Masters of Disruption: How the Gamer Generation Built the Future [18]
In part two of my conversation with John Carmack: the Holodeck, emails from Zuck, and the S curve of success.
This post is part of a longform project I’m serializing exclusively in my newsletter, Disruptor. It’s a follow-up to my first book, Masters of Doom: How Two Guys Built an Empire and Transformed Pop Culture, and it’s called Masters of Disruption: How the Gamer Generation Built the Future. To follow along, please subscribe to Disruptor and spread the word. Thanks!
David Kushner: At this point in your career, you could do whatever you want. Why did you decide to devote your time to working on Artificial General Intelligence?
John Carmack: A lot of it does come down to sort of a leverage argument. This is kind of a weird situation where, yeah, I do have my pick of problems to work on, and a lot of things that I would find interesting, I can go and do. But I feel still some degree of desire to have an impact on things. When I feel like I'm no longer at the top of my game, my actual powers are waning to some degree, I can retire and write retro games or something. There's all these things in older time computers that I'm attracted to. I'd like to go work on simpler systems that can be optimized. You can feel the whole breadth of them, rather than the monstrously complex things of today.
And you know what? That would be fun. But I still think largely, I'm a world-class engineer and developer. I can make a difference in some things. So maybe I should be working on these harder, global-scale challenges. That's been, I think, one of the powerful ways about how I look at the world. I can look at almost anything and see the opportunity for improving it. But then you get to the quantification of how much value you can provide. I could be working on some tiny little thing and make some hundreds of people really happy by improving it, or I can be working at a global scale.
Like, the games affect millions and millions of people. Gaming, in the larger sense, is something most of the world is impacted by. Virtual reality, in the larger scheme of things, as we start making distance less relevant for more things than just the chat and video communications, that's a powerful thing. I think that will have an increasingly large effect on the world: this ability to defy distance and let people from all over the world participate in more ways than they do now. That’s one of the very largest possible factors for growth of human value in the future. It's such a shame that so many people are just in the wrong place at the wrong time and can't contribute to the best of their ability.
This gets tactical at just how horrible the hiring process is at companies. And I'm terrible at it, trying to make a snap judgment on people about how valuable they might be. Facebook gets so much shit for their positions on things, but connecting the world really is a virtuous thing. There is trillions and trillions of dollars of value just to be had by getting the right people talking to the other right people, and then interacting with them in different ways.
“With VR, you can have any experience you want, and bring anyone together in any environment they want. Clearly that's enormously valuable.”
David Kushner: When you talk about value, it makes me think of a quote from Masters of Doom that gets referenced quite a bit. You and I were talking about the Holodeck, and you spoke about the moral imperative of creating virtual reality. Do you still feel that moral imperative now with VR?
John Carmack: It’s kind of funny, the longevity of that quote. Because nowadays, everybody misses the fact that that was a Real Genius movie quote at the time. But yeah, I do feel compelled to pull the future into the present for things that I consider to be very valuable. I am analytic about quantifying the value of what I'm doing and what I'm trying to accomplish or what other people should be focusing on. Because I think that I had a good gut sense for a long time, turning a heuristic into something more procedural.
It's like when I look at things like Oculus, about what should we be working on, it's amazing how many people can just make a very bad value judgment about stuff, like how this feature is going to be super critical and important. And you break it all down, and it's like, “no, that's only going to be helping 1% of our user base. It's not going to be something that adds a lot of value.”
Very much over the last decade, I've become this champion of user value, about look at everything through the lens about how many users are going to get how much value out of what are we doing. Multiply those together, divide by the effort required to accomplish that, and add in your error bars, and then make your decision on maximizing that value. So much of machine learning is about a maximization function. You have a loss function, you back-propagate error, and it all comes down to making this number move.
A lot of people will decry that as too reductionist and simplistic, but it is the most powerful way of looking at so many things. You can say everything is equally important, but then you'll make a terrible decision. You need to figure out what your weights are on this. It's like, ‘no, this is much more important than this’ because of whatever process lead you to that decision. So, yeah, the idea of the moral imperative was sort of a pre-quantitative view of mine. It was about saying “yes, the whole world will benefit if we can virtualize it like this.” If you can have that sense of the Star Trek Holodeck, where you can have any experience you want, and bring anyone together in any environment that they want, clearly that's enormously valuable.
People have lifelong desires to go and see certain things or do certain things. And the bottom line is in a world of eight, nine, 10 billion people, we can't give everything physically to everyone. But there's a very real chance that we can virtually give all of those things to everyone. While of course people will say the virtual is no substitute for reality, that's kind of the wrong way to look at it. The virtual will not usually be as good as reality, but for people that have no chance of getting the reality of an experience, a virtual experience, even if it's only 10, 20, 30% the veracity of the real thing, that's still enormously valuable. And many of these things might get up to 80, 90% of the quality of reality.
