John Carmack: VR Developers Should Stop Chasing After More Ultrarealistic Graphics Because "We are running out of Moore's law"

Here's a fun new Wired video in which Oculus CTO John Carmack explains the concept of virtual reality with ever greater levels of complexity. It gets really interesting about 9:35 minutes in, as he's talking on an expert level with Aaron Koblin, CTO of the VR company Within. Surprisingly, Carmack starts advocating for VR experiences with abstract or even cartoonish graphics, as opposed to high-end ultrarealistic graphics he became famous for, primarily through his Doom games. Instead, Carmark argues that the aesthetics are more important than high fidelity per se.

As he puts it: "Some of [my favorite VR experiences] are clearly very synthetic worlds where it's nothing but these cartoony, flat-shaded things with lighting but they look and they feel good." At first he describes this as a practical, market-driven approach for developers, but very quickly after that, makes an interesting point about the coming limits of graphics simulation:

"We are running out of Moore's Law. Maybe we'll see some wonderful breakthrough in quantum structures or whatever, but if we just wind up following the path that we're on, we're going to get double or quadruple [the graphics quality], but we're not going to get 50x more powerful than where we are right now. We'll run into atomic limits on our fabrication."

Which sounds exactly right to me, though that's notably different from something Carmack once famously said in 2004: "Because of the nature of Moore's law, anything that an extremely clever graphics programmer can do at one point can be replicated by a merely competent programmer some number of years later." Then again, the limitations of Moore's law weren't as evident 13 years ago.

Comments

[Joey1058]

I'm of the opinion that the more realism that one is immersed in, you start to consciously seek out the imperfections. I have blogged that the Japanese are totally comfortable relating to anime styled creations, be it mascots or avatars. Just human enough without the creepy uncanny valley not quite right detailing. SL avatars are good examples as well. I'm comfortable because it doesn't make an overt effort to do things that the tech can't do. While I appreciate the efforts, I don't see a personal need to immerse myself in a crazy level of detail that I can simply step outside to experience without tech.

[Medhue]

Personally, I'm not really in agreement with John. It's not 1 system that runs a game. It is many completely different systems. Each 1 of those systems could and will see advancements in the future. Now, will it all as a whole continue to advance at the rates we've seen in the past? Likely not, but does that even matter? In many ways, when an improvement comes along, it is sometimes a tiny improvement, and sometimes the next big thing is a HUGE improvement, in some way or another, which is difficult to even calculate, as a measurement. With so many different systems, it is also practically impossible to understand what new improvements in each of those system will evolve, and why.

It sounds to me, like John's goal is to get more developers to drop the realism for what he and his people regard as better VR experiences. In a way, it almost seems as if they are trying to set up a scapegoat, if VR fails.

As a creator of 3D objects, my life's goal is not to create blocky, low poly items. There is not a whole lot exciting about creating such things.