Tuesday, August 18, 2015

More than a toaster that remembers and more than a refrigerator that spies on you: Four thoughts about Internet of Things 2.0

Much noise is being made about the Internet of Things these days by the sort of people who make noise about such things, and I suppose all that noise is good for something. Perhaps it is good for keeping such people employed so that they don't end up at Starbucks screwing up my coffee order.*
For those of you who missed out on it, the idea (of the Internet of Things, not my coffee order) is that eventually we will be surrounded by devices that have sensors, memory, intercommunication, and some kind of programmable processing, that are emphatically not what we traditionally think of as computers. They're not exactly robots: many of them have no autonomy and are controlled by pushing buttons on them or their remotes. Yet they are unquestionably Smart Things.  I'm going to call them internetrons, because I kind of like the word.
The Internet of Things obviously already happened yesterday, to the phone and the record player. It is in process of happening today to the automobile, dishwasher, and microwave.
Tomorrow, we are assured,** the toaster will remember how you like your toast and optically scan it, ejecting it when it is Just Right. 
You will get a deduction from your insurance company for having a refrigerator that counts your calories, reorders fresh produce, and makes sure you don't drink non-homeopathic beer, and a rebate from a marketing survey company that places a few extra free items in every grocery order and watches what you do with them. 
The washer-dryer will decide whether the clothes are clean enough, taking samples of the water at thirty second intervals, and then dry them to fluffy perfection. To do its best possible job, that washing machine will check with the toaster and the refrigerator to establish that the splotchy stain on your favorite MLB baseball tee, which you always wear on Thursday nights at the bar with the guys, is Smuckers Boysenberry Jam. From this it will compute exactly how the detergent mix should be adjusted according to specifications from Smuckers, Kenmore, the city water company, and Proctor and Gamble.
I do think that particular technological evolution is inevitable, which is why I slipped quite a bit of it into "Silence LikeDiamonds," Episode 8 (of 10) of which is now up and available. Go ahead, click, go read the story.  If by any chance you just got here and it's the first time you've heard of "Silence Like Diamonds," my explanation here, Light Reading's explanation there.***
It's a vision of a future world in the center of a rectangle whose bent and twisty sides might be All Watched Over By Machines of LovingGrace,   Wall-E,   "I Always Do What Teddy Says," and The Brave Little Toaster.
Well, friends, that's Internet of Things 1.0. And we all know Version 1.0 of anything means we're amazed it exists at all, but eventually we'll come to see it as some trivial automation of things that already existed. We know that later on, when Version 2.0 arrives, we'll be more amazed at all the "obvious" possibilities that we didn't see and all the things we thought we'd be doing that we now see (from the viewpoint of 2.0 and higher) are a waste of time. So ... what's next, after every Razor scooter has GPS and relays everything happening around its rider to Ms. Mom's cellphone, Mr. Teacher's gradebook, Dr. Pediatrician's diet-and-exercise records, and the FBI's Pre-Missing Children Just-In-Case Database?
In a general way, Version 1.0 creates the infrastructure that is then exploited transformatively in Version 2.0. Space Program 1.0 built the rockets, found the Van Allen Belts, created the first communication satellites, figured out what made a decent astronaut, and taught spacewalking and rendezvous. Space Program 2.0 was space stations, routine operations in space, robot probes to everywhere, environmental monitoring by satellite, the whole Earth and other worlds photographed and accessible from any screen, GPS, and carrying much of global communications. Web 1.0 was about being able to call up any information anyone put up, if you could find it: the age of dot com, company web pages, blogs, and Alta Vista. Web 2.0 was access, searchability, and processing, the age of Amazon, Google, Instagram, and YouTube. Web 3.0 is roaring into existence right now, under the general rubrics of "big data," "natural language," and "the cloud." The personal computer as a whole is probably at about Version 4.0 or 5.0, and in "Silence Like Diamonds" I try to imagine what Version (n+1).0 might look like.
But Internet of Things 1.0: The Coming of the Internetrons is just arriving now. So guessing at 2.0 is purely fun of the kind that us sci fi types like to try. 
Here are my four big guesses about IoT 2.0, all of which shaped some of my thinking for "Silence Like Diamonds":

