Gabe Newell: On Productivity, Economics, Political Institutions, and the Future of Corporations

We are obviously very honored to have Gabe Newell, founder of Valve Corporation, here with us today. I know that for most of you, Gabe Newell does not need any introduction. Valve is an immensely successful gaming company, and Steam has evolved into a major electronic market with 55 million customers around the globe. Today, Mr. Newell will discuss his views on productivity, economics, political institutions, And the future of corporations, Mr. Newell.

So I'm Gabe and— so Valve was founded in 1996. We're a privately held company and that's actually a consequence of other decisions that we've made that I'll touch on later. We made single-player games like Half-Life, multiplayer games like Team Fortress. We built an internet services platform called Steam that, as Cherie said, has 55 million users and 2,500 different applications running on it. So the question would be, other than my charming personality, is Valve interesting, right? Like, why is Valve an interesting company? So if you use super gross metrics, Sounds like it could be interesting. We grow about 50% a year and have since we started the company. And we tend to make people very productive and they add a lot of value. So if you look at Apple or Google or Microsoft, we have much higher revenue and profitability per employee than they do. And, you know, we generate more internet traffic than most countries, right? I think we're the 4th largest bandwidth consumer in the world right now. So that says something is happening here, but is it interesting beyond just being another company that makes a bunch of money? Well, my path, which I'm going to use to sort of illustrate— I'm going to sort of walk you through what— how I learned and what I learned that applies to Valve, and hopefully use that to motivate your own understanding of the choices that we've made and what, if any, significance they have. So I was at Microsoft from 1983 to 1996. I started off working on the Macintosh applications and then switched over to operating systems. Now, one of the things was that in 1983 there was a company called Businessland. Businessland is sort of an answer to a trivia question nowadays. At the time they were sort of the biggest seller of PCs to small businesses in the world, and they would take 50% of the gross revenue of a sale. So somebody like Osborne would sell them a machine for $1,000, and they'd mark it up to $2,000. If you flash forward to 2012, oftentimes our games are sold at negative margins at retail. So we'll sell in a product into retail, and they'll actually sell it to consumers for even less. So why— they're hoping that maybe they'll generate some foot traffic that will help offset the negative margins they're taking on us by, you know, selling something that is likely to attract people's attention and then maybe you'll buy, you know, some blank CDs or something. So what's happened? Why has this part of the value chain lost so much control over the whole process? And is that generalizable? Is it just a unique instance or is it something that's symptomatic of a broader set of trends? Well, in the early 19— my own experience in answering that question started sort of in the early 1990s. So Microsoft had no idea what was happening in the real world with our products. We'd have reseller lists. So you'd sell your product to somebody else who would then take a pallet that we would physically put together. They'd break it down and then they'd sell the individual products out to lots of little stores. And those resellers represented only one small fraction of the total product being sold. They had nothing to do with anything other than the North American market. And so, but we would post these reseller lists where they would tell us what products were selling out of their warehouses, and that was pretty much the complete insight we had into how our products were performing in the real world. So product managers would obsess over relative motions of 1 or 2 spots on a top 10 list, but the idea that you would know what was happening through an OEM channel or happening with corporate accounts or happening anywhere outside of the United States was more or less science fiction. This was, you know, a well-run company in the early 1990s. '90s. So at one point the Applications Group had to make a bunch of decisions like, you know, hey, these crazy Windows guys, they're saying that Windows is doing super well, but, you know, Lotus is telling all their corporate accounts that Windows is all— the sales numbers are all fake and that people are not adopting Windows in any significant way. This was going to materially impact how quickly they tried to move all of their products over over to Windows rather than continuing to be character-based. So Microsoft commissioned this really large study where rather than looking at the sales side of it or the reseller side of it, they'd actually go out into the real world and see what people were actually doing. It was like the first time Microsoft had done it, is to, you know, look at, you know, 10,000 people's machines to figure out what they were actually using PCs for. Well, it turns out they were using them for porn and video games. And that part of the study was immediately ignored. But the good news was that if you extrapolated those numbers, was that Windows was actually being used on 30 million people's PCs in the United States. And this was cause for enormous celebration. Thank God they're actually using it rather than just getting it shoved onto their Dell or Compaq machine and immediately uninstalling it and going back to character mode DOS. But the thing that was really striking to me at the time was that Windows was the number 2 product, and probably most of you know what the number 1 product was. Anybody?

Doom.

Early— yeah, it was Doom. So Doom was one of the first first-person action games shipped by a company called id that's located in Mesquite. So somehow the largest software company in the world was being outdistributed by a 12-person company in Mesquite, Texas, and there was no conception, there was no model for why this made any sense. How could they possibly have done that, right? I mean, it took years to build up enough distribution strength to go from number 7 to number 6 in the word processing category. much less come out of nowhere with a product that would be on more desktops than the most important product that Microsoft had. And at the same time that Doom was out there with a completely different approach to connecting users with value, people inside of Microsoft were arguing, oh, we need a 500-person— you know, Rob Glaser, who eventually started RealNetworks, was saying— We can't be competitive with Novell NetWare until we hire a 500-person sales force, and all they're going to do is sell LAN Manager to the resellers. So they weren't even going to sell to customers. They were investing in an indirect channel and making this giant— building this giant organization that continues to live on within Microsoft all these years later, who were not actually directly involved in making that sale. So that was sort of the landscape at this point in time. And there were 2 other interesting data points. There's another company that was called Ventura, who made Ventura Publisher. Have any of you ever heard of that?

