A doorway opens in the wall. Beyond it, there's a gleaming city. A robot flies through, blasting missiles at me. My hand holds a controller, but I see it as a laser blaster. I squeeze the trigger on my controller, sending energy beams at the robot. It collapses against the ottoman. A missile streams by, gleaming and a bit ghostly, and I turn to watch it fly to the opposite wall, where the CNET camera crew, a host of Magic Leap employees, and CEO Rony Abovitz watch me duck and move. The missile passes over them, unnoticed, because only I see it through the Magic Leap One headset I'm wearing.
I'm not used to being watched this much when I try new things, but maybe that's the future. In the kind of augmented reality created by headsets like the one I'm testing, we're all performers in a split experience -- one part real world, one part illusion.
I'm trying the Magic Leap One for the first time, putting the headset through its paces at the company's Plantation, Florida headquarters just a few weeks before its public debut. For years, the startup has been shrouded: It's received $2.3 billion in funding from the likes of Google and Alibaba, but the company has released only a small handful of experiential videos to the public, raising questions about the company's veracity. But now, the hardware has a price and a release date -- the $2,295 headset is available now -- and Magic Leap is ready to show its creation to at least a few outsiders.
I can tell you this: the Magic Leap One isn't vaporware. It's real, and it works. Whether it's more than a developer prototype, and whether it amazes you, is another story. My initial experience didn't blow me away, despite Magic Leap's promises. And yet, I came away thinking it's the best AR headset experience I've had to date -- including my Microsoft HoloLens escapades. Even though it's not all that fundamentally different from the HoloLens, which has been available for developers to purchase for $3,000 since 2016, the Magic Leap One feels better in terms of display, controls, graphics and immersiveness. And by immersiveness, I mean the things I see and interact with feel more real in front of me. Still, though, there are significant drawbacks to Magic Leap's AR hardware, mostly in terms of its limited field of view.
This AR system is a step forward, but not a game changer. Not yet, at least. It all depends on what comes next.
I'm a tough audience. I've tried all of Magic Leap's AR competitors over the last few years, including the HoloLens, the Meta 2 and prototypes such as Avegant's mixed reality headwear. I've tested all of its VR cousins, too. VR puts you right into another world and blocks out reality, but AR juggles a far more challenging task of trying to place virtual things into the world around you. But it's AR that many tech giants, from Apple to Facebook, are banking on as the next revolution in technology, so Magic Leap faces high expectations.
I was floored by my first experiences with the Oculus Rift and HTC Vive. And, ultimately, that's probably why Magic Leap feels, to me, like a familiar stepping-stone more than a revolution. It's also a headset that seems much more targeted at developers exploring the ever-evolving future of augmented reality than it is anyone else.
And there's one other thing that became extremely clear to me: This isn't made for everyday customers yet. But, here I am, wearing it.
Let's start with a frustrating caveat: We had no way of capturing point-of-view footage as we saw it through the Magic Leap One headset. Magic Leap shared some footage for our videos, which don't look as good as what the headset's display feels like in person. According to Magic Leap, the One uses a proprietary light field technology created through small lens chips in the goggles to project images. These chip displays are manufactured downstairs in the Magic Leap's Florida headquarters, which is a former Motorola factory. I got a chance to tour the entire place. Sealed-off bays, dust-free suits, assembly lines with robotic arms, all sit just yards away from the creative team's open office, full of comic books and board games.
Wearing the headset feels not unlike what Microsoft's HoloLens accomplishes, but brighter, crisper and with better graphics. The Magic Leap One visual experience reminds me of Avegant's retinal projection headgear that I wore years ago in its surprisingly bright, vibrant imagery. (Avegant has its own prototype light-field AR headset I tried last year, too, and it feels similar in spirit to Magic Leap's displays.)
Ghostly is a term that echoes in my mind. Images floated and displayed coherently, creating an impressive sense of depth, but the images seem more like a video game than real. Particles, smoke and other semitransparent effects look the best. The room in which I tested the Magic Leap One had slightly soft living room light, high ceilings, furniture, tables and wall art. I used Abovitz's better lit office, too, but didn't get to walk around anywhere else with the headset.
In those environments, objects like a floating, swimming sea turtle, hanging in the air, look charmingly convincing. Small objects fared the best. Bigger things revealed the Magic Leap One's biggest flaw: The display's small field of view doesn't cover everything you see in the room.
VR headsets don't have fantastic field of view either: they're like Scuba masks. But the fact that I can't see anything beyond that mask works in VR's favor to help complete the illusion.
AR headsets have a different challenge because your viewing window isn't as large as the rest of reality. Microsoft's HoloLens feels like seeing ghosts through a small window in front of your eyes. Magic Leap's One suffers a similar fate. Get close to a virtual car, and the edges start to disappear once they hit the limits of your field of view. Magic Leap's field of view is a bit better than the HoloLens, and Abovitz is quick to explain the field of view more as a "cone" of vision: put a virtual thing 50 feet away, and it can look like it's large and filling a hallway. But not being able to see a fuller view of the room's virtual objects is a serious drawback. Sometimes I lose track of things I can't see, and require sound to help me track where the augmented things are hiding, and where to turn.
The Magic Leap One system is made of three components: The Lightwear headset is wired to the Lightpack, a small circular clip-on piece of hardware that houses the guts of the computer, as well as the battery pack. There's a wireless hand controller, too.
The Lightpack PC is about the size of two paperback books and looks a bit like an old Discman. Inside, there's an Nvidia Parker system-on-a-chip, a Tegra X2, with 8GB of RAM, 128GB of storage and volume controls. I clip the device over my pants pocket, and push it down till until a safety clasp clicks in place. The rest hangs outside my pants.
