John Carmack Oculus Keynote 2017

The main keynote gets obsessed over by a big team of professionals for weeks but here I get to be a lot rougher and more opinionated and even a little bit critical. I actually wrote up all of my notes on the plane flight over here. I know it drives a lot of production people crazy but I think that's part of the charm of this. We are four years on. I get reminded of this quite a bit when I started at Oculus as a new employee I did what a good new employee does and I adopted the company coding style which was Michael Antonovs which included headers at the tops of files with creator and date so now I still run by all this stuff and see author John Carmack created four years ago.

It's been a long time that we've been been working on this and VR hasn't taken over the world yet but it is showing up in a lot of surprising places. Just last week I was at the grocery store and I looked over and there was a gear VR on a poster for some immersive wine tasting experience or something which is very strange thing that four years ago we probably would not have predicted that we'd be seeing it in those sorts of circumstances. However, there still is this sense that you know maybe this year we are out of the heroic age of VR about shipping commercial you know consumer commercial VR out to people where the people that signed up for this bold new journey to go out and build these products that you know that hadn't been out before.

I am and some people expected it to be magic right off the bat that it would go out and they would be a strike a bolt of lightning and there would be this huge amount of success but the truth is we're product competing against hundreds of other very good products and very finely tuned products and it's gonna take a whole lot of work to to kind of reach that level of success but I do think I'm optimistic in so many ways and much of what makes me optimistic about VR now is that it actually feels like all of the pieces, all of the ingredients that we need are already really here they're just not stir cooked and seasoned exactly how they need to be when we look at the fact that we have over two thousand applications I in our store and there are no there's a broad range of quality but there are some very good things that almost anybody should be able to find something good and interesting in the store and I look at and say Facebook is largely an advertising company we should be good at connecting people with the content and that they're going to enjoy but that's something that we're not doing a great job at yet but I think that there's huge room for us to improve that there's a lot more that we can do, and the fact that 360 media has probably exploded beyond anything I would have expected four years ago a lot of Facebook having 360 in the feed both eye photos and videos is huge the fact that we have thousands and thousands of high production quality 360 videos and millions and millions of 360 photos I am 360 photos are by far the most the most popular form of user-generated content for virtual reality or for immersive media and terminology wise I have tried to settle on using immersive media more often.

Whether it's 360 or 180 or stereo whether it's got 60 frames per second or narrower fields of view and I lump all of this together and just try to say you know immersive media is something that's better than looking at media on a flat screen and I think that there's enormous value that's already there that we're again just doing a poor job of exposing to people the fact that I make this comment a lot if you look into oculus video and you say I want to watch 360 videos you could be forgiven for thinking that there are only 80 videos as you page through sort of the top featured things not that there are thousands and thousands of them available just kind of waiting to be discovered so these are some of the things that I am you know that oculus should be doing a better job at and in our partners we do have some world-class brands like Netflix and minecraft that everybody in the world knows that that are on our platform and the web experience that we're providing is is getting pretty good and in many ways the web is the backstop for all of the apps that you don't have where there's a webpage or an application for practically anything that you could want to do.

It feels like the pieces are there and in many ways Samsung has been doing has done a better job at being user-focused and I feel oculus has in many ways when their applications first launched the original I am Samsung VR or milk VR as it was at the time or the Samsung Browser I was fairly critical of the initial applications I would look at them and say you know they should be doing this in a time warp layer this should be filtered better this wrong option is on here but I am you know they went they released version after version and we look over some time number of versions of later and Samsung had the number one application in VR for a while and I used that to berate a bunch of people.

We need to step up our game you know Samsung is out there listening to users doing what they want providing the features that that they actually find valuable while I think oculus has in many cases focused on we want to build the platform we want to build the infrastructure for some of these things but it's easy to kind of fall into a trap there and not judge yourself on value and impact to the users when you're saying well I'm building infrastructure that's going rely on somebody else to extract this value here so I do you know push oculus very much to to try and say well let's actually do things for our users not necessarily for our platform you know often I'm like in the voice of Tron we should fight for the users but many of our applications in the store got 2,000 plus applications but so many of them are stuck on sort of that first version like the original Samsung applications and they don't really get to the point of being able to be iterated on.