No.

It was the same kind of thing as Doom, and they were the first people that I ever met with who explicitly said retail distribution and marketing is a commodity and we're not even going to bother investing in it. Instead, we're going to build a great product and once we have a great product, we'll turn around and then find— we'll put it out to bid. Who wants to give us the highest percentage? They ended up going with Xerox and Xerox eventually proved that sales and marketing aren't quite as commodified as the Ventura guys had thought they were, so they probably underperformed what they could. But they're the first people to say, let's just ignore all of this other stuff. It'll actually make us worse. We'll just focus on the thing that we do super well, and this other thing is going to become declining in value constantly. At the flip side of it was the IBM PCjr. And the interesting thing about the PCjr at the time is, I guess 1987, was that the burdened cost of shipping an empty box by IBM was over the retail price of the PCjr. In other words, they couldn't possibly be profitable ever selling products in that category. Their choices about the organization, their investment in direct corporate sales, the way that they had structured themselves financially made it impossible for them to even move an empty, well, a nothing through their inventory process and into a customer's hands. So both of these things make you start to think about what's going on. I'll sort of do a shortcut to the conclusion, which was there were structural changes that were affecting the relative value of line of business functions, and that whole categories within typical corporations were essentially being made obsolete. That communications to your customers were on an exponentially decreasing cost curve and everything that was associated with talking to a customer or delivering a product to a customer. And this would be true whether you're doing virtual, you know, digital products or physical products. It would just be more obvious when you also had low marginal low incremental cost, low cost for incremental production, right? All the costs associated with building a video game tend to be associated before you sell your first copy. Selling the copy plus one is relatively cheap. So even though this thing would be true, just as true for John Deere tractors as it was for video games, it would probably be more testable and provable in spaces where you could easily see that the that the low-friction delivery of a good didn't involve significant incremental costs. So what did that mean in terms of how you were designing a company? This is the question I was trying to answer in 1996, because we'd go out and we'd meet with all these companies. You know, you meet with insurance companies and airline companies and startups in Silicon Valley. And you had to come up with some way of thinking about what the design was. And the thing that we became convinced of, Mike Harrington, who's the other founder, and myself, was that everybody was going in the wrong direction. There's sort of a movement towards outsourcing. And outsourcing is essentially, where can we find the lowest-cost English language speaker somewhere in the world? And we'll give them a job and they'll do it just as well for a lot less money. To us, that seemed exactly the opposite of what you should be doing. And what we decided was that we were going to buy the most expensive talent that was out there in the world, that the opportunity was to— that those were the people who were least correctly valued. By talent, which is a word I hate, I just mean the ability to be productive. Now, one of the things that sort of helps people in the software space think about this is that it's— it's relatively easy to look at relative productivity of people within that space. So at IBM in the 1980s, typical productivity would be 1,000 debugged shipped lines of code per year, right? That was the metric that they used for sort of their median employee. Whereas when we were shipping Half-Life 1, one employee, Yann Brunet, was shipping 4,000 lines of code per day. So there's pretty clearly, you know, it's easy to see that there's— even if you're not arguing about whose lines of code are more useful, and arguably Jan's lines of code were way more useful than somebody developing a 3270 terminal emulator for OS/2, even if you're just looking at the raw production, it's easy to see that there's huge variation in productivity. One of the sort of tenets that we used was that that was probably true across a much wider range of functions. Right? So we started from the assumption that even though it's easy to see that there's huge variation in productivity in software programmers, there was probably that same variation in a lot of other roles. So when we designed Valve, we said, okay, that's what we're trying to do. As you know, everything goes back to this fundamental question of how do you attract and retain the most highly productive people in the world? Because if they're— if our thesis is correct, Right. That's where we're gonna create our greatest incremental value, right? Any, you know, we could hire a whole bunch of— there was a big thing a couple years ago where everybody was trying to hire low-cost content producers in India and China. And we were trying to do the opposite. We were saying, if somebody somewhere, you know, somebody working on a feature film production in New Zealand is making $200,000 a year, If they come to Valve, they should be making $500,000 or $5 million, creating that much value. You know, we'll split the difference and go from there. So there are consequences of those choices. Nobody's asking questions, which has me a little bit worried. There are consequences. One, Valve is not a publicly traded company. Being a publicly traded company adds a bunch of headaches, and it didn't really solve any problems For us, right? It meant that control around decision-making now involved third parties. So for a developer or somebody building something at Valve, it's like there's the customer and they're the person that you're trying to make happy. And from the time that you make a change in a product, it's 15 minutes worst case before a customer is actually using that. There's no approval process. If you make a bad decision, it's 15 minutes until you've fixed that problem. You don't go to board meetings where the board argues about what the series, this third series of venture capitalists are worried about, you know, dilution and hitting certain targets, right? There's no notion of a distribution channel who says we're really looking for something that fits in this particular particular slot at this particular time. The whole point of being a privately held company is to eliminate another source of noise in the signal between the consumers and producers of a good. One of the other things is that the kinds of people that we thought fit this model don't need titles, right? That titles are actually the enemy of their enemy rather than being something that makes them more productive. Oftentimes their solutions in one generation are different than their solutions in the previous generation, and that an explicit organization captured in titles makes it less likely for them to have the insights and to build the next generation of whatever it is that's going to be exciting, right? Whether it's— The great example from Half-Life 1 is that the Uh, guy who designed the skeletal system— animation system also had a Bachelor of Fine Arts and could also build environmental art.

