Introduction

I apologize for being so long between blog entries. I had a bunch of travel, followed by being sick for the a long time. With a mountain of things I wanted to discuss, I had a bit of a writer’s block on what to cover first and spun my wheels. I have photographs and information on many companies and technologies collected from CES, Photonics West AR/VR/XR conference, and other sources.

This blog has been reporting why from a technical perspective on the massive problems with Magic Leap since November 2016. At that point, Magic Leap had “only” raised about $1.4B. They were able to raise another $1.2+ billion since this blog started reporting on their hype. Magic Leap’s total VC raise of over $2.6B dwarfs the measly $700m raised by the infamous Theranos.

The later part of this article will recount some of this blog’s coverage of Magic Leap.  Before I get started on the whale in the gym, I would like to give some clues as to what I hope to cover in the coming weeks.

Often Asked “When Will Magic Leap Implode”

The question I seemed to be most often asked is, “when will Magic Leap implode?” It was not “if” but “when.” Anyone in the industry with half a brain knew that Magic Leap was spending too much money on a product that was always going to cost way too much to make for its intended market.

Some simple calculations which (explained in Magic Leap Patents Have Been Assigned) showed that Magic Leap was going to need more money in 2020. The question was, could they find another rich person with more money than knowledge? It looks like Magic Leap is running out of suckers.

I have been told that back in 2019, Scott Henry, then CFO of Magic Leap, had found a company willing to buy out of Magic Leap. But that Rony Abovitz thought the price was too low, and that this resulted in Henry’s departure. It looks like Magic Leap missed the chance to dump a mess of a company on someone else.

Magic Leap rates an 8 on the 10-Point KGOnTech Theranos Hype Scale

I’m not the first, nor will I be the last to compare Magic Leap to Theranos. Theranos set the gold standard of a high tech company that was all hype and lied to investors, magazines, and customers. Theranos, like Magic Leap, had a “vision” of what they wanted to do that was impossible to achieve. While not as bad as “Theranos,” Magic Leap did some Theranos-like behavior.

Magic Leap led by Rony Abovitz, who had no background or knowledge of display devices and optics. For a hypester, a lack of knowledge can be an advantage as they can make outrageous claims without being burdened by physics. Selling the Magic Leap fantasy cost lot of money such that Magic Leap had burned through the $2.6 billion and was needing more. Even a slimmed down Magic Leap is likely going to need more than $10 million per month.

Magic Leap’s “core technology” was originally the fiber scanning display. As this blog pointed out in December 2017, at least as far back as 2013, the Fiber Scanning Display was Magic Leap’s “Core Technology” that never worked out, much like Theranos’s one-drop blood detection.

Rony Abovitz also was pushing his word salad “Photonic Lightfield Chip.” It was A) not a “Photonic Chip” which implies it generates photons which it did not. And B), the Magic Leap One never was going to produce a “light-field” as the term is commonly understood.  Magic Leap was just going to use a passive diffractive waveguide, just like many companies had already developed. Magic Leap did stack two sets of these waveguides to give a poorly executed attempt at Vergence Accommodation Conflict (VAC) but never was going to support a true “light-field.”

In 2015, Magic Leap released three very deceptive videos, Everyday Magic with Mixed Reality, Waking Up with Mixed Reality, and “ILMxLAB and Magic Leap “Lost Droids” Mixed Reality Test.” They labeled these videos as “Shot Through Magic Leap Technology.” Only these videos had nothing do do with the technology Magic Leap was using to build the Magic Leap One. These videos used a different display technology and completely different optics than were used in the Magic Leap One. I don’t count the obviously fake 2015 Wale Video and the more accurately labeled Concept Video from Weta that others have ridiculed.

Much like Theranos, Magic Leap used the vast sums of money they raised to turn around and buy existing technology to go to market with something less than revolutionary. Magic Leap hired a bunch of people to make another version of Hololens (shipped in March 2016), with different, and often worse, tradeoffs. According to The Information, Abovitz was initially telling investors Magic Leap would sell 1 million units in the first year before being talked down to a more “modest” 100,000 units. In reality, they sold less than 6,000 units (and I am told that many of these were giveaways). With the unit costing so much to make and with a $2,300 price tag, they claimed the Magic Leap One was a “developer unit,” but the fact was trying to be a consumer product, only one that costs way too much to make.

