Emeritus software engineer, Warren Craddock has participated in many tech projects that have made the front page of the specialized press in recent years, in particular Google Glass, the Google Clip and Lytro, the “3D depth of field camera” of the future… which never made it to the market.
Of course, the most crunchy is to be found in the reasons why Google Glass, in particular, turned into a failure. But, you will understand, a very particular leitmotiv unites these three projects: none is really useful for the public supposed to be the target.
Google Glass lacked a real “killer app”, among other things…
Thus, behind the scenes, the engineers of the Google Glass project were desperately looking for a “killer app” which would have allowed the accessory to impose itself. Warren Craddock worked on the photo / video part of the connected glasses for 16 months. He tells : “no one at Google has ever found the application in question, but the corporate culture has led the teams to refuse the necessary conclusion”explains the engineer.
Its technical characteristics were far from helping the situation, since the developers had to be satisfied with a tiny 640 × 360 pixel screen which was “simply too small, and awkwardly placed in the corner of the eye”. Between its price and a usefulness that seemed to boil down to asking the occasion “Ok Google, how big is the Eiffel Tower”, it is easy to understand why the product has never been marketed to the general public.
However, the firm never quite gave up on the project, and even found its second generation of Google Glass Explorer Edition to be of some use in business – in particular to simplify maintenance in industrial environments or manage inventory in warehouses. .
He also says that internally, the teams found the design of the glasses “ridiculous”: “I’ve never seen anyone wearing Google Glass in the office. The units sat on our desks, plugged into USB, and recharged endlessly” . Despite all this, explains the engineer, Google continued to push the project, without ever directly acknowledging the presence of these major problems.
Recognizing flaws in an idea under development is not always easy
More confidential, Google Clip has taken a similar route. It was a device supposed to record moments of life automatically thanks to artificial intelligence. The device could be worn around the neck or on clothing and the product was well marketed. But failure was not long in coming.
As Warren Craddock explains: “The bottom line is that humans only like photos that are taken from the perspective of other human eyes. It is a characteristic at the base of our psychology, and it is not surprising. We tend not to like what emerges from photos taken from tables, bags, labels or dog collars”.
Finally, the engineer talks about Lytro – a startup that seemed very promising at the time. Lytro was developing photography technology to capture depth of field. Thanks to an ingenious system of lenses and algorithms, it became possible to create new kinds of images in which it is possible to refocus by clicking anywhere or obtain a slight 3D effect by dragging the mouse over it.
However, here the problem was again related to technology and a target problem: “we could only take pictures of small subjects, for example sushi. For taking pictures of larger subjects, especially people, the advantages were non-existent… but most of them were trying to take pictures of people”.
To improve this point, the technology would indeed have required, according to the engineer, a lens of a size that could reach 1 meter which would not have been feasible. What lessons can he draw from these experiences? Obviously, you have to be wary of the corporate culture, recognize the presence of problems before they lead a potentially very interesting project to the edge of the precipice, and not hesitate to talk about it…
I’ve worked on a number of high-profile failures:
– Lytro lightfield cameras
– Google Glass head-mounted computer
– Google Clips automatic photographer
They all had a fatal flaw. Everyone saw the flaw. But the culture that arose in these teams purposefully ignored the flaw. 1/n
— Warren Craddock (@warren_craddock) October 10, 2022