Mm-hmm.

So if you go to a movie nowadays, the idea that you would have a single person doing even as much more than a subset of any of those roles is mind-boggling. But there are a set of, uh, experiences that you get in Half-Life 1 that are entirely a consequence of the person working on it being able to change the environment change the code or change the animations depending upon what was the most tractable way of solving the problem. So titles and organizations keep people from sort of properly encapsulating a problem at a point which allows them to be most productive. Does that make sense? Yes.

Was that something that you had at the outset of the creation of Valve?

Yes.

Or was that something that you decided along the line that was how it should be after having gone to a traditional organization structure?

No, we always— this was part of the design of the company from the beginning. One of the other things along those same lines is the thing that people sort of— to us, we've been working this way for a long time, so it doesn't seem nearly as peculiar as it does to people on the outside. So another example that people tend to be a little bit surprised by is that management is a skill. It's not a career path. It's— everybody is sort of— we assume that everybody is a mix of individual and group contribution. And there are a set of tasks that look like, you know, that evolve over time related to project organization and just sort of keeping things going. And usually people refuse to do it twice in a row on back-to-back projects because it very much is, you know, a service job. It's like, my job is to entirely define myself in terms of the productivity that I enable in other people. The next project, I just want to not— you know, that's a very stressful job and it's hard to measure your own productivity. Like, everybody else says, hey, you know, Jay, you did a great job, you should do it again. And Jay's like— Screw you guys. So usually we look for some younger sucker to give the job to who think that, you know, who have some old notions about management and, you know, it's authority within a hierarchy related to decision-making. And then they find out that it's, you know, working really, really hard to make other people be more productive. If that doesn't blow your mind, we have no QA department, right? We have no marketing department. We have one guy who calls himself a vice president of marketing because it confuses people if he doesn't tell them that when he's talking to them. But everybody at the company is— we assume that their job is to talk to customers. Yep?

Outside of the company, how are you able to to hire and retain the sort of talent you were looking for?

So a lot of times it was just personally pitching people, saying, you know, this is why we think this is an interesting opportunity. Some of the people that we thought were going to join us at the beginning didn't. You know, we were like, let's go, let's storm the castle. And then there's like, where'd everybody go? So once you convinced people that you were doing something interesting, then they would start to rope in their friends. So a lot of it ends up being social networking at the end of the day. John Carmack was pretty helpful at getting this going. Michael Abrash was another guy. We were also willing to— we were willing to take what other companies would say were risks. So one of the first programmers at the company, his previous job was manager of a Waffle House. And he was one of the most creative, and still is one of the most creative programmers in the industry. In the back of the room?

I was wondering, Has your flat company structure changed since the days of Half-Life 1? Because with the increasing cost of software development, one person can do, I guess, less now than they could before. Has that changed the way your company works at all?

Well, our company evolves all the time. So I'll try to give you an example of how we evolve. I am a big believer in one person, one office. DeMarco's Peopleware, if anybody's read that, you know, I sort of treat that as a bible. And so I was like, I hated cubicles, like, with a passion and a vengeance. It's like, you know, they grew out of this Data General philosophy of if you're screwing your employees, you're probably doing something right. And so I'm— I was like, okay, everybody always has their own office. And, you know, and you've got a do not disturb button on your phone and you have a door you can close and you can decorate your office however you want. And people kept sneaking into other people's offices and then they started tearing doors down. So everybody had their own office, but nobody was ever in it, right? So they were all congregating in our conference rooms. And so we had all these empty offices and all these clustered—

People.

