The Apple Vision Pro is a weird duck. As a piece of technology, it’s extraordinary – the pass through is great, the hand/pinch gesture sensitivity is great, the hardware feels nice and is (mostly) comfortable. The big thing for me is the forward-heaviness of it, which is almost instantly mitigated by just tucking the battery into the rear strap. I get that they wanted strap options, and that they want you to comfortably rest your head while using the headset, but a well-designed counterweight strap would make this a LOT more comfortable with almost no downside. I don’t get why Apple didn’t learn this lesson. It’s not weight that matters in a VR/AR headset – it’s where the center of balance is. If you put the weight directly over the neck, it can be much heavier, and still totally comfortable. The big problem is forward-heavy headsets exert pressure on your cheeks that eventually becomes intolerable. But I can hold that same weight easily on the top of my head for hours. So that’s weird.

 

I’m struggling a bit to figure out what this is for, to be honest. I’m trying it with a keyboard and trackpad to use as a general computing device (writing this post, even) – and for that, it’s nice. But with limited battery life, and as delicate as it is, it can’t compare with a laptop. As a home-use device, it’s also weird – at home there’s usually people around, and “shutting them out” with the headset feels awkward and isolating – even with the pass through. The eyes on the outside of the device are so utterly bizarre in practice I’m kind of baffled that they released it this way. Cutting the external display would have saved weight and allowed the device to be quite a bit thinner, and that definitely would have been a trade off I’d have made. Vs. how the eyes are *supposed* to look, maybe it’s better – but as it is currently, it’s almost zero value.

I think the biggest thing for me is this emphasizes how important robust object recognition and tracking is for a device like this. I need to be able to “attach” virtual objects to real objects and have them move in total lockstep. Apple tries to do this a bit with the Magic Keyboard and a little virtual widget they attach to it, and it’s neat to see the widget track the keyboard in space, but there’s a significant delay, and the object locking is quite … imprecise. That said, it’s a hell of a lot better than anything else trying to do the same thing I’ve seen.

There’s also some weirdness to VisionOS, and I’m sure this kind of thing will be improved as time goes on – but I find that I want apps to “snap to object/surface” by default. I don’t want random windows floating in midair most of the time – I want them to snap to cabinet doors, or wall surfaces, or to lock to my (physical) desk top. Being able to “pull” windows off of physical objects would be necessary, and being able to have them float in space is obviously a must – but by *default*, I want windows to snap to objects and be tied to physical things.

Still – the pass through makes “doing things” in the Vision Pro WAY more comfortable than in VR, or even with the Quest 2’s low-fi B&W pass through. I can look at my phone, or a clock, or see someone approach – and all those things make AVP feel way less isolating than any headset I’ve tried before. But there are a LOT of straightforward improvements that are almost certain to come, and the gen-3 version of this thing will be incredible. I’m still struggling to think of something to make for AVP – for the most part, my personal lack of understanding about how object-recognition might work on a platform like this hamstrings my understanding of what’s possible. I should probably dive into the SDK’s documentation or something. But the other big thing is, “How are users going to use this platform?” And that’s something I still don’t really have an answer for.

This is definitely an interesting platform. It’s incredible tech. It’s just that I can’t easily imagine when I would naturally put this thing on and dedicate time to be in AVP, vs. using a laptop or iPad, and having unfettered access to reality. I think it’s clear that that’s the end goal – that this pass through solution is a placeholder. But this is an interesting start, and it’s nice to see what feels like a significantly different take, with a very different focus, on AR/VR that’s not Meta. If that’s literally all the AVP accomplishes, it’ll still be worth it.