Om Malik asserts, the iPhone X FaceID is a perfect illustration of Apple’s not so secret “secret sauce” — a perfect symbiosis of silicon, physical hardware, software, and designing for delight. Their abilities to turn complex technologies into a magical moment is predicated on this harmonious marriage of needs. post
Apple’s ability to design custom silicon — its new A11 Bionic processing unit, a new GPU, new image processing chips, S3 watch processor and low power wireless chips (W2), True Depth Camera System — is at the heart of every new product launched today.
The reason Face ID works is because of some key silicon innovations — yes, there is that TrueDepth camera system made up of a dot projector, infrared camera and flood illuminator and a seven megapixel camera. Face ID projects more than 30,000 invisible IR dots.
The resulting IR image and dot pattern is then used to create a mathematical model of your face and send the data to the secure enclave to confirm a match, while adapting to physical changes in appearance over time. What decodes the data captured by this camera (for lack of a better descriptor) are neural capabilities of its A11 Bionic chip.
Microsoft and Intel were attached at the hip. Microsoft released new versions of Windows. Intel made ever new powerful processors that made using Windows easy as a breeze. And that in turn prompted Windows world to figure out how to use up that extra oomph that Intel was sending their way.
This created a flywheel of innovation, demand and ultimately wealth that turned Microsoft into the behemoth it is today. Intel and Microsoft sucked most of the profits from the industry, leaving enough scraps for their feeder army of PC makers.
The iPhone represented a fresh start for Apple.
Just imagine Apple having to depend on Qualcomm to supply its chips — it will be tied into Qualcomm’s ability to come up with new technology — and thus will be working on a timeline defined by the San Diego chip giant.
This doesn’t factor the harsh reality of paying Qualcomm premiums and having to worry about losing chip supplies to someone with bigger orders and desire to work on razor thin margins.
Controlling your own destiny is a smart business strategy and Apple isn’t the only one doing it. Today Google and Facebook are designing and producing their own highly optimized hardware for networking and data centers, mostly because the industry vendors like Cisco Systems made gear that had to fit the needs of many companies.
Google is designing its own chips — especially to conduct resource hungry machine learning and artificial intelligence tasks. When they were thinking about introducing voice-control system on its Android OS, they would need twice as much compute infrastructure as the company owned if every Android user used it for three minutes a day. Instead the company created specialized machine learning chips.
It all starts with Silicon. Unlike software which can be written, discarded and rewritten at a rapid clip, the law of reality makes it hard for a chip to be designed, tried and manufactured at scale. So in a sense, chip and hardware teams have to peer almost two-to-four years into the future, predict what could be possible, what they can make possible and then make it work.
Carver Mead forecast this symbiosis speaking of the tall-thin system designer working at very large scales of integration. He even identified neural nets as a promising application area worthy of fresh analog design rules. What didn't happen is ongoing fabrication by email by corporate giants. Giants don't need that.