clusters of people in conference rooms. And, you know, that's when I realized, you know, they were making the right decision, right? They were deciding that the— there were benefits to being in closer proximity to other people and that Gabe's rigid adherence to one person, one office was actually hurting their productivity. So even, you know, after being hit on the head, long enough, I realized that everybody else was making a decision that was correct for them. And so now we just say everybody designs their own office space. So our largest single room right now has about 80 people in it. I share an office with 2 other people. But each person, you know, has to— that's one of the things that you think about. You think about how you communicate with other people. You think about what's an optimal work environment for the goals that you have. Desks are all on wheels. We only have 2 contact points with the floor, so you can pack up and join a different group of people in about 15 minutes. Some people are working on the design of— if anybody's done paired programming, there are some people who are trying to build a desk that's more optimal for paired programming and— Giving you a shared thing. And they have to make the decision whether it's more worth their time to work on something like that than it is to do the work that they already are. So we just sort of said, okay, each person has to make their own decisions about what they work on, who they're working with, what are the tools they need, and what, you know, what their offices ought to look like. But that's how we evolve, usually by people deciding that I have a bad idea. Another thing that became very important sort of later on, but it is pretty much part of the design of Valve now, is if you're not making quantitative predictions, you're probably doing it wrong, or you're probably not doing it as well as you can. And that's sort of become kind of critical to how we operate. You have to predict in advance. Everybody can explain anything after the fact. And it has to be quantitative or you're not being serious about how you're approaching the problem. Yes, in the green shirt. So I'm at a web company that's actually experimenting with the bell model.

What are some gaps or problems that you've encountered with the kind of flat hierarchy, some lessons that have been learned?

You have to be really aggressive about firing people. We haven't done a really good job with interns or new hires. It's kind of a sink or swim thing. You know, people have to take it seriously, right? It's an engineering problem, right, in the sense of you have to make decisions, you have to measure outcomes, you have to make changes as a result of it. So, you know.

Thank you.

You know, I would have trouble working any other way now. I think most of the people at Valve would have trouble. There's something we call the— we sort of somewhat unkindly call the beaten wife syndrome, where people come in from other industries and really struggle. The worst are people from the feature film industry, where they've been sort of taught that anytime they show initiative, that somebody's gonna leap out and smite them for doing that. And so that can— usually takes about 6 to 9 months for people to really sort of internalize the working model of the company. So those are some of the things we've encountered.

Yep.

Do you think that that might be the reason more companies don't follow your plan?

It's sort of a personal responsibility type of thing?

I'm going to come back to how I think things are changing and why they're going to change in a little bit. And maybe if we come back to that question, you may have a different way of asking it. So to get onto that subject, these issues are going to collide in a couple of slides. I'll start off with something. So how do we make a single-player game, right? A single-player game was the challenge that we first had when we were building Half-Life, which was how do we make any decisions? People would say we should make games more realistic, and we sort of went down that rabbit hole. It turns out it's a horrible, horrible design trope to use. And so we had to come up with this notion that games are things which respond to agency on the part of the player. If the player performs an action or has a state, you need to recognize that and respond to it. That's a fairly abstract way of saying, if I shoot at a wall, something should happen. So when we started with Half-Life, your gun would fire but the wall wouldn't change. The wall was essentially ignoring you. So then we'd put a decal in and people— testers would go and spend an hour just drawing things on the wall with their machine guns. And we said, okay, so this looks like a useful—

A useful thing.

heuristic for making game design decisions. In Half-Life 2, a piece of technology that was really hard, but turned out to, in our mind, to be super valuable, was modeling how people's eyes actually work when they're looking at you. So eyes are not spheres. So when your eye is rotating, you're going to see something that's like the highlights, the glints on your eye are going to move differently than if you model them as spheres.

Right.

And people, turns out you have part of your brain that's hardwired to know where somebody's looking, right? Even though somebody's in the second row, they can tell that I'm looking at somebody in the third row. And so spending a bunch of time to make them aware of where— the characters aware of where you were in space and that they're looking right, correctly at that person, it turns out that people suddenly say— Oh, these characters are nicer. These characters are more like me. These characters are smarter, right? So something as seemingly irrelevant to a game as making eyes look— work right turns out to generate the sense that this game is more fun. You know, and that's how a heuristic leads you to interesting conclusions that you go and do. Now, of course, all this breaks down when you're trying to build a multiplayer game.

Right?

You know, a single-player game is sort of like a feature film where your lead actor is kind of retarded and autistic, but you can sort of think of it as a feature film. In a multiplayer game, you— these rules that you've come up with sort of don't work anymore. So an example would be in Counter-Strike. We put the riot shield in and our player numbers go up. We take the riot shield out and our player numbers go up. So how do you explain that? You need a different way of thinking about what you're doing. You start to think after a while that multiplayer games are all about externalities. They kind of look more like operating systems or a sport. So in terms of how they behave, they behave a lot more and value is created a lot more like a spectator sport than a feature film. So that's pretty cool. The problem was, jump forward a couple more years, that this great model that we had for thinking about single-player games or multiplayer games seemed to be breaking down again. So there are a couple of interesting examples of that. One is free-to-play games. So on the surface, free-to-play games sound like a horrible idea. I will give my product away for free. You know, most people just sort of stop right there and go, okay, that's horrible. But in a free-to-play game, what you're really doing is you're creating a lot of goods that are related to status and affinity and hierarchy, right? You're creating—