So that is a world-scale level of value: being able to bring the set of experiences that we can simulate virtually to, eventually, almost the entire world. The cost curves on these things are actually very, very good. The idea of a VR headset can be cheaper than a big screen TV. There is no reason that this doesn't follow the same price curves that a cell phone does, where eventually, you get down to some $100 and then some $50 headsets. Nowadays you can buy a $25, $50 Indian or Chinese smartphone that works better than any of the computers that the best people in the world had decades ago. And I think virtual reality can follow that trajectory.
“Zuck takes my emails. It’s interesting being in the vicinity of power on that scale.”
David Kushner: To me, it still feels like a big ask to get anyone to wear something on their face if they don’t have to. How far away are we from goggles going completely away?
John Carmack: It’s been interesting, you’ve got the whole S curve of product acceptance. And absolutely VR did not take off as fast as we hoped when Oculus was founded in 2013. We would have thought that seven, eight years on, it would be further along than it is now. But we actually did see an acceleration in the last couple years with the Oculus Quest finally picking up steam and getting mainstream. It was very much to Facebook's credit that they stuck with the product, continued investing in it for a lot of it.
We have some product directions that are very much like, “oh, people don't want the big shoebox on their face.” But then how much is it worth? Is it worth a $100 more to shrink it down to pancake lens sizes? Is it worth $500 more for some never-before-seen optics to get it down to a kind eyeglasses level? I don't claim to have a real answer on that. The market will decide, and it might surprise people.
I don't think that the bulkiness of the headsets is a terminal factor. You could still find a billion people willing to do that if we provide sufficient value for them inside the experience - as long as we can make it better inside the headset than what it is outside. A lot of people in the technology press are not self-aware of the bubble they live in terms of that. If you're rich and you can go half-court, and you can go behind backstage, and you can do all of these experiences that are pretty much for the elite, then yeah, VR is not going to look all that impressive to you. Because you can turn the world into whatever you want it to be.
I think back to when I was helping a developer working on something. They were having a problem with a 360 photo. They were in, like, a dorm room in the Philippines, and their test case was their 360 photo of their place where they lived and worked. And I'm like, “yeah, VR can be a hell of a lot better than what this guy's situation is in reality there. We can give him his beachfront view and big screen TVs and all of that that can be a genuine improvement.” So, yeah, I still feel pretty good about VR.
Working with the large company has been a real experience for me. All the politics and large team stuff was grimmer than -
David Kushner: But you never left Dallas.
John Carmack: Yeah, I did refuse to move to California. Maybe I could have been more effective. Less things would have happened out of my sight. I could have levered on the political decisions. But I was still trying to just make all my marks on the technical matters, and do the right thing.
David Kushner: After leaving your full-time post at Facebook, you described yourself as a “Victorian Gentleman Scientist.” Are you happier that you've branched off to do your own thing?
John Carmack: I still nominally work one day a week for Oculus and it winds up being a bit more than that as I do touch base fairly often. Zuck takes my emails. It’s interesting being in the vicinity of power on that scale. Zuck listens to me. I send him a couple emails a year when it's something that I think is really important to us, that's worth his attention.
Whatever I do, I want to do a good job at it. I still write some code to do experiments, but I haven't landed anything into production in a long time. I'm going through a transition with about accepting that. Almost all of my output now is text intended for humans rather than text intended for computers. It's all posts, emails, chats. I have documents. I have a few meetings that I attend and it is a different way of working than it's been in the past. But interestingly, if anything, I may have even more influence now than I used to. I hear that, that sometimes happens. People leave and become consultants and they're actually listened to more than when they were employed.
Facebook built this product called Workplace that they sell to other companies, but they use it internally. It's basically Facebook groups for internal stuff. Because I've been remote, because I wasn't in the smoke-filled rooms where decisions were getting made, behind the scenes things, all my communication largely was through these persistent posts. Now anytime something new comes up, I’m like, “Well, here's what I said two years ago and four years ago, and they were right then.” That pile of evidence has been increasing, so it's given me a good track record.
So, I'm pretty happy with the way things are going right now. I like having contact with consumers and everything that Oculus gives me. We have millions of people using VR all the time now. I'm open about it. I joke that I'm customer support’s last resort on Twitter. If something is fouling up and people can't get a real answer out of Oculus, I’m okay. You can hit me up on Twitter and I will try to get to the bottom of it.
I’m closing in on a million followers on Twitter, but I have a remarkably good Twitter experience. I know lots of people, even with far, far smaller audiences, just have a soul-destroying experience on Twitter. There's probably a whole lot of factors that go into it, but I like to think my earnestness comes through. Even if people disagree with something that I might say, I am coming from a place of earnestly trying to do the right thing for the things that I'm doing.
