The Abundance Puzzle
What Does It Mean To Live When Things Are Made By The Billion?
[The following appears, in a slightly different form, in a monthly newsletter I write in my job at Google Cloud. You can subscribe here.]
A few years ago, the head of a chip company told me, “This is the first time in history when you can make something and expect to sell 100 million of it.” He was talking about contracts for the chips that go in smartphones, products made by just a few companies that sell by the billions.
What he didn’t see at the time was what else that would mean — all sorts of new products that result from the new scale, perhaps including engineered telepathy.
Welcome to this month’s Speed Read, the abundance issue.
Engineered telepathy is no joke, even if it is at best several years off. Last month I saw it described in a fascinating presentation by an engineering whiz named Mary Lou Jepsen. She is a cofounder of the “One Laptop Per Child” effort to crush the cost of computers, and has also worked in cutting-edge display design at Google X and Facebook. Now she runs a company called Openwater, which aims to crush the size and cost of magnetic resonance imaging, or MRI, machines by 1000 fold. And maybe solve this telepathy thing.
Here’s how it works: Red laser light easily and harmlessly passes through flesh. Using smartphone components like cheap laser chips used for facial recognition in smartphones, along similarly cheap with cloud computing and Artificial Intelligence algorithms, Openwater plans to look at blood flows in places like the human brain or breast. Since tumors use a lot of blood, they show up as big anomalies.
If the costs really are low, it might be possible to monitor the development of a tumor with this new fangled cheap MRI in a cap or a sports bra, rather than going in for vastly more traumatic surgery. And, as fidelity improves over the decades, look at the flow of blood into various speech and image processing centers and eventually read minds. This is already done on very crude levels.
There’s more to it, and you can listen to Jepsen’s talk here. For our purposes, the thing to think about isn’t mind reading, it’s the strange new level of innovation that is being created by cheap and abundant processors, cameras, and other technical goodies created by our global involvement with technology.
When you can sell 100 million or more of something, particularly for demanding, low margin/high scale businesses, peeling off another 100,000 units for the telepathy thing (or camera doorbells, or speech-parsing home devices, or new screens and data storage on all kinds of things) is almost a trivial exercise. And engineers looking at these components are limited only by their imaginations.
So, the question becomes, what else is cheap, at scale, and ready to be developed? All kinds of things.
Data and data storage, for one, as these sensors connect to various clouds to do some of the heavy computation required by their new products. The data is cheap, abundant, and potentially connected to other data for new insights (Jepson talked about “streaming healthcare” as we begin to model the complexity of a human body moving through space and time).
Possibly specialty chips, particularly the Graphical Processing Units (GPUs) and TensorFlow Processing Units (TPUs) that power much of AI. As AI becomes part of everyday life (it’s already in your Google searches, parts of your phone, and in your home device experiences), that demand will also move high enough that chips will become something to hack with.
Most interesting: geniuses. You may have heard of Battushig Myanganbayar, a kid in Mongolia who at the age of 15 aced MIT’s online course in circuits (he’s now a researcher at that university). Even 10 years ago, and for all time before, it’s likely that a Battushig would have had an impossible task taking his genius to the world. Einstein wasn’t just smart; he was lucky to be in Zurich, with access to publishing. In our connected, 100 million of something world, the inventory of geniuses will almost certainly skyrocket.
What else might go up to planetary scale, then become cheap and utile enough to be repurposed? Solar panels have done a nice job in terms of getting cheaper and more efficient, and there are probably other low-impact ways of powering devices. Genetic engineering will probably create new materials at a cheap and scalable rate (though spider silk in goat milk still hasn’t done much).
Going back to one of my running themes — the cloud, and all that’s created around it — I’d add, good ideas, and the ability to leverage them to a greater number of people. Those will certainly be more abundant. So will the other kind.
How will we sort it all? There’s still no good algorithm for taste.