Status.

a whole bunch of goods there, and the marginal or the incremental value of an audience member is greater than the incremental cost of making that person an audience member. So typically what we see with the free-to-play game, which on the surface sounds like suicide, is that your audience size goes up by a factor of 10 and your gross revenue goes up by a factor of 3. So since your— the incremental cost of another audience member is fairly small, just the cost of distributing those bits to those customers, your profitability tends to increase a lot more than a factor of 3. Now, okay, that's sort of puzzling. And then we start to see this thing occurring in lots and lots of games where you have markets or auction houses. So you have trade in goods between different customers. And there's this sort of appalling thing that happens where somebody will play your game 20 hours a week for 4 years. And then the value of that all goes to zero, right? So it's like you bought a house and you made a bunch of improvements on the house, and when you move to your new house, you have to start over. You get no value from the investment that you've made. Clearly, these markets and auction houses are valuable, but— or the assets that you've accrued were valuable to you enough to justify this tremendous expenditure of your time. But games seem to have this really sort of whimsical notion of your property rights. We're also seeing this huge uptick in user-generated content. So to be really concrete, 10 times as much content comes from the user base for TF2 as comes from us. So we think that we're super productive and kind of badass at making TF2 content, but even at this early stage, we cannot compete with our own customers in the production of content for this environment. So the only company we've ever met that kind of kicks our ass is our customers, right? We'll go up against Bungie or Blizzard or anybody, but we won't try to compete with our, with our own user base because we already know that we're going to lose. Once we start building the interfaces for users to start selling their content to each other, we start to see some surprising things. So right now I think the, the, the most anybody has earned in a single year is about $500,000. So they're making content, selling it to other customers, and we have a revenue share with those people, and their takeaway is $500,000. The first 2 weeks that we did this, we actually broke PayPal because they didn't have— I don't know what they're worried about, maybe drug dealing. They're like, nothing generates cash to our user base other than selling drugs. Like this. So we actually had to work something out with them and said, no, they're making hats, not— But now that we had this notion of— oh, and another thing that was weird is we knew that people at other companies, other game companies, had employees who were making more money being users in our framework than they were being employees at their company. So we're like, okay, this is weird. And then it got more complex because we started to see things like inflation. We started to see deflation. We started to see users creating their own versions of currencies, mediums of exchange. Countries started to create regulatory structures. So in Korea, you actually have to create the equivalent of a W-4 form for your players to account for the virtual income they get in playing your game. And so you're like, okay, wait, should we increase the drop rate for those customers to offset the tax, you know, the sale, the implicit, the sort of sales or income tax that they're having? And what do we do about, uh, uh, purchase price parity, right? Should we adjust drop rates to like provide welfare to people who are playing in low income, or, or do we drop the value of their drops to reflect the fact that they can trade their hats for cheeseburgers for, you know, fewer— I mean, do we— how do we decide how we are, are creating productivity in this, this environment? We started to see trade. The trade is exploding, uh, even though currently it's, it's a, it's a, it's a barter economy. We started to have problems in liquidity where at different times of day certain mediums of of exchange sort of dried up and all of a sudden you had sort of weird little mini financial crises erupting in your TF2 hat exchange community. So when we started to see these kinds of issues, we said, oh, we should probably start talking to an economist. Yes?

Where did the value of hats come from? Because it literally is Like hats or like small stat boosts and things like that. Like, what makes somebody spend $5 versus $32 for a cosmetic change in like a video game character?