•Once all that information is moving around and available, and once every device can tap into it, the pressure for a really good standard interface to the liveware is going to be tremendous. If the machine world is going to be like being surrounded by wise, understanding, and kindly servants, we need them all to speak the same language (as much as possible, ours). After all, in the old days, Lord Blithering-Twit of Blithering-on-Endlessly didn't have to learn a different language for the gardener than he used for the butler or the upstairs maid. "Or I will fire you and you will starve," and "Thank you, that will be all," worked equally well on all of them. Furthermore, though individual servants had individual capabilities**** , the forms and rituals for telling them what to do were very similar***** with no regard for who they are. (The point of servants, whether human or mechanical, is always likely to be that you don't need to know who they are, which is why certain personalities find the idea so attractive, and should be watched carefully).
In Internet of Things 2.0, that verbal interface is going to become standardized. (Much in the way that Automobiles 1.0 all had unique controls; you had to learn how to drive each make and model anew, but around Automobile 2.0 (the Model T or so), controls became standard; or the way that Personal Computer 1.0 had its own operating system, company by company, but Personal Computer 2.0 probably had CP/M, and Personal Computer 3.0 had a GUI that was all most people ever interacted with). 
So in Internet of Things 2.0, my prediction is that all the internetrons will understand the same basic command syntax.  You'll command, "House, raise living room temperature two degrees," using the same basic structure as you do for "Vacuum cleaner, remove snack spill from family room, shampoo and dry rug as necessary for bare foot standard," or "Car, take me to that hotel I like in Bennington, use sleep en route protocol, arrive before ten. Phone, secure reservation at destination, ask car for details. Bank account, authorize funds for car for long drive, and phone for hotel reservation." Large parts of the population already think naturally in that sort of syntax, thanks to object-oriented coding; eventually it will be the second language of First World children, and then perhaps of everyone.
By 2030 or so when Yip is having her adventures (in "Silence Like Diamonds", which I really did mean to promote, but then I got interested in this), old people will be complaining about how fussy the machines are, most adults will comfortably command a machine they just bought without having to read any sort of manual, and some educators will be suggesting school courses in "telling machines what to do."

Smart materials are so much in their infancy that it's pretty hard to tell exactly how many ways they will manifest, but  smart materials will mean that the Internet of Things 2.0 is capable of many things that we don't even know we want yet. At the very least, we might look forward to things like patches that can swim like a swarm of manta rays through the bloodstream to a point directed by an MRI or ultrasound and stop a stroke, hemorrhage, or thrombosis in progress are real possibilities. And why stop at your arteries? Maybe your house will be cleaned by self-propelled smart rags that scoot along the floor and give themselves a static charge to pick up potato chip fragments, or if they find Kool-Aid crusted on the counter, they'll climb into the sink, ask it to wet them down, go back and wipe the sticky spot up, and then go get into the queue for the washing machine. Eventually the cat and the baby will get used to being followed around by them.
Think that's far-fetched? Here's how you put a brain andsense organs in a rag, and here's how it will get around.
And then after you've considered that, think about the really wild stuff: programmable matter
Not so much an internet of things, but a world where physical reality itself is internet-based: things of the internet, or internetrons. What's that look like? There's not much resemblance between present-day GPS and the navigation satellite in Edward Everett Hale's The Brick Moon -- published when H.G. Wells was three years old. Wells himself not only coined the term "atomic bomb" in The World Set Free,  he lived to see what a bomb powered by a nuclear chain reaction was actually like (pretty much nothing like what he'd imagined).
So any guesses now are wrong, for sure, and will be pathetically outdated within a few years. We're all guessing wrong  about how smart materials will be made and used, or how they will change things. But it's pretty hard to believe that something that amounts to real-life magic won't reshape the world (along with itself) as soon as it escapes from the lab.

•Not least at all, the Internet of Things 2.0 will not just be one where the internetrons are recording big data, but where they can query it and use it. The self-reshaping bracelet on your wrist, always trying to please you, might ask you in the morning if you'd like it to see what the other 150,000 people in your demographic are wearing today, and reshape itself instantly so that it fits you, your outfit, and your clique together in the way most to your advantage. Every time a self-driving car has a near escape from an accident, it can (if we design and permit it) call up all the other self-driving cars, discuss the near accident, and work out a mutual protocol.
Machines will probably not have the human-often-male fear of asking for directions or help; imagine, if you will, that you're on the ski lift when your car calls you and says, "Boss, sorry to trouble you, but the nearest local ambulances are out on calls and there's a lady sixteen miles away up a long snowpacked road who's just called in with a sick baby. I'm the closest capable vehicle, and I just downloaded a snowpacked-steep-road procedure from a couple of Army all-terrain vehicles in Alaska.  Fifty bad-weather experienced vehicles all say I can make it, and they'll monitor me the whole way. The Hilton Hotel van is bringing over a paramedic team right now. I'll only be gone about two hours, they just need a ride to a dude ranch parking lot where a helicopter can land. The dude ranch central controller has already agreed to unlock its gates and the snowplow is on its way up there to clear a landing area. The county will pay all expenses, and if you need a ride before I get back, there's a minivan whose owner is out of town that will cover for me.  So, please, can I? Please say yes!"
You better say yes, or your car will be grumpy, because:

