The mixed messages of Microsoft’s Hololens2: very few corporate use cases and lots of limitations

I’m a bit bothered by the overselling of mixed reality, or spatial computing, at least short term (long term my kids’ lives will be dramatically improved by them, we all can see that, but that might not happen for quite a few years). Notice that Microsoft says its Hololens2 is for enterprise uses only, yet to demonstrate it they had a piano on stage. And that’s just the start of the mixed messages I saw.

Microsoft is still overselling the technology. Why? Well, it demos amazingly well and positions Microsoft as being an amazingly cool, futuristic, company. Even though I bet it’ll only sell a few tens of thousands of these, just like the original Hololens.

Most of that oversell, or mixed message, is due to the “God view” in its on-stage presentations. You get to see every virtual object on stage. But when you actually get one on you realize you were sold a bill of goods: that you can’t see that view in the Hololens, but a small little view port.

Even that is oversold. “Greatly Increased Field of View” the Hololens website promises. That’s like saying having two pennys today is greatly increased wealth when you only had one yesterday.

Thanks to my lot in life I’ve gotten to travel to see a lot of jobs. Just last summer I visited the factories of Boeing, Tesla, Ford, and Louisville Slugger.

Where will we see Hololens2 being used? Not in many places. Where will it be used?

The corporate customer experience centers (every company has them, these are multi-million-dollar efforts to look impressive). Why? Because, like it’s doing for Microsoft, it could be used to make a company look cool. That’s the magic of augmented reality. Also, because they hide Hololens’ weakness: that you can’t really wear it for hours. Or, if you try, you really don’t want to.

But will the line worker at Ford wear them? Hell no. Too heavy. The optics will cause eye strain and block too much of the real world. They are too bulky. Some worker who is putting your dash in place inside a Tesla or a Ford would constantly be hitting his or her head and the device. And even if it was being used for, say, training, or tracking of parts, that use case requires millions of dollars of custom software to be written. Software that the current development team building flat, 2D UIs in, say, Visual Studio and C#, simply doesn’t have the skills to build (you need people who have experience with video game engines to do that).

To see what I’m saying, look at all the videos up on the Hololens YouTube site, or on the main site. They look impressive until you look closer. They are all visualization scenarios that only show things that are appropriate for looking at for a few minutes, at most. Even the surgery video is oversold. Let’s say you are a surgeon doing open-heart surgery and you’ll be working on a patient for more than an hour. Do you really want a device on your head that weighs more than a pound? No, and if you get itchy, or want to move it to adjust it, surgeons tell me that doing something like that will cost $1,500 because you’ll need to rescrub your hands and that’s what it costs to do that when you touch something that’s not sterile (due to the costs of everyone else waiting the few minutes for you to go through your scrub proceedure).

But it gets worse. If we are going to really do real work, rather than just amazing visualizations, we need real tools. Note what they demonstrated in the user experience demos: a few sliders and a few buttons. There wasn’t any real work being done. Like what you and I do on our computers a lot, like in, say, a CAD tool (note that Autodesk wasn’t included in any of these demos, Autocad’s leaders told me they were burned by the Hololens team before and are skeptical of Microsoft’s efforts) or even video editing, which would be a great thing to do in spatial computing. Why not? Because finger controls, even though they are much better in Hololens2 than in the original version, aren’t precise enough to be productive. You are better off using a pen on a Wacom tablet or on an iPad screen or a mouse.

Spatial computing glasses do have some major promise. Because you can see through them they could be used in customer service, for instance, or nursing but Hololens2 simply can’t deliver on those use cases, because of the social problems of wearing a big, ugly, black thing on your face, and because they are so heavy that wearing them for hours will end up hurting your neck.

Let’s talk about optics. Note what Microsoft didn’t talk about: multi-focal-point optics. Why not? My friends say these optics don’t do that. And the operating system for Hololens doesn’t yet have support for such a thing. Magic Leap does, and that was the core reason investors gave Magic Leap $2 billion. Why is this important? If you want to really work on virtual items you must be able to get close. My Hololens only lets me get a foot or so away from items and even then if I try looking at items that close for long my eyes get tired because the images aren’t refocused like a real item in your hand. And I’m told by optics experts that the accomodation and vergence handling isn’t nearly as good as in Magic Leap. (Accomodation is the technical term for how your eye changes shape and refocuses on things close to it, and vergence is the term for how your eyes get crosseyed when looking at something close to them).

Think about curling up in bed with your phone or tablet and how close that gets to your eyes. Hololens2 can’t do anything like that. Now, think about a worker who is putting in electrical systems into, say, a Ford truck. I’ve watched them work. They often are within six inches of their work to make sure that things get snapped in properly and, even, they are working in such tight quarters, like underneath a dash, that they don’t have much choice to be far away anyway.

And the optics still aren’t bright enough, nor sharp enough, to be comfortable reading, say, Tweetdeck or the New York Times in bright sunlight. It might be good enough to see CAD files laid over a building site, but, again, that doesn’t require much hard focusing on text or doing much real manipulation. In the video demos they are pretty careful to stay away from that kind of work and more on the “look at the cool visuals” kind of demo.

After the demos were done on Sunday I started hearing that Microsoft is going to be careful about who they sell these to, making sure that buyers actually have a real use case and they won’t try using them to do something outside of a small set of use cases. I don’t yet know if that’s true, but note that they aren’t setting expectations on shipping dates on the website yet.

At least Microsoft has been pretty consistent at saying these aren’t for consumers, although I wish it had been consistent and tried demonstrating on some enterprise machines, or designs, instead of having a little virtual angel flying around stage and a piano. That sends mixed messages to the market that Hololens simply can’t meet yet.

That said, the real battle over the future of computing has barely even begun, which is what I said in an analysis of what it means for Apple and Magic Leap, on Sunday.

Already, since then, Rony, the founder of Magic Leap, has been promising a new pair of glasses with mind-blowing optics and much better use cases next year.

Until then expect the Hololens to be used on limited corporate projects: things that are fun for the CEO or CTO to demo, but aren’t really used much to do real work. Hopefully that changes with future versions, but we need much better optics, much lighter weight, and much more software to do a wider range of use cases and make it easier for 2D software teams to move their old apps into the spatial computing world. I don’t expect all these problems to be solved for many years, do you?

Until then, I wish Microsoft would be more realistic in demos and stop using the “God view” so much in its demos so people get a real feel for what it is like to use these. It oversells the technology and that’ll hurt its credibility with the people it needs most: the evangelists who will need to help companies put them into use and the CTOs and developers who will be asked to build projects with them.

And start giving us a road map for how much effort Microsoft is going to put into Hololens in the future. We need a lot more tools to help turning our factories into spatial computing-ready workplaces. We need global mesh abilities that Microsoft hinted at, but really doesn’t have a strong vision like what Magic Leap lays out whenever it talks. No discussion of privacy, for instance. That punt is acceptable due to saying “these are enterprise only” (we all know we don’t have privacy at work) but real workloads require much more than what we’ve been told here.

One real positive step Microsoft made for workers: the flip-up screen. That shows that Microsoft understands that these devices are only useable for short periods of time and then you want to flip the screen up to go back to other computing devices, or to talk with other people, or use other tools.

Can MagicLeap, or others, take advantage of any of these mixed messages? Possibly, but Microsoft has such a strong lock over most of the computing used in enterprises that I’m skeptical. But Magic Leap, by having a more consistent vision, and one that’s free of having to cowtow to 2D customers, can really set itself apart for consumers and creators. Microsoft has left that door open so far. Will it shut it next year and stop sending mixed messages?

Advertisements