I don't think that the value for that is any different than the value of lots of other kinds of status goods or, you know, why do people buy Porsches? So I don't think that there's anything super different about how value gets created. I mean, money. I mean, if you look at— people say, but these things aren't real. And then you go, well, what about money? And they go, that's not fair. So it speaks to— so one of the things that you're gonna hear me say is that we're trying to optimize productivity. And that requires you to start thinking in terms of productivity in terms of what. And so it is an issue that we're struggling with right now. How do you think about what is the value and how do you feel like you're not creating, you know, like asset bubbles are something that worry us. And, you know, this is where, you know, hopefully an economist, we also believe, we're also particularly, I'm worried that there are a set of initial conditions that have, different outcomes and that by putting in the right kinds of structures and disciplines early on, we can avoid a lot of turbulence subsequently. Yes? What about crafted items? I'm thinking EVE Online. It all starts with ore. You want to build a pipe or whatever, you've got to provide it in some way. So what we think is that you need some sort of market and that what our job is, is to maximize the productivity of users in creating digital goods and services, that markets will determine what the marginal value-add of each of those activities are. But the kinds of ways in which people create value are— creativity in creating frameworks for that are going to vary. Now, you can't just define productivity in terms of shit you give to a customer, right? Then you've sort of missed the whole opportunity for, you know, if you're just raining Hats onto your customers, eventually you're probably going to suffer huge inflation and so on. So the way to think of it is that there's probably going to exist a central— I'll just use the term economy— and that games are all sort of instance dungeons hanging off of that. And within a game, I'll be able to create goods and services which I exchange with other people in other games. Some of the things that I will do is, hey, I got this hat, right? And then somebody else will say, I actually designed that hat. And the designer and creator of the hat is probably going to be a higher value person than the person who simply traded or acquired this asset in some other form. So you're going to have a bunch of different ways that people are going to be creating things. One group of people will simply be arbitraging. looking for trading opportunities and creating liquid markets in hats, right? But other people were saying, oh, I'm going to build a game and my game contains animations and models and environments. And the steps that we're taking right now are to plumb this notion of both ownership and authorship throughout the entire system. So I'll be able to create something in one game and exchange it for value with somebody else. So everybody, I think, can understand this notion of creating a model and exchanging that. That's a fairly easy way to think about productivity. But another, you know, taunts or animations are also ways in which you create value. But we actually think something like being a really good player is a super valuable thing for the community. And the challenge isn't that you've created value. The challenge is coming up with a monetization method for those people. So in Dota 2, we say, here's Na'Vi. Na'Vi is one of the best teams in the world. They're from the Ukraine. In particular, one of their players named Dendi is like one of the most charming people in the world. How many people know who Dendi is? Okay. And so when Dendi is playing or when Na'Vi is playing, You can purchase a banner and depending upon the outcome, you may or may not be rewarded with an item for showing your support. And a percentage of the banner revenue goes to the team. Turns out to be a really good way for the team to monetize the entertainment that they're creating for a bunch of people, right? So we're looking for ways to say, Dendi, you're awesome, you're entertaining, and so on. Trying to sell ads on your YouTube channel, here's a much more direct way for you to engage with your audience and for them to also benefit from waving your banner while they're, you know, logically waving your banner while they're playing, right? So, you know, pretty quickly we'd expect somebody like Dendi to be making, you know, $100,000 through banner sales for his team a year. And that to us says, okay, We think this is valuable. We've done it. People are spending the money. Dendi is now going to even spend more time doing entertaining things like playing Crystal Maiden as a carry with, you know, 5 Blade Furies, which if you're a Dota nerd, that's like— sounds ridiculously cool. And if you're everybody else, it doesn't. So the things that you do in these games need to have persistence and they need to be exchangeable. And we need to look at more ways for users to generate content, to be productive in this audience, to accumulate assets that are exchangeable and retain their value. So an example of that is, you know, I'm going to actually jump ahead in my own slides. To talk sort of concretely about ways that we instantiate this new understanding of what we're doing. So right now Steam is essentially a curated store, right? It's a bunch of other things, but you can also think of it as a curated store. So we have these really hardworking people that other companies call up and say, hey, would you put my game on Steam? They're like, oh, you know, We're putting up, you know, 3 games a day right now and we have to create these capsules and blah blah blah. So essentially, whether we want to or not, we're sort of becoming a bottleneck in terms of content being connected with users. Now there are reasons why you might want to create an artificial bottleneck between content creators and consumers. For example, if you want to shift where relative value is towards controlling distribution, it's great if you can create artificial shelf space scarcity. But that's not really what we're trying to do. So rather than having this curated store, we're going to say, okay, if we're thinking about this correctly, it really should be sort of a network API. There should just be this publishing model. And yes, you have to worry about viruses and malware and stuff like that, but essentially anybody should be able to— You know, make, you know, publish anything through Steam, right? Steam is just a whole bunch of servers and a whole bunch of network bandwidth, and if people are interested in consuming the stuff that you're putting up there, then a good, you know, a collective good is, you know, is going to be there. So rather than us sitting between creators and consumers, we're going to get as far out of that connection as