We're in a post-vid slump growth phase.
Nvidia is, as you said, a spike. One driven by AI but will probably be irrelevant in a few years. They're still great at engineering and CUDA will almost certainly continue to dominate, but the AI boom won't last forever. Once ASICs hit the scene it will be game over for GPGPU vendors, like it was with crypto in 2021. They'll still have gaming and industry markets to cater to, but their spiking days are numbered.
So with these recent influx of capital to me that's the market saying that it believes Nvidia and Apple are going to bring in profits in the future so here's money now to grow in order to capture those profits that come in later.
I'm probably misunderstanding aspects here though.
Yes, it's not as simple as "the company will make profits, therefore it's worth a lot".
Apple's expected forward p/e is ~28 over the next five years, which means if you believe that and buy now, you can expect an average return of 3.5%. How much you expect future earnings to grow determines if buying now is a good deal or not.
More seriously, if anybody launches a platform that can't weather a recession, it wasn't going to make it long-term anyway. Windows 3 launched around the 1990 recession, IBM PC launched into the 1982 recession, iPhone launched during the 2007/2008 recession. Platforms take time to take shape. A recession occurring early in a platform's life might even be beneficial, as it will make management more forgiving of "missed" sales targets in the early years.
1 - The USA
2 - Seriously, we're not in a recession and haven't been for 3 years. The economy grew at this rate most of the last decade and there were no complaints about us being in a recession. Money costs more now is the major difference.
For people who think they'll be designing and coding on Vision Pro virtual screens all day might be disappointed.
I’m perfectly comfortable spending 8+ hours in a day writing code on a non-4K screen (widescreen with the pixel density of a 27in 1440p) and based on what I’ve used in the past could probably handle anything as good as a 27in 1080p monitor with the right fonts.
That said, I do find the screen door effect on my Valve Index rather distracting and can’t use it for longer than an hour or two without risking a headache, especially if I am reading any text (tried watching a movie with subtitles and oh boy).
So I don’t think retina resolution is necessary for work (at least for me personally) but it does seem like having the screens so close and completely covering vision does increase sensitivity to picture quality so I may be wrong about that.
What's the killer app for this device which other devices on the market don't support today?
I don’t know that a killer app for it exists, but with its built-in developer base it’s inheriting from iOS thanks to using the same APIs (a dev base that’s notoriously fast to pick up new features, at that), I believe the chances of a third party dev finding a killer app are significantly higher than with prior AR platforms.
Also worth noting this is not a “headset” but a computer (“headtop”?) in a different format. Think of it not as buying a headset for your gaming rig but as a laptop replacement.
If you’re talking about additive displays like xreal , magic lens or HoloLens, then the Vision Pro has a much higher field of view of the content and the ability to display things that are darker than the real world , in exchange for looking at a screen of the world instead of directly at it. Those trade offs are subjective so I won’t focus on them because the tech is so different.
If you compare to other VR headsets with cameras feeding video to the screen, there’s a lot:
1. Clarity is higher. If you look at the footage from their product video which they said is shot from device, it’s a lot more stable and clear than anything up until you get to the $6k Varjo XR-3.
2. Latency is lower apparently. Most other headsets have more of a delay between photon in to photon out based on people who’ve tried it.
3. Better reprojection of the image. If you look at videos of pass through from the Quest Pro, as you move, things warp in the view. From videos from the vision pro, it doesn’t seem to have any of that. (Obviously I’m just taking each company at its word that videos are honest)
Supposedly the Quest 3 is significantly better passthrough than the Quest Pro but very few people have tried it and said anything in depth about it, so it’s too early to judge.
What about mRNA vaccines? Centimetre-accuracy differential GPS navigation systems that fit the palm of your hand, like the uBlox f9p? 3D printers like the Ultimaker S5 that can feed carbon fiber thread into the print? High-end cameras like the Sony A1?
What is the defining feature(s) that makes a new VR headset so much more advanced?
I’m not saying that they didn’t make something great - it does seem great. And given it’s Apple, they’ll likely have great polish and attention to detail, that make a HUGE difference.
But it’s not a best thing since sliced bread. It’s not a technological breakthrough. It’s “just” a new generation of customer electronic device, that’s better than previous one.
And of course Apple is buying the displays for their headset from Sony, so who is doing the envelope pushing there?
While calling it “the most advanced” might be a bit of a stretch, the sophistication of the hardware was something that I wouldn’t have previously thought possible. Some of your counter examples fall short of that complexity. The actual mRNA vaccines are pretty technologically simple, the complexity in differential gps is in its algorithms rather than hardware, and is that camera really groundbreaking (genuine question).
Apple vision is even more technologically simple. You get the screen, attach some sensors to it, sprinkle cameras, add CPU, battery, lenses and you’re done.
> the complexity in differential gps is in its algorithms rather than hardware,
I could somehow understand your ignorance when it comes to mRNA - it’s outside of typical software engineer field of expertise. But GPS?
Yeah, complexity is just in algorithms. Building a satellite that can survive decades in space, building and attaching extremely precise clock to, some solar panels and antennas, thrusters and few other minor details, then getting that onto the rocket and injecting into very precise orbit - that’s an easy part! And especially doing that in 1970s!
All those things are easy compared to leet code! /s
You mentioned some cool tech stuff, but it's all too categorically random and distantly unrelated for a fair and meaningful comparison.