Magic Leap has an over-the-shoulder strap in case you'd rather wear the computer. You won't forget it's there, but it wasn't a gigantic distraction, either. The bigger issue was getting the headset's cable, which is tethered to the Lightpack, to drape behind my neck as Magic Leap's staff has encouraged me to wear it.
The headset doesn't look that massive on your head, but it certainly feels as weird in person as it appears in photos. It looks part steampunk, part space goggles, part Snapchat Spectacles, with a dash of some biomechanical costume. The headband stretches out to fit better over my noggin than I expect, but the goggles rest high up in the back of my head, and angle down on my nose. None of it feels uncomfortable.
The controller looks familiar: Much like those for the Oculus Go or Samsung Gear VR, it's a one-handed trigger-enabled controller with a circular top trackpad. Magic Leap's controller feels heavier and has vibrating haptics. Its circular trackpad has glowing LEDs, too. It's comfortable and responsive, but the nonclick touchpad can get a little confusing to use, and the vibrating haptic feedback feels too subtle.
There's one huge drawback to the entire experience of putting a Magic Leap One on my face: It doesn't work with glasses. My handlers asked for my prescription before I arrived in Fort Lauderdale, and pop-in prescription lenses were supposed to be provided for my demo. But it turns out my prescription broke the mold. I'm -8.75 in one eye, -8.25 in the other -- too strong. I had to come wearing contact lenses instead, which I haven't done since reviewing Google Glass in 2013.
When wearing contacts, my world seems larger compared to my glasses. My eyes focus a bit differently, too.
I was handed three different Magic Leap One headsets, each with a different mix of demos onboard. Here's everything I tried.
Finally, I'm invited over to Abovitz' office again to check out something he apparently thought up the day before I arrived. He hands me a Magic Leap One, and I put it on. He guides me to look out the large angled corner window of his office, looking into the hallway. A dinosaur, a tyrannosaur type, stands upright, made of what looks like either balloons or multicolored bits of pastel candy. Seen from a distance, it seems large, tall, well placed and somewhat convincing. The illusion from a distance is really good.
This is what Abovitz is trying to demonstrate to me: While the Magic Leap's field of view is limited, he argues that the 3D cone of view extends far back, and allows for large-scale effects. Now Abovitz goes out to the hallway, and stands next to the dinosaur for comparison. He walks behind it, too. I'm asked: Can I see him?
I can (his shoes are visible). Also, through the bright glow of the dinosaur, I make out slight hints of him. If I hadn't known to look, though, I probably would have missed him.
The illusion reminds me of the classic magician or stage trick Pepper's Ghost, which uses a half mirror to make a hologram ghost appear. Only this time, the Magic Leap One's doing all the work. It's like a demo of a future Disney Haunted Mansion attraction. Who knows, maybe it will be.
As I said earlier: No, it didn't, really. And I kept wondering why.
The display technology at work definitely looks better in-headset than any of the video demos seen on YouTube in the last week or so. The light field display tech feels brighter and more vivid, and the 3D placement feels solid.
Magic Leap is a better real-world experience that looks worse in the videos you've seen. It's funny, because in comparison, phone-based AR hasn't been perfect, but it's been able to produce some damn good-looking shareable videos.
But also, this isn't really new to me. I've seen examples of AR headsets for over three years. Microsoft did it first with the HoloLens. I've seen similar ideas at work in the Meta 2, and in Avegant's prototype AR headset using light field tech. Now, I'm looking for nuance.
The Magic Leap One seems best for using at installations, places where the environment could be seamlessly layered to maximize the effect. In fact, the Royal Shakespeare Company was meeting with Magic Leap the day I visited, and the immersive entertainment company Meow Wolf is already working on Magic Leap experiences.
Magic Leap's hardware also creates an experience that's less immersive than the best VR hardware. I can really start to feel like I'm somewhere else with an HTC Vive. The Magic Leap One's in-room immersion is bound to have moments of disconnect. My field of view suddenly ends. Or, it's not perfectly recognizing that chair. Or, a slight tracking hiccup doesn't line ideally with my movement. Or, people in the room with me suddenly walk through my personal holograms.
That's a challenge for AR in general, and what makes me think AR is much tougher to crack than anyone's letting on. After all, the recent AR experiences I've tried haven't been perfect. And it'll take a lot more for Magic Leap to be as good as the best immersive theater experiences I've had.
The Magic Leap One feels like a solid first step, but it isn't quite there yet. It could be that Magic Leap's real intention is to get this hardware out for developers, show the world that it can actually ship working hardware and then work on crafting a more fully realized follow-up version. One that tackles field of view. And glasses. And ease of use. And price. And better controls. There's a lot more I want, and much more that the market will demand if Magic Leap -- and AR in general -- is ever going to go mainstream.
Abovitz, of course, with a performer's touch, has a table in the Apple Store-like product display room that's under a cloth. He tells me that the Magic Leap 2 and 3 are under there, and he's only been showing them to investors.
Abovitz compares the Magic Leap One to the Apple 1. That was the first computer that Steve Jobs shipped in 1976 -- the one that was essentially hand-built by Steve Wozniak. The Apple 1 wasn't a machine most people even got a chance to own, but it was a proof of concept. A year later, the Apple II -- the first real, mainstream home computer -- arrived, and the rest is history.
I think the comparison is intentional. This is the trial run, the collector's item.
But if this is the Apple 1, what will Magic Leap's next jump look like?