possible, right? So Steam stopped being this calling up Jason Holtman and yelling at him until he puts your game up on the Steam store and instead just becomes a network API. So that's a consequence of our perception of the direction that the industry is going. So on top of that, we'd also say right now we have this— so in Team Fortress 2, we say, you know, anybody can make content. So people make goofy, you know, Yushankas for For all the characters, you know, which I didn't even know that word until I saw it appear in TF2. There's no notion of privileged content, right? Right now in Steam, the store is privileged content, right? It's like a store is like a collection of editorial perspectives on stuff. And what it should be is user-generated content. That means other companies might create their own stores that are connected to the Steam backend. But anybody would be able to create a store. And right, there's some market-based mechanism for determining the price that a store gets to impose. So anybody who tries to charge too much for the goods that are flowing through that store will get priced out of the market. But if you have a collection of games that you own and you play, and one of your friends decides to buy a game through your trivially created store, then you should get a percentage of that revenue. you. Now, most people won't have interesting collections of games or interesting friends, I don't know, but some people will go to a lot of effort to create, treating a store as another type of experience. And some people, like the guys that I would have loved, Old Man Murray would have done an awesome job. Yahtzee would probably be another person where either through affiliation or because of his editorial process, I'd actually purchase a product through him rather than through some other way. So you take 2 things which are tended— you tend to think of as super valuable assets that have to be guarded really carefully, you know, deciding who gets to be on Steam and deciding how stores are presented to consumers, and this is how we rethink them, right? It's a generalized network service. And a store, rather than being some unique special thing that represents, you know, Valve having control of this, instead it turns into, oh, lots of people are creating really entertaining— will add a lot of value to that process. And through the market mechanism, it'll justify— or the audience will reward or not reward people for building entertaining stores. Does that make sense where we— take these sort of high-level concepts about how we're going to change and turn them very concretely into a set of product changes and system-level changes. As I said before, one of the things we have to do is plumb through this notion of authorship and ownership. If I build a texture and somebody builds it into a model and somebody takes that model and builds it into a level and then somebody else sells that level to a customer, whatever revenue share that I should get depending upon what are the terms that I've provided to other people to use it, all that has to be tracked and accounted for. And it means that we need to spend a bunch of time building these kinds of frameworks that help enable people to be productive and to be rewarded accordingly. Even with our primitive versions, we're helping people at other companies be more productive than their corporations are doing it. But we're constantly thinking about what are the ways and what are the most useful fundamental ways of enabling that productivity. One of the topics that Ioannis and I are working on right now is making prediction markets something that are easy for people to author and easy for people to participate. We think that there's a huge amount of value in On both the entertainment side and on the information capture side to having prediction markets. And we also think that once everybody gets used to having prediction markets, they'll be very beneficial tools. I am absolutely certain that within 3 months of prediction markets going up, that the prediction markets will be vastly more reliable than a typical game publisher's own predictions and market forecasts for their own products. So, I bet you could take 20 gamers, build a prediction market, and nail the next Call of Duty: Black Ops 3 sales numbers much more accurately than any of the market services that we use. So, there's an example of something that we're currently missing that would be hugely valuable to economists if one of the prediction markets is Nominal GDP level targeting, my personal favorite. But it would also be entertaining if some of the prediction markets are how long until Lindsay Lohan goes into rehab again. We have we have. Lots and lots of data, but analysis of data is hard. Not in terms of— we want to take all that data and make it as widely available to people as possible, right? If building a hat is valuable, analyzing data from Steam is going to be a lot more valuable. So anything that we do, we're trying to figure out how to systematize and create a framework for our audience to participate. Right? I do my job, but if I can figure out how to get Reddit to do my job, they'll do it a lot better, right? You've been waiting with a question. Yeah, I just wanted to see if you were going to touch on security. It sounds like with building these worlds that it would almost be easy for a hacker to figure out how to use your world to do something that you really, really, really don't want them to do. First of all, there are multiple kinds of hacking. So one of the biggest mistakes game developers is to mistake— one of the biggest mistakes game developers make is to mistake hacking or to mistake a genuinely entertaining action by a player as hacking. So the example game developers tell each other is when Lord British was killed during his first speech in Ultima Online, they rolled the world back rather than recognizing that it was the coolest thing that had ever happened in Ultima Online, and they should have let that instance turn into whatever an instance turned into after they had killed off Lord British. So you have to, one, you have to be really careful. Second is, yeah, you have to have really, I mean, it's a general characteristic of social systems that people have to have confidence that in the future, right? And that's true whether it's your house and the stuff that you have inside of your house. If property rights are provisional, then your value of property changes accordingly. If you think— Yanis and I were talking about communities where the homicide rates are incredibly high compared to say, North America, and you'll see the consequences of that lack of security throughout the entire thing. So, yes, it's important. Yes, there are a whole bunch of technical problems to it, and the degree to which you fail to provide adequate security, you know, those consequences will ripple throughout the entire world, meta-world.

