Invest Like the Best with Patrick O'Shaughnessy
· Patrick O'Shaughnessy
[00:00:00] S0: Has there been a most surprising thing about building now 2 things and more that are very big?
· Palmer Luckey
[00:00:00] S1: The biggest surprise was that Oculus became big, when I set out to build Oculus. I didn't have any of this perspective. That was big by chance. Obviously, I believe that VR was going to take over the world. I believe that it was going to become the dominant form of computing. But if you would ask me to honestly bet the odds that Oculus was going to be the one that managed to pull it off, that was a tougher bet. I started working on VR as a hobby, and I was able to turn it into a job and able to turn it into a company. That was a crazy bet. And so I guess my other advice to people would be don't look at things like Oculus and think that, that is the optimal way to make an impact in the world. It's not to say, what am I passionate about and what do I like doing? I'm going to turn that into a business and I'm sure it goes well. It's actually to completely separate the impact you have from what you want to be doing. Put it another way, I don't want to be doing Anduril, not really. I would rather be making virtual reality headsets. I'd rather be making video games and toys, fast cars, spaceships. That's what I want to be doing. I'm doing what I'm doing with Anduril because I think that it's important, and I think that it's going to be more impactful. When I was starting Anduril I specifically said, "I want to do something big. I want to do something impactful." And I don't know if you've ever heard me talk about this, but there are 3 things I was considering One was solving obesity, which is the biggest killer in growing that I think we can feasibly solve. The other is private prison reform. I think there's ways we could eliminate the private prison system by out competing them with better incentives and align those incentives towards…
· Patrick O'Shaughnessy, Palmer Luckey
[00:00:00] S0: Recidivism? [00:00:00] S1: Recidivism, I am sorry, I was just drawing on myself. And then the other was trying to fix national security. And so this is the one I ended up in. But none of those 3 things sound like fun things to work on. But they are important things. [00:00:00] S0: When you approach those problems, I have heard you talk about those other two, which is totally fascinating how you approach things. Maybe even before I ask this question, I'm just curious what your method of invention is. So you find an interesting problem either that you want to work in this case, don't want to work on necessarily. Is your method iterative? Is it more theoretical? Describe the way that you start to invent in a field when you approach something for the first time.
· Palmer Luckey
[00:00:00] S1: Well, it depends on if it's a field that I know a lot about or don't know a lot about. If you know a lot about something, it's easier to get right into the iterative side of things and know that you're probably on a pretty reasonable path. In that case, iteration is a valuable tool to move very quickly, find out what works, find out what doesn't and then continuously make it better. The risk with going with a strongly iterative approach in areas that you maybe don't understand, and you might even think you do, but let's say you truly don't is that there might be much better approaches that you should have started iterating on. Or that you should have examined before you committed to one particular path. I talked about this earlier, but it's really about going back to the future. I love to go and see what everyone else who solved this problem thinks about it, not in the recent times. I don't want to know what my competition looks like because when I started Oculus, I wasn't looking at what existing companies were doing in VR because clearly, they were all doing it wrong. Whatever they were doing was not working. I was not going to look around at the handful of VR companies that existed in that time and learn anything except how to fail. So I wanted to look into the past, what were people thinking when they were thinking bigger, when they were willing to look at wackier paths. When they were willing to consider things that have been eliminated often because technologies just weren't ready. There's a lot of technologies that have been discarded because they weren't practical at the time, and nobody ever revisited them and said, "Hey, I actually think the time has come." A good example is with the Rift. Doing real-time distortion correction is not a new idea. It existed in the 1980s and 1990s in the virtual reality community. They have been discarded even by NASA. There's a fascinating NASA paper where they talk about doing real-time geometry distortion correction on a virtual reality headset that made the optics lower distortion and allow them to, therefore, use wider field of view. Lighter weight optics than would have otherwise been required to have an optically perfect image. And the conclusion was, yes, this is a really good way to save money and to save weight, but it's too computationally expensive.
· Palmer Luckey, Patrick O'Shaughnessy
We're using most of our processing power to warp the image in real time rather than render this wire frame image. And so this is not a good approach. You should just do it optically, and then you don't have to have a more expensive computer. But back then, compute was the expensive part and the optic transform used up a lot of your compute. Nobody reexamined this idea even as computers got better. Nobody went back until me and said, "Wait a second, you can do real-time transform on a modern graphics card for like 1% or 2% of your render horsepower." And also your graphics card doesn't cost $100,000 anymore. It costs a few hundred dollars. And so if you're worried about that 1% or 2% impact, just buy a graphics card that costs a few dollars more, so you can save hundreds or thousands of dollars on the VR headset itself by using optics that have geometric distortion or in chromatic aberration, for example. That was an idea that had been discarded and nobody ever came back to it. And most of the things that made the Rift successful were ideas like that, there's a few others where I was just going back to the future and realizing, "Wait a sec, these ideas, they were actually pretty good. They were just a little too early." [00:00:00] S0: It makes me think of the Nintendo lateral thinking with whether technologies idea, you can always recombine stuff.
· Palmer Luckey
[00:00:00] S1: We talked about that a lot at Oculus. It's true. Sometimes it's also combining new technology with old technology. I actually had a background of thinking this way before I started Oculus. I don't know if you know, but I ran an Internet game console modification forum called ModRetro. The whole purpose of ModRetro was to combine modern technology with retro game consoles. So the primary focus was taking old game consoles, cutting them up, removing a lot of the old legacy components that can be replaced with better modern versions like the power regulation side, the audio management side and then building them into handheld portable devices that were much smaller than the original consoles, but provided full functionality. So like, for example, building portable Nintendo 64 that could fit in a cargo pocket. I started that forum I think when it was 14 or 15. And actually, if you look at my Twitter Bio, even today, I listed as the founder of ModRetro, Oculus and Anduril working on projects in that forum was actually what gave me this mindset originally, where we were very much looking at what these old consoles were? How they worked, what they had done and then speculating what modern technology would they be using today to solve the problem if it had been accessible to them. It's usually quite clear. And so taking old ideas and combining them with the latest implementations of modern components, Nintendo has done that very successfully. They waited until motion tracking got just good enough, and we're able to use pretty limited compute and pretty basic ideas to do things that, of course, have been talked about for many years. But Nintendo, they were like right on that edge, right as MEMS-based motion controls and image sensors that were appropriate became possible. They immediately turned that into something that was a lot better than some of the parts.