Meta (formerly known as Oculus and as Facebook) just had their yearly conference. I think right now is a good time to take a brief break from worrying about AI timelines, and take a few minutes to notice the shortness of VR timelines. VR isn’t going to determine whether humanity lives or dies, but I think it’s close enough to have immediate-term impact on decisions about where to live, and how to structure organizations.
Concretely: I think we’re 6 months from the crossover point where VR is better than in-person for a narrow subset of meetings, specifically, 1:1 meetings without laptops, where both sides are knowledgeable enthusiasts using the latest headset in combination with a high-end desktop PC. I estimate about 2.5 years until the crossover point where VR is better than in-person for most meetings.
I think most people are wildly underestimating how impactful, and how sudden, this is going to be. I think they’ll continue underestimating how impactful it’s going to be, right up until the second crossover happens.
Basically, everyone tried out Zoom meetings and discovered that they sucked compared to in-person meetings, for reasons that are slightly hard to put your finger on. Some people tried VR meetings, and discovered that they were worse. People then updated towards believing that remote work (including both videochat and VR chat) would not be transformative, but did so without a model of what concretely was going wrong and when/whether it would be fixed. But, zooming in on the technology more closely, I see a fairly short list of known problems that seems to fully explain why people prefer in-person meetings in practice, and I’m watching that list slowly shrink. I currently think every item on the VR-meetings-suck list has line-of-sight to a full solution, and that at that point, VR meetings will be straightforwardly better than in-person ones.
For context: Oculus announced a new headset, the Quest Pro, at their yearly conference yesterday. It ships in two weeks and costs $1500. Consumer sentiment about it is probably going to be negative; I expect a lot of grumbling about the price and the battery life, in particular. But this is going to be misleading. The important thing about it is that Quest Pro is that it contains a first-pass version of all but one item on my why-remote-meetings-suck list. These solutions are going to have problems at first, but the problems are not fundamental.
In particular:
Color passthrough cameras solve the problem where entering VR requires giving up the ability to see the real world, which makes it very costly compared to joining a call. There are some minor visual-quality caveats, and some major software-not-ready caveats about how the worlds mix.
Inward-facing cameras with face and eye tracking. Under research-prototype-demo conditions, it looks like this solves the problem of not being able to see peoples’ facial expressions. Under realistic conditions, what I’ve read suggests that they’ve got framerate problems, which is why I specified “1:1 with gaming PCs” as a setup that will work before other setups do.
Audio latency. I believe Oculus is currently pretty far ahead of Zoom on this one, due to a combination of console-like known hardware, expertise crossover from video-latency being a core requirement of VR, and engineering prioritization.
The one caveat that this doesn’t attempt to address is that the panel resolution isn’t high enough to have small text, which in turn means you can’t bring a normal-sized laptop with you into VR. As far as I can tell this is not a fundamental problem, denser panels do exist elsewhere on the market.
Transformative VR Is Likely Coming Soon
Meta (formerly known as Oculus and as Facebook) just had their yearly conference. I think right now is a good time to take a brief break from worrying about AI timelines, and take a few minutes to notice the shortness of VR timelines. VR isn’t going to determine whether humanity lives or dies, but I think it’s close enough to have immediate-term impact on decisions about where to live, and how to structure organizations.
Concretely: I think we’re 6 months from the crossover point where VR is better than in-person for a narrow subset of meetings, specifically, 1:1 meetings without laptops, where both sides are knowledgeable enthusiasts using the latest headset in combination with a high-end desktop PC. I estimate about 2.5 years until the crossover point where VR is better than in-person for most meetings.
I think most people are wildly underestimating how impactful, and how sudden, this is going to be. I think they’ll continue underestimating how impactful it’s going to be, right up until the second crossover happens.
Basically, everyone tried out Zoom meetings and discovered that they sucked compared to in-person meetings, for reasons that are slightly hard to put your finger on. Some people tried VR meetings, and discovered that they were worse. People then updated towards believing that remote work (including both videochat and VR chat) would not be transformative, but did so without a model of what concretely was going wrong and when/whether it would be fixed. But, zooming in on the technology more closely, I see a fairly short list of known problems that seems to fully explain why people prefer in-person meetings in practice, and I’m watching that list slowly shrink. I currently think every item on the VR-meetings-suck list has line-of-sight to a full solution, and that at that point, VR meetings will be straightforwardly better than in-person ones.
For context: Oculus announced a new headset, the Quest Pro, at their yearly conference yesterday. It ships in two weeks and costs $1500. Consumer sentiment about it is probably going to be negative; I expect a lot of grumbling about the price and the battery life, in particular. But this is going to be misleading. The important thing about it is that Quest Pro is that it contains a first-pass version of all but one item on my why-remote-meetings-suck list. These solutions are going to have problems at first, but the problems are not fundamental.
In particular:
Color passthrough cameras solve the problem where entering VR requires giving up the ability to see the real world, which makes it very costly compared to joining a call. There are some minor visual-quality caveats, and some major software-not-ready caveats about how the worlds mix.
Inward-facing cameras with face and eye tracking. Under research-prototype-demo conditions, it looks like this solves the problem of not being able to see peoples’ facial expressions. Under realistic conditions, what I’ve read suggests that they’ve got framerate problems, which is why I specified “1:1 with gaming PCs” as a setup that will work before other setups do.
Audio latency. I believe Oculus is currently pretty far ahead of Zoom on this one, due to a combination of console-like known hardware, expertise crossover from video-latency being a core requirement of VR, and engineering prioritization.
The one caveat that this doesn’t attempt to address is that the panel resolution isn’t high enough to have small text, which in turn means you can’t bring a normal-sized laptop with you into VR. As far as I can tell this is not a fundamental problem, denser panels do exist elsewhere on the market.