Theranos was scummier than Magic Leap and risked the lives of people, so it sets the standard with a full 10-points on the Hype Scale. A one on the scale would be a solidly delivered product that meets all claims with no hype. A typical product introduction on the hype scale would be between 2 and 5, where they may exaggerate somewhat, particularly on their specification. IMO, Magic Leap rates about an 8.

As I wrote in Magic Leap 2, Many Unsold ML1s, Mortgaging Patents, & Needing More Money:

“File this under learning from WeWorks and Theranos. I have personally seen good high tech startups be put through the wringer in due diligence for a few million dollars. Then you see companies get literally billions in what I call “lemming financing.” All it seems to take is one big name and big ego’ed lead financer does poor due diligence and people with more money than knowledge follow suit.

Magic Leap Layoff, is it 1,000 people or is it About Half?

As reported by multiple news sources, including The Information and Bloomberg reported that  Magic Leap laid 1,000 people or about half its workforce. My confusion is that by my count Magic Leap appears to have had about 1,800 people in late 2018. As this blog reported in November 2019, Magic Leap had been quietly laying off people in 2019 with the layoffs accelerating the last quarter of 2019. So by the time of the latest layoffs, Magic Leap was probably down to 1,500 people or less. So if they laid off 1,000 people, it would be more like 2/3rds rather than half the remaining workforce.  

One can expect that most of the rest of the people will be distracted by seeing their colleagues leave and polishing up their resumes in case they are next. So we can expect to see some more people leaving voluntarily.

Magic Leap Ill Suited to Pivot to Enterprise – More Like Spinning a Narrative than Pivoting

Magic Leap started saying they were pivoting to enterprise in December 2019. The “pivot” was more a rebranding of unsold consumer devices with a $700 higher price tag and with a technical support contract. Unfortunately, the Magic Leap One is very ill-suited to most of the enterprise market, particularly when compared to the Hololens 2.  

As I wrote in Hololens 2 First Impressions: Good Ergonomics, But The LBS Resolution Math Fails!, “Hololens 2 makes the Magic Leap One Look Like a Poorly Designed Child’s Toy.” As the article points out, the Hololens 2 has many serious flaws, but Microsoft clearly focused on better ergonomics.

The Hololens 2 is self-contained and without Magic Leaps dangerously dangling cord. Hololens 2 has many ergonomic advantages over the Magic Leap One including enough eye relief to support wearing ordinary glasses, flip-up visor, much more transparent, digital interpupillary distance adjustment,  a single unit that adapts to most people, much better peripheral vision (a very important safety consideration), less darkening of the real world, and appears to be more comfortable to wear for long periods of time.

Then you get to the not headset factors like having Microsoft backing the design and developing databases for the Hololens. And Microsoft is not laying off about half their Hololens workforce nor about to run out of money. How is a slimed down Magic Leap going to compete with Microsoft?

History of Magic Leap on KGOnTech

Back in November 2016, almost 2 years before Magic Leap One was available, I decided to figure out what was going on with this company that raised at that point $1.4B. I knew a lot of about the difficulties of making AR headsets, many of which are summarized in a December 2015 Display Daily guest article:).

I knew “the physics” of AR was extremely difficult and I doubted that anyone, even with over a billion dollars was going to solve enough of the problems to have a consumer product anytime soon.  With the America Invents Act of 2012 (more on that later), Magic Leap was leaving a trail to what they were doing in their patent applications. Some others with less knowledge of AR and insight into the technology had also looked at some of these patents but drew the wrong conclusions. So I decided to write a series of articles on Magic Leap, not knowing where it was going to lead.

The first articles used Magic Leap’s patent filings to figure out what they were doing

The first Magic Leap article was titled Magic Leap “A Riddle Wrapped in an Enigma”. I followed up with Magic Leap – Separating Magic and Reality where I identified the most likely display technology that Magic Leap was going to use. Almost 2 years later in August 2018, iFixit did a teardown of the Magic Leap One and asked me to help analyze the optics which confirm my prediction of about 16 months earlier (see figure below).  

