based on use, regardless of whether the underlying IP belongs to him.

Thus, Valve's decision to hire him has always puzzled me[3]. Did they do it simply because of his claim on the trademark and his relationship with the DotA community? I have no idea. Based on this article, it seems there's at least one other possibility: they were too trusting and didn't have the organizational infrastructure to do a thorough background check (and possibly still don't realize where IceFrog stands in the history of DotA development). We'll never know what Valve's real motivation was, but I[4] at least have lost a fair bit of respect for them since this whole thing started.

Edit: I agree with you that neither Blizzard nor Valve should have the trademark. Ideally I'd like to see it ruled generic.

1. For all that it matters at this point, Feak did this without Eul's permission, and has profited enormously from infringing Eul's work.

2. Though trademarks are different from other forms of IP and it's not impossible that Valve will end up with the mark anyway. But regardless of who owns the mark, the characters and such in DotA aren't IceFrog's IP, and could be used by other people in other games.

3. This because, aside from these considerations, he's a not a great coder and has no art or other skills. So Valve can't have hired him for technical expertise. If he was a superstar developer I'd happily chalk it up to that.

4. Disclosure: I was heavily involved in the Warcraft III modding community from before the original DotA up to the release of StarCraft II. I watched this all play out in real-time, which absolutely colors my take on it.

codexon• 14 years ago

People overestimate Valve's competence because everyone loves their games.

If you've ever tried hosting a TF2 server, you will find how riddled with bugs it is. Every other patch causes crashes.

There was a power outage at a single datacenter that knocked out all of Steam.

They also had a security breach. Several months later, they RSA encrypted people's password with javascript on top of SSL.

http://kotaku.com/5884430/power-outage-knocked-out-valve-ste...

http://kotaku.com/5884092/valve-hackers-obtained-old-steam-t...

egypturnash• 14 years ago

sends an application to Valve even though she's a 2D artist through and through

llambda• 14 years ago

Possibly the best job advertisement I've read, and I don't mean that in a disparaging way. Very well crafted and inherently worthwhile. After all, who couldn't be excited about Valve after reading this?

mkramlich2• 14 years ago

What Abrash described as Valve's management style sounds very much like 'adhocracy'. The simplest definition of it is that it is a system which is the opposite of bureaucracy.

jes5199• 14 years ago

Ok- this reminds me of Xerox PARC, or Google in the early days... What other companies these days have this level of trust, confidence, and creativity? Any?

haberman2• 14 years ago

"I had met John [Carmack] in person just once before, but he knew me from PC graphics articles of mine he had read when he was first learning to program the PC"

Ha, if that's not geek cred I don't know what is.

sbierwagen• 14 years ago

Abrash is a Big Name in the computer graphics field:

https://en.wikipedia.org/wiki/Michael_Abrash

http://www.amazon.com/s/ref=ntt_at_ep_srch?_encoding=UTF8...

haberman2• 14 years ago

I know that, I own several of his books. :)

Still, to be the person that John Carmack learned graphics programming from... that's pretty badass.

drieddust• 14 years ago

Ycombinator came to my mind when I was reading about their work style.

It is hard to believe this model is not adopted by more companies.

fijal• 14 years ago

So in my experience this works incredibly well as long as everyone is generally willing to cooperate. A single person that has dictatorial inclination can ruin it all, because you can no longer reach consensus on pretty much any topic. You have to either decide to give up on your core values (introduce voting or management) or have to think how to get rid of an otherwise interesting team member, which might not always be feasible.

surrealize• 14 years ago

The reason for the Half Life gap is a little clearer, now. :)

Valve does sound like an awesome place to work, though.

DanBlake• 14 years ago

Preface: Valve is one of those companys I admire and wish I could buy stock in them.

Saying that, Its too bad they have said they will never go public/sell. Working there would be even more awesome if you knew there was a possibility to become a millionaire before working there for 20-30 years or getting lucky and advancing up the corporate ladder to where you could make 200k+ a year.

I wonder if facebook would have been such a hot place to work at if zuck said right at the start he would never go public and never sell the company.

Its obviously in a startup guys DNA to look at day jobs and retirement plans in disgust ( we are all trying to make tons of money now, right? ). I accept that type of thing is fine for most people and thats cool with me. I could personally never work for a company knowing that the most I can hope for is a higher salary though. (Unless you climb to the truly top of the ladder, where bonuses and whatnot await you). Working at valve, I wonder if there is going to be any stories of chefs/masseuse's who become millionaires because they took the risk early on. Probably not.

Again, not that there is anything wrong with that. Valve is never claiming to be a startup. Just saying, I wish they were.

dhimes• 14 years ago

That's a tough game- the startup lottery- and there are professionals playing. Sure a few people manage to win and become millionaires by being there when it all started. But far more common is the scenario where the employees take "equity" over salary based on this dream, and the equity evaporates in later funding rounds.

I think a system that is much more fair to the employees is this what Peldi's doing at Balsamiq: http://blogs.balsamiq.com/team/2011/09/12/profitsharing/

Here, employees get regular bonuses based on how well the company is doing. Everybody is sitting on the same side of the table. And if the company is profitable, employees win- a situation that is not always the case with stocks.

brown9-2• 14 years ago

I think it depends on what you define as "hot" - if a company is sought after to work at because of the working conditions or the perceived monetary upside.

I think it's fair to say that the pressure to make investors and Wall St happy each quarter would ruin the kind of long term focus Valve has.

lifeformed• 14 years ago

I think that's a good thing, because it makes sure they get people who are in it for more than the money. I'm not saying that wanting to strike it big in a startup is all about money or is a bad motivation, but rather that it unambiguously paints Valve as a place where employees don't even have to think of it as a career. People are there because they want to make cool things with cool people, not because it's a good stepping stone to another opportunity. I think I read that their retention rate for employees is like 98% or something?

I don't think they need promises of money to attract the kind of people they're looking for. Plus, I'm sure they pay very well, being a 200-300 employee private company bringing in billions in revenue.

Schwolop• 14 years ago

From the article: "(although there are generous raises and bonuses based on value to the company, as assessed by peers)".

stashdot• 14 years ago

"I had a great year or two learning about natural language before figuring out that it wasn’t a problem likely to be solved within my lifetime."

That' s a bit disheartening.

Abrash on Valve: How I Got Here, What It's Like, and What I'm Doing

Related Stories