From a development point of view, AR is taxonomically a rank higher than VR. A fully 3D AR HMD with hand tracking is VR if it blacks out the screen entirely and shows you zero transparent pixels. VR is a subset of AR.
VR is just that magical AR rig without the ability to draw transparent pixels. Everything else is the same, from a user/use perspective.
Likewise, jacking into the Matrix is one step above AR taxonomically, and AR is a subset of the Matrix. A high end AR rig would be the matrix if only it had full-body mortality haptics.
But we don't have any of that tech today.
What we have is now is junk
The Matrix from the movies is crappy tech. I'm sure they COULD rewrite the Matrix to generate an AR experience (just let some meatspace light signals in!), but they don't because you need a copper spike in the base of your skull and be strapped down to a chair. You'd just be looking at ceilings all the time. The Matrix, as we know it, is not the pinnacle of that tech. I'm sure someone (Samsung?) is working on a competitor to the Matrix that is wireless and lets you walk around meatspace while using it, and they'll finally figure out that bug where you can't disconnect mid-session. One day that will be the #1 matrix product, but until that launches we still need that copper spike.
Likewise, AR of today is crappy. Products like the Microsoft HoloLens is very large, clumsy, and unwieldly, but has the market distinction of being one of the only 3D AR HMDs out there. Pokemon Go could be (poorly) argued to be an AR app, and it's probably the worst of the examples. iPhone ARKit and Android ARCore solutions are incremental, but they are a far cry from fully immersive technologies, and they don't even have hand tracking. AR has a long way to go, and some say the math required to properly solve visual occlusion (hiding a 3D character half behind a real-life object while properly occluding the relevant bits from sight) is too hard to solve on todays CPUs, nevermind on mobile CPUs. Heck, anything not mounted to a head has significant ergonomic problems when you are talking about long-form content and not quick toys or fun tools.
And that's not to say VR is perfect, either. We've figured out how to do full-body-tracking but it's far from a market norm, wireless hasn't fully taken flight yet even though it's obvious we all want it, and the market is flooded with really crappy mobile-VR toys that ruin the experience for everyone forever. Nevermind the whole "it's stupidly expensive" problem.
We are currently holding the tool that will build the future
If you want to be a designer for the Matrix (either the current cheap one or the fancy new Samsung Wireless AR Matrix) and you want to make fully immersive content, what should you be practicing on today?
Clearly, hands-down, you should be working on a high-end desktop VR rig. It's not even a serious question in my mind. The software design lessons from current-gen AR won't translate to the Matrix because of the intensity of the obfuscation layer - nobody is going to stop in the middle of the Matrix and squeeze their fingers mid-air HoloLens style to pick something up. They aren't going to swipe or tap. They are just going to bend over and pick the damn thing up. That's something you can only do in VR right now. Granted, there is still the obfuscation of a controller in your hands, but new developments such as the Valve Knuckles EV2 controllers is slowly removing this obfuscation.
We are learning so much from VR right now;
- Basic design patterns
- Lessons about haptic responses
- How to handle "virtual swords" that never physically clang against anything
- User interfaces/menus/huds (or the lack thereof)
- Area and space design/architecture
- Spatial design density
- Diegetics and skeuomorphics
- Human locomotion and how that ruins your best laid plans
- The causes of simulator sickness and how to remove this threat
Every day we are learning new methods of design abstraction in VR that will directly translate into the Matrix being a viable product on launch day, instead of launching with a bunch of shitty software and slowly making it great over the following years.
One day, hopefully before the Matrix launches, we'll get a Real AR rig. Something more than a "reality overlay" mode on a phone, which is fun, but not the same product. Once this new AR Rig comes out, every single VR developer will be able to seamlessly switch to the new platform and start making new stuff (or porting their old stuff over).
A good AR product will have a ton of masterfully designed software from VR developers available on day one.
Present-day AR developers will more work to do, comparatively, because they'll have fewer of those design lessons in their toolbelts.
The argument for present-day AR
That's not to say that AR development is worthless. Seeing first-hand how hard it is to place 3D objects into a real world environment can be daunting, and something that all developers will have to contend with in the future. There are definite lessons to be learned there, and diversifying your experience is always a good idea.
Financially, AR is currently riding the "hype" curve, and VR is climbing out of the "trough of disallusionment". Present-day AR tech is kinda crappy (or just plain different) compared to present-day VR, but there are a LOT more consumers that could boot it up in an instant on their phones, and it is hot and new and investors love the crap out of it. VR is kinda hard to make a buck right now.
AR might be the path you need to take if you want to make money. Just know that it doesn't necessarily lead as nicely into the next-tier-human-experience as VR currently does.
There's also an argument for staying within the phone ecosystem for a variety of strategic reasons. The first MegaGood AR rig will need to be mobile, which means it needs to have content optimized in a way that only mobile devs are good at. All those tips and tricks I know to keep VR running quickly on a desktop are just a sliver of the tools present-day AR devs have, and investing in those tools might be necessary.
AR is also fun and new. Maybe you just want to try out something different, and don't really care about the Matrix? Cool. That's not why I'm in the game, but your goals are valid too!
Also there is that one person on reddit that's always like "well actually AR is better than VR" and you'll make them happy by going that route.
This is one of a handful of VR advice pages I've written. Check out the index for more.