If you’ve been paying attention to the headlines today, you might have seen a recent TechCrunch report where Marc Andreessen makes a crazy statement:
I think VR is going to be about 1,000 times bigger. In the Valley right now, this is a very contrarian view. The general theme that you hear is that AR will be bigger than VR, and obviously it should be. If you can do things overlaid over the real world, that should be inherently more interesting than having to construct a synthetic world.
I just think that’s only true for people who live in a really interesting place in the real world. But only something like .1 percent and 1 percent of people on Earth live in a place where they wake up every morning and think, Wow, there are so many interesting things to see. So for everyone who doesn’t already live on a college campus or in Silicon Valley or in a major other city, the new environments we’re going to be able to create in VR will inherently be much more interesting. And there will be a lot more of them.”
My initial reaction was “What’s he, the web browser guy, drinking?!” I thought there was NO WAY he understood the technology or what it entails, and it bothered me that such a smart and respected guy would make and record such a stupid proclamation. So I came to my new little blog here to write down my counter arguments and basically call him Rian Johnson. But unlike how I feel about the guy who can’t see a plot hole if he were standing in it, I’ve thought through where Marc may have been coming from and can kinda, sorta see his point.
Mind you, I disagree with him, but only on two points:
VR won’t be 1000x bigger
People who live in uninteresting places stand to gain the most from AR and NOT VR.
But before I make my case, let me take a second to explain how I define what these two letter acronyms mean, because I’m pretty sure a lot of people get them and their potential confused (possibly the co-founder of Andreessen Horowitz, too).
Briefly; Augmented Reality is a display of overlayed static or moving images that are aligned with real objects within the field of view of a viewer. The hardware requires a mono or stereo display, a series of cameras, and sensors to capture the native scene and determine the angle of the display to ensure alignment and orientation of the overlayed images to the viewers’ field of view (meaning, their eyeballs). This can happen via reprojection of the entire scene or, more ideally, via controlling the transparency and occluding the view of CGI objects into the field of view.
Virtual Reality is a display of 3D rendered images into a digital ‘virtual’ space, where the orientation of the 3D objects are automatically and synchronously manipulated in response to angle changes of the headworn display device. It requires the same types of sensors as an AR device for orientation, but it does NOT require alignment information because all of the projected imagery is computer generated.
The two images above really can help explain the differences the best. With AR, you can see through your device as if they were a pair of glasses. With VR, you can only see what your device projects.
Now, what does this mean from a user experience point of view? One could envision an AR system that makes a Toyota Corolla look like a Ford Mustang GT. The cameras on the AR hardware see the surface of the vehicle, and the computer (say, like those on the Magic Leap) can through machine vision identify the steering wheel, the shifter, the windshield, and so on. The AR system can then visually block your vision of these specific objects’ surfaces with reprojected CGI into their place. The Toyota’s plastic steering wheel can take the appearance of leather, and your car’s flat fenders can take on the sexy curves of the Mustang GT. While you may not feel the supple textures of the real Mustang’s interior in your hand, the visual illusion can be quite convincing.
For a VR system, you’ll be wearing a headset looking at a projected CGI or 360 captured video of a Mustang GT. Surely, you can sit in your car and hold a steering wheel in your hands, you won’t be able to engage with it to the same extent, and you certainly won’t be able to see through the windshield and drive! Even with the most portable envisioned systems, the experience is entirely designed around the digital world created by the VR scene’s author.
So, when it comes to hardware sales, Marc cannot be correct. AR hardware will absolutely outsell VR hardware because AR headsets can already do VR. All you need to do is 100% occlude the field of view (instead of just the steering wheel) and voila, you have a VR experience.
That said, Andreessen Horowitz’s motto is:
so it’s pretty clear to me that he means software development for VR will be more popular than it will be for AR. Ummm, Maybe,
I could concede that VR CONTENT will be more popular than AR CONTENT because you can do a ton more things with a blank canvas than you can repainting the world we live in. Machine Vision is computationally intensive (requiring power, giving heat) and therefore not yet portable. Nonetheless, I believe he is dramatically over-estimating when giving his 1000x figure because I think his statement about people not living in “interesting places” is TOTALLY WRONG.
As I describe in my thought exercise above, AR enables customization that VR simply cannot do. It can make your existing boring place MORE interesting. Let’s take these ideas to the next level: you can skin your Corolla into a Mustang, but I’m going to make my Subaru look like a Millenium Falcon. Instead of parking it in front of my rickety old apartment building, I’m going to dock at the Regula 1* space station. And when we hang out, you (and my future Patreon-AR subscribers) will see my custom tattoo and crazy hair color that I feel like digitally wearing that day.
Does this sound crazy? Maybe, but it isn’t. Custom skinning of 3D objects has existed in many contexts for years, with real, profitable economics back them up. For just a few easy examples:
All of these marketplaces sell real digital properties for real dollars for 3D rendered environments, including those that can be projected in VR.
My bet is that these marketplaces and exchanges are what Marc Andreessen is thinking of when he starts talking numbers like “VR is 1000x of AR”. That said, I think he is making a bad assumption thinking that there is a massive customer base that will want fully replace their ‘uninteresting’ existence in a Sensorama, Johnny Mnemonic, Tron, Matrix, ReadyPlayerOne kind of way.
At least god I hope not.
My opinion is that developers will want to shift away from pure VR 3D creations into the more life-enhancing (or interesting-making, in Andreessen-speak) once the AR hardware options start getting better. In fact, the big guys like Unity3D, Steam, and UnReal have already moved in that direction, and they will take their marketplaces with them. Hopefully Marc will clarify what exactly he meant in some future tweet or other response!