Analysing Magic Leap Videos  and Realizing that they were deceptive

As discussed earlier, in 2015 and 2016 Magic Leap release three videos Everyday Magic with Mixed Reality, Waking Up with Mixed Reality, and “ILMxLAB and Magic Leap “Lost Droids” Mixed Reality Test.” They labeled these videos as “Shot Through Magic Leap Technology” which was at best highly deceptive. In Magic Leap – The Display Technology Used in their Videos on November 9,^, 2016, I started to analyze these videos frame by frame. In a follow-up article,  Magic Leap CSI: Display Device Fingerprints December 16, 2016, I then when through and explained how I came to my conclusions a stated that the videos were deceptive quoting:

“it became obvious that while the “Through The Magic Leap Technology” videos were NOT using the same “Magic Leap Technology,” as Magic Leap is planning to use for their production product.”

The image below was taken from Magic Leaps April 8, 2016 video. Note particularly how the things that are white in the image stay white throughout the image where they don’t in the picture through the Magic leap One. This is proof that the optics were not the diffractive waveguide (aka, “Photonic Lightfield Display”) as was used in the product.

Below is a picture I took through the optics of a Magic Leap One. Note how the white circles in the image change color indicative of the diffractive waveguide being used. None of these problems can be seen in the deceptive April 8th, 2016 video.

BTW, my best guess today is that the videos were shot using an OLED display with “birdbath” optics similar to Nreal, but the prototype in the video uses a much lower resolution display. BTW, the Nreal prototypes blow away the Magic Leap One on image quality, cost, and weight as I discussed in this article.

December 2016: Reed Albergotti of The Information (and over 20,000 others) were reading this blog

It turns out that Reed Albergotti, then of The Information (now with the Washington Post), was a follower of my blog. On December 8, 2016, The Information published his The Reality Behind Magic Leap (behind a paywall). Reed was nice enough to give this blog as one of his sources (many other articles use this blog for content but don’t credit it):

“And Karl Guttag, an entrepreneur and expert in the field of augmented reality, has also devoted many hours to analyzing the company on his blog.”

Magic Leap’s 2013 Investor Presentation – As Discussed in Dec. 2017

In December 2017, I went through Magic Leaps (formerly) confidential 2013 investor presentation that was part of a US patent filing. Once again, thanks to the America Invents Act.

The most Theranos-like part of Magic Leap’s presentation is the Fiber Scanning Display (FSD). As I explained in Magic Leap Fiber Scanning Display (FSD) – “The Big Con” at the “Core”, using “4th-grade math” one can show that FSD would never work at any decent resolution. It appears that it was this Theranos-like FSD technology that they used to convince people, particularly the people at Google, to make their initial investment in Magic Leap

Analyzing the Magic Leap One rollout and building a 3-D print model to check out some of the ergonomic issues

In February 2018, the Magic Leap hype machine was on full display with a video by Recode Media. Magic Leap brought out “experts” like the NBA commissioner and basketball star, Shaquille O’Neal. I then wrote Hype-to-English Translation Of Magic Leap’s Recode Interview.  

In March 2018 while waiting on the Magic Leap One that would not be available until August, I built a rough 3-D prototype to explore the view through the Magic Leap One and some of the ergonomics. I wrote a series of articles on my conclusions. BTW, the 3-D printed model turned out to match the view through the actual device quite well.

KGOnTech Evaluations of the Magic Leap One (ML1)

In September 2019, I purchased a Magic Leap One (ML1) so that I could review it and wrote a series of 3 articles. The first article concentrated on the view of the real world through the ML1, the second looked at image quality, and the third was a shootout between the ML1, Hololens 1, and a Lumus Vision. Below are summary images and links to the various articles.