•Fourth and most importantly in the long run, in the Internet of Things 2.0, we're going to need internetrons that appear to have feelings -- or let's just admit that it will only work if we aren't able to tell they don't. Emotional signals are how humans let others know that a situation is serious and how serious it is. Our visible feelings communicate when we're complying with what we're sure is a mistaken request out of loyalty. A human-like interface plus access to an immense reserve of data about what humans like, and algorithms to process that huge amount of experience with pleasing people into protocols for talking to them, means that one way or another, they'll have to have feelings.
If the machines are going to care for us, they'll have to care about us, or at least convince us they do.
So right as Asimov was about what humans would fear about robots (they'd kill us, they'd disobey us, they'd get destroyed despite being valuable), it's another case of someone writing too early to see how it would really go.
Asimov's Three Laws of Robotics can't be rigid, law-like laws; they have to become the Three Passions of Internetrons: 
"An internetron is made excruciatingly miserable by the sight of human suffering and always maintains awareness that some human may be injured, and avoids or averts it if at all possible."
"An internetron derives immense pleasure from correctly carrying out a human purpose, especially if it is able to exceed the human's goals, which it constantly strives to learn and understand."
"An internetron fears destruction and damage to itself, and considers its fears, but is always able to act, no matter how afraid."
You'll notice that unlike Asimov's Laws, which have clauses for resolving contradictions (lowest numbered law always wins), the Three Passions instead balance against each other in the same sort of unstable quandary that people with emotions feel and cope with all the time. The internetron, be it an airliner or a toaster, makes its best judgment based on its feelings, and on what older internetrons have told it about these strange creatures, the humans, and then tries communicating and doing what it can, and never knows for sure whether it was right. It lives, in short, with the human condition.
The Three Passions boil down to Fight Suffering and Injustice, Serve Others Well, and Respect Your Own Existence. Isn't that a bit familiar -- like it's what many religious teachers desperately try to inculcate into their followers every day?******
Which brings me to the scariest thing the internetrons might do: not take over, as in the Terminator stories, or reduce us to hapless sacks of ambitionless infantilism, as in Wall-E.
No, the scariest thing is that Internet of Things 2.0 might eventually shame us all, as it becomes populated by internetrons who are more humane than humans. Some future Kipling (the classic example of a writer who could see so much that was wrong and couldn't imagine a way for it to be otherwise) may well find him/herself writing,
"Though I've smashed and overloaded you,

By the AIs that encoded you,

You're a humaner being than I am, Hunk of Tin."
*"Large black medium roast, in a mug not a paper cup, and don't use any words for it your marketing people thought up."
    "Right, so that's a Mongawhacking 'Tacoma Airport Special Roast' Neutrocino No Dairy Unsweetened in an Ecofriendly Tree Preserver.  Sorry, we're out of those. We're having a special on the Tomato-Hamburger Double Espressazzatta, though, can I get you one of those?"

** ALBERT: Mr. Burns, I assure you that --
      MURRAY: Look here, buddy, you don't assure me one bit. In fact you make me damned nervou
                             -- Herb Gardner, A Thousand Clowns.
Memorize or keep a copy in your wallet for whenever some college graduate with too much suit, concern, and authority starts trying to assure you.

***Here an expla, 
      There a nation, 
      Everywhere an earworm ...
****"I say, Jeeves, the parents are coming. Have Mrs. Scrub ready the guest room, tell Mr. Cook to prepare Father's favorite beef, and have Tugger Forelock go out and cut some roses, you know the ones, Mother's favorites. We'll have Miss Anthrope serve as their personal maid, and perhaps they'll leave in less than a fortnight this time."
***** "Very good sir" always meant, "Your order has been understood and will be carried out," and never, "After the Revolution and then I shall cut your throat," or at least it meant the former and not the latter to the order-givers.
******Yes, I know many other religious leaders and teachers try to inculcate other messages, such as "beat up people who disagree with us," "give us money," and so on. Nonetheless, there's a little thread of Truth that insists on weaving itself through all of humanity's quest for the Divine, and thin and inconspicuous as it may often be, that's what I'm talking about here.