Yeah? Is the desire to create a new internet kind of crowdsource a lot of the labor or as much of the labor as possible. Is that— was that part of what led Valve to release Source, like releasing Source Engine as a product available to people?

I don't think— I mean, I would say that Source is kind of a— releasing Source if your goal is to enable productivity has not been one of our most successful efforts. It's the— it's just not a It limits the ways in which people can participate and add value. It's hard, right? It's too hard for most people to incrementally get into it. So the way we sort of think about being a member of this economy is to think of it much more like an MMO. Like you need to walk people through stepwise, like here are your quests, like, you know, sell, you know, buy a hat, sell a hat, okay, sell 10 hats, okay, oh look, you got an 11th hat when you sold your 10th hat. Okay, now make a color change to a hat. And then sort of walk people up in the same way that MMOs walk you up how the experience and the ways that you participate. But that's how we think of it, is that growing people within this economy looks a lot like walking people through an MMO experience. In the back of the room. Can you spell your team?

I was wondering if you could comment on motivation theory and how that plays into your economics here. Like, are the users generating content, are they so much more productive than your internal team because they're intrinsically motivated, because they're self-directed, or are they extrinsically motivated because it's a commoditized product that they're making?

They're— they do it for both. They get a lot of cash and they get a lot of other rewards as well. We're not trying to separate those rewards. So it's also the way we think of it is that money needs to flow as a signaling tool, that it's a much more— if people want to do something because their friends think it's cool, that's still going to be there. But in order for people to really assess like what they're doing is valuable or not, you need to have the— you need currency to— so they go, oh, wow, I made 10 times as much money doing this as I did this. So the signaling becomes more explicit and comparable. So why do companies— why do corporations exist in the first place, right? I mean, if you're an economist, one of the challenges is to explain Why do we need corporations at all, right? Why isn't the price mechanism sufficient to allow good organization of production and allocation of capital?

I hate to interrupt, but for those of you who have questions, I'm going to ask you to hold them, and Mr. Newell has graciously agreed to do a repeat performance And we have a whole other crowd downstairs waiting for him. So if we could take about 5 more minutes and wrap this up, because a whole lot of your colleagues are waiting to hear from him, and he has agreed to speak with them next.

So thank you. Okay.

So why are the PC and Internet important? They create frameworks for innovation. and competition. And one of the most important consequences is that there are sort of variations of Moore's Law which are making not just computation but communication and access to data exponentially cheaper. And those are continuing to accelerate. We all know that when things get computationally cheaper, that's valuable. But there are also consequences to communication and access to data getting cheaper. I had lots of anecdotes about that, but I think people understand it. One of the things from an economic perspective that's important to realize is that cheap information reduces moral hazard, agency costs, and information asymmetry. The way— the point that I make to people who are of your generation is you can't lie to Reddit. It's remarkable how many people try, but they don't understand That Reddit's ability to detect bullshit is insanely high. But that, from an economist's perspective, translates to something that's really incredibly important about information asymmetry and moral hazard. You'll hear me— people get confused when I sort of rag on consoles because, oh, well, you have a competitive platform, blah, blah, blah. The thing that I dislike about closed platforms is they— I can go into this in ridiculous amounts— is that they increase sort of the chunk size of competition, which I think is a bad thing. They increase costs of market entry, so people who have good ideas, it's actually a lot more expensive for their productivity to be monetized essentially in closed platform spaces. And they also delay standardization. So standardization is amazing. mechanism by which— well, if you ask me, and I look at most of the behaviors in corporations, it looks like rent-seeking behaviors on top of friction, which is a really dark way of thinking about the roles that people have in corporations. Standardization is a great way of eliminating the ability to rent-seek. This is probably an entire presentation on itself. But I would sort of say that to us in this room, it seems fairly obvious that the internet does a better job of organizing a bunch of individuals than General Motors or Sears does.

Mm-hmm.

So, you know, and that a lot of these companies, things which were perceived as being capital, assets end up disturbing their decision-making so that they actually end up dooming those companies. I don't know how to do that in 5 minutes. So the sort of bottom line is that corporations essentially look to me like pre-internet ways of organizing production and allocating capital, and that when you see things like the makers, when you see the emergence of— so I have a million-dollar CNC, right? It's a hobby that kind of got out of control, right? And it's a bunch— and so I thought, oh good, I'm going to have a hobby, it's going to have nothing to do with my day job. Well, the reality is that all the problems related to machining are software problems. They're problems of, you know, how do you have engagement angles that look a lot like rendering problems. It's like the difference between carving something out of aluminum and drawing something in 3-space are astonishingly similar. And the way that you keep this thing busy is you standardize a bunch of stuff and create a network interface so jobs can flow through this thing. And you end up looking— you end up making this hobby of mine ends up looking a lot like Counter-Strike in in terms of the sets of decisions that you would make to improve it. And I think there's this temporary intermediate state where people who are machine tool builders actually stop selling their machine tools and end up selling machine capacity to a much wider audience. So machine tools are really cool. And if you're adhering to a set of standards, right, if there's a set of interfaces just like TCP/IP provides a standard internetworking protocol, It turns out that you can actually get a lot more stuff through one of these devices. So my hobby annoyingly looked a lot like my job and sort of, you know, let me use it in presentations like this, but sort of took a little bit of the, I'm going to do something completely different, out of it. And I actually think that this—

Yeah.

notion about machining, which is sort of the most real-world thing I could think of. With 3D printing, now we've gone all the way to microtransactions. It's like, oh, it's even more like video games now, right? Where on demand you're building custom items for an audience on a worldwide basis, right? It's like there already are companies that don't even bother to sell you the hardware. They sell you the 3D models that you send to your 3D printer and then they, you know, the thing you actually get from them is the software that's running on top of that. I think you're going to see a lot more businesses that look like that. I think the pre-internet way of thinking of communicating to users as this sort of broadcast, one-size-fits-all, the way of thinking of products, like right now everybody thinks that we're all going to get the same product. Our children are going to say that is the weirdest thing I've ever heard of. Like every one of my products are unique and different, and that that transition means that the sets of lessons that we're learning today in the video game space are probably going to be true of a much wider range of industries tomorrow. Thank you.

So,