Magic Leap Review Part 1 – The Terrible View Through Diffraction Gratings

Magic Leap Review Part 2 – Image Issues

Magic Leap, HoloLens, and Lumus Resolution “Shootout” (ML1 review part 3)

KGOnTech breaks the news that Magic Leap’s I.P. “Mortgaged” to J.P. Morgan Chase – Also That ML Had Been Laying Off People

In November 2019, KGOnTech was the first to publish an article that Magic Leap had mortgaged all its IP to J.P. Morgan Chase. I read a post on Reddit that asked whether the patents had been mortgaged and realized that nobody had written about it. After verifying that J.P. Morgan Chase had taken over the assignment of all Magic Leap’s and their surrogate company’s patent this blog broke the news. Then starting two days AFTER KGOnTech, repeated the story. The original story on KGOnTech made #1 on hacker news and there were over 20,000 unique visitors that day.

Conclusion – How to Waste $2.6B

It truly is amazing how Magic Leap raised $2.6B based on a flawed concept. Magic Leap raised so much money that they could paper over the cracks by hiring smart people to make something that “worked” using well-known technology rather than the miracles promised by Rony Abovitz. But even these smart people could not violate the laws of physics and economics.

When I started casting technical doubt on Magic Leap back in 2016, I had people saying “how do you know that they won’t hire people smarter than you to solve the problem. My response was, “what do you think Microsoft, Apple, Google, and Oculus/Facebook haven’t hired many more smart people?”

In 2012 Google Glass was introduced, Facebook buys Oculus for $2B in March 2014, and then in September 2014, Microsoft goes public with Hololens (but does not ship until 2016). Some of the people at Google that were excited about AR decided to play for them a small bet of $500M (petty cash for Google) on Magic Leap. With Google investing but doing little due diligence, a series of lemming investors followed. There was a lot of industry hype that AR was the next big thing after cell phones and the “fear of missing out” (FOMO) was running strong.

I have tremendous respect for innovative companies. It is rather sad to see so much money wasted on something that was so massively over-hyped.

Some Future AR Topics I Plan to Cover

I that information collected over several months should help put into context events such as Facebook’s exclusive arrangement with Plessey Semiconductors on MicroLEDs. Plessey is the second MicroLED company that Facebook has locked up having outright bought InifiLED in October 2016. Still, Facebook has not cornered the MicroLED market, at least yet. Perhaps the most interesting MicroLED company is Ostendo, which has been demonstrating a full color on a single device to selected people. Jade Bird Display(JBD) of China has also been showing many advanced MicroLED devices similar to Plessey. I have seen and learned a tremendous amount about MicroLEDs and the optical matters related to working them over the last several months.  Covering MicroLEDs and the optics that go with them are going to take several articles. A big part of this story is why some companies that use pupil-expanding-waveguides, both diffractive (ex. Hololens 2, Waveoptics, and Digilens) and reflective (ex. Lumus) are looking at using lasers in addition to MicroLEDs.  

After studying the light throughput efficiencies of various optics with MicroLEDs, I have developed a new appreciation for non-pupil-replicating optics. Both Tooz (spun out from Zeiss) and Oorym (startup founded by the founder of Lumus) showed interesting optical solutions at the Photonics West AR/VR/XR conference. It is also time to update what I understand about pin-mirror optics by Lentin AR and Kura.

Appendix: The American Invents Acts (AIA) passed in 2012 hurts startups but help my blog

I have some experience with patents with my father, and two brothers have been patent lawyers. Furthermore,  while I am not a patent lawyer or agent, I am an inventor on over 150 patents and have served as a technical expert on about a dozen patent litigation cases. This background causes me to keep tabs on the evolving US patent system. The “America Invents Act” (AIA) of 2012 radically changed the U.S. Patent System for the worse. The justification for the AIA was to prevent patent trolls, but in reality, it was a way for large companies to keep innovative companies from getting good patents.

Thanks to the AIA, tech companies, particularly smaller ones, now tend to leave a very public trail of what they are doing before they can produce a product. The AIA does the exact opposite of what the title says. The AIA was lobbied for and crafted by Google and several other tech giants to prevent (mostly US) companies from getting patents. I discuss these issues in more detail in  https://www.kguttag.com/2017/12/22/2013-magic-leap-confidential-presentation/ and particularly in https://www.kguttag.com/2019/12/12/the-lack-of-apple-ar-patents-and-apples-leveraging-the-aias-ptab-banana-republic-courts/).