In fact some developers go through series of applications that if they build a VR application do something interesting then they build another VR application and do something interesting but they're all only taken to that first maybe 80% and or sometimes like in game development you talk about how the second 90% is always the hardest and not many of our applications are going that extra mile where all the really important magic and user value happens and there's a school of thought where that could be the you know the intelligent direction where if you believe that you need a killer app that it's something nobody's ever seen before then it can be completely rational to be sort of point sampling the application space trying lots of different things or throwing things at the wall and seeing what sticks so if your thought is that nothing that anybody's doing is even on the right path to being the winning solution then you just try shooting around randomly in different places but I'd like to present an alternate view point there where there are a whole lot of nonlinear value thresholds where it's not a matter of adding a little bit more work here always results in a little bit more user value sometimes there are a very nonlinear or very sharp cutoff points where if you look at many of the the things about winning and succeeding in lots of different areas and even the the revenue streams at our stores it hurts to look at these sometimes where you have a handful of applications that go make you know million dollars here and are very successful.

You have hundreds of applications where you've had developers put developer years of effort into something and wind up with five hundred dollars of revenue from it and that's really sad but it's unfortunately that's reality and almost all of these consumer-facing industries where sometimes the difference between something that does almost nothing and something that is really huge are relatively minor things and things like Facebook and Instagram are examples where they had competitors they had competitors that were doing very similar things and in many cases the differences between what became the multibillion-dollar juggernauts and the companies that went out of business were doing you know dozens of small things somewhat better so I would like to push everyone towards trying to work harder on existing applications and there's another argument for that and that because each new generation of Samsung devices especially we bring in huge new tree of users and that's an opportunity that an older application cleaned-up improved has an opportunity to make brand new fresh impressions on people and I'd also say that again in terms of application discovery and presentation to users we're not doing the best possible job where an application could go out vanish without a trace beneath the surface not necessarily because it maybe didn't have a kernel of something something right in it but because it never got exposed to the user's right eye you know it needs the different guerilla marketing tactics and all the things that people do to be successful on other platforms also have to be applied to the platforms in VR.

If you have a brilliant idea, if you think you've got that lightning in a bottle the killer app you know by all means go ahead and work on it but otherwise you can always spend time improving the existing applications and those are good muscles to exercise so even if it turns out that that wasn't the magic app I am going through the the disciplined work of making it as good as it can possibly be is what's going to need to be applied when you eventually do get the magical application and when I complain kind of bemoan internally about how much better our VR applications could be about the things that they could be doing better why the graphics should be better why the responsiveness should be better why the load time should be better you know I have people that I you know that could know and agree but they'll say but all the devs they don't have your experience you know to which I'd reply well go get it grow stronger you know I'll help you know there's a lot of people that do have this experience you know you can find old-timers that you know that know how to do this type of optimization and even the new people that know how to do this type of user testing or even marketing these are skills that you know that are going to be necessary for success and I some people would like to think that this type of disciplined coding and design within very tight constraints that maybe if it's like in the old days you could just wait a few years and you're lazy design whatever just works on the new computers that was the way of PC development for a very long time but we're in a situation now with with mobile being important.

The end of Moore's law kind of drawing nigh that those types of skills are absolutely going to remain important for the foreseeable future this future device that we all imagine where we have AR sunglasses that billions of people are wearing that's probably going to have at least as tight of a power budget and design constraints as you have on gear VR. I mean it's going to have lighter weights smaller battery, less thermal mass it'll be expected to run for 20 hours instead of three or four I am you are not going to have some magical thing that comes in between now and then that's going to just solve all of our problems and this distinction between I raised the hypothetical question to a lot of people if you had a choice between getting magic hardware and magic software you know which would you choose if I could say well I could take I you know gear VR are stand-alone and automatically give it 4k displays give it twice then the power performance that it's got right now magical optics that are perfectly crisp across everything or you could get software that actually does things to the limits of what the hardware we have today can do and I have always taken the side that right now I would rather have magic software over magic hardware I think that we have enough extra value that we can put into the software on the platforms we have today to make those kind of critical differences and I think we could take what would be thousands of dollars of hardware magically put it down to this low level and the current software that we have still wouldn't be incredibly awesomely engaging just because some of the the hardware magically got a lot better.