Ubicomp, and why I think it's broken
A discussion (well, more of a rant) in Teemu's blog warrants repeat here - why I think the traditional vision of ubicomp, that is, the Internet of Things and the Smart Spaces is broken.
You see, I think that intelligence is relative. You feel stupid when you're with people who're smarter than you, but you could rule the land if you were the only person in a village of idiots. Kinda like social status or wealth - you don't have to be rich: it's enough to have more money than your neighbour.
The thing is, that when you make the space around you smarter, you make the people more stupid. And most people just don't like that. For example, I sometimes like to stay late in the office ('cos it's nice and quiet). The automatic system does not see me moving around, so it shuts off the lights and air conditioning. So you have to go an push a button or run around the corridor and wave your hands, hoping that the system will pick it up. And I just hate that.
People want to feel smarter, and in control. When you are overwhelmed with choice, you feel stupid. When you have five options, you can weigh them in your mind, and make a choice which you feel happy about - you feel both smart and in control. Apple gets this - the reason why iPhone is so cool is because it makes you feel powerful and in control as an user: you understand the options (no geekery involved), you can use it with ease, and you get to go wherever you want. Granted, your array of choice is limited, but that only exists so that you can feel smarter.
Mobile phones are an extension of you these days. Jan Chipchase notes that most people are very Maslowian: they carry means to a shelter (their keys); means to purchase food (money or credit cards); and means to connect to their circle of people (their cell phone). They give you power over bad weather, hunger or loneliness.
So I believe that the logical extension of putting smartness are the things that you carry with you. The idea of "googling for your keys" is alluring, but that does not warrant a full-scale smartification of the entire world. It's much easier to make the keys smarter so that they can talk to you and let you know where they are - not Google.
Same goes with money: it's increasingly becoming smarter. The chip cards are essentially small microcomputers of roughly the same scale as a Commodore 64 (though about 20 times faster).
One of the things about Near Field Communication that really fascinates me is how it takes the money and keys and puts them into your mobile phone. So it's real convergence of the most important things that most people carry. But more importantly, it's something which does not require the environment to become smarter. It makes you smarter because you have the power to use the mobile phone in any way you want.
The second big reason why the ubicomp vision is broken is cost.
Building infrastructure costs money. Maintaining infrastructure costs money. Making your environment smarter means that it needs to have maintenance. Yes, it can be smart and call a repairmain to come by - but as long as it's not a legal citizen, it can't pay for the repairs. So who's going swipe the cards?
Is it really ubiquitous, if it works only in very selected patches of the world where people can afford it?
However, consider your personal electronics - like the mobile phone. You get a new one every two years. The carriers will essentially force one down your throat. It's got more power than a turn-of-the-century PC. It's already with you. It's connected almost everywhere (third world countries and USA notwithstanding ;-). You get immediate, concrete, even life-saving benefit from carrying one around. The infra is already laid down, there is no need to bootstrap. Corporations are making loads of money from the infrastructure - but would they make money out of googling for your keys?
Personally, I think the iPhones and Androids and Limos and Nokias of the world have a lot more claim to the ubiquitous computing vision than the internet-of-things folks. They're already connected, and they're everywhere.
The third thing that I find broken in the whole thing is how the human factor has been cut from the equation. Yes, it is said to transform our lives, but I've yet to hear one good reason what exactly would make two home appliances want to talk to each other? And note - I am specifically saying want. Because at the moment, they don't want anything. They do as they are told, without any personality or desires. We need to figure out what a toaster wants (and not ask the one in Red Dwarf) to understand why they would need to network - and if they do, why aren't they talking to me instead of each other?
Yeah, I know that sounds weird. But consider this: we already speak of cars like "it has a tendency to understeer" or "why won't it go?" We are attaching emotions to things which don't have them - and does that not create them? Does it matter if they have any, if we treat them like they had?
Because in the end, it's my life, and all this stuff should be out there to make my life easier, more bearable, and in general nicer. And of course, all my fellow human beings.
(Ha, the lights went out while I was writing this. Damn you, smart environment! I am still here!)
P.S. Yes, I understand the desire behind the ubiquitous computing. I'm just saying that I think it's just mostly harmless tinkering, until either of the two things happen:
- the Singularity arrives, or
- someone figures out a really good business case for it and can solve all of the logistics issues around it.
My cynical self says to bet for option 1. Until then, I think it's just better to help you become smarter, which in turn makes the environment dumber.
Back to weblog
|"Main_blogentry_300608_2" last changed on 30-Jun-2008 19:48:13 EEST by JanneJalkanen.|
Going off on a tangent (and I've always wondered why this is called a 'tangent' when 'normal' would be so much more correct, but I'm a nerd and I also digress...) I remember when IPv6 was young and some people did recommend it because 'then every light bulb would have an IP address'.
No, it wasn't always a metaphor as somebody really proposed that the bulbs could monitor their internal state and tell somebody through wireless internet that they've broken down. I didn't know why I just couldn't look at the bulb and see that it's burned out - in my own home, that is. There might be some places where you don't look at the bulbs but this would just be introducing unnecessary complicated (hard|firm|soft)ware into places where it really doesn't do any good.
Another thought: I think I'd like a *really* smart machine, who would know what I want. Much of the current-day smartness just doesn't work as it should. I tried to look for an old friend with Google and with default settings searching for 'Tomi' gives 'Tommi' and 'Tommy' as hits, without even acknowledging that it fudged with the search string. Not fun.
I think the problem is that even *we* don't know what we want, less so with other people (if we understood other people better, there would be a lot less trouble in the world). So when you say a really smart machine, I think you're talking about smarter-than-human, as in the Singularity. ;-)
Yeah, smart thinking. Great read. Good article. Put a fan one of those double oscillating ones. Not just the fan, but the head as well...yes, away from you and RIGHT NEXT TO THE SENSOR (which is typically near the ceiling) and you get the idea, that way, the fricken lights won't turn out and you won't have to jump up every now and then to do your silly little chicken expletive dance to get it on again. Still, you might consider, it's good, because it keeps you from getting carpal tunnel...got to get out of that chair and stretch every now and again eh? I SO want the ubiquitous very clever smart phone and am forever harangued by my phone company to get fios (I live in a non fios area, yet they still send ads for me to get it.) and you know, you mentioned how there isn't coverage and it's very spotty at best, etc. Yes, we won't be around in 2045 the year Kurzweil indicates in his Singularity Is Near book, that the "Singularity" will show up. Well, I'd just love to have a phone that lasted more than 2 1/2 hours and played/stored music, had vid/cam capability, button keypaid, and oh, a slew of other things, but you know. They've created the need for me to have this crap yet yank us around relentlessly. New IPHONE? 3G? Can't upgrade the camera. WTF? I don't like all touch pad and I love my work discount of 18% off my bill from Verizon...don't want to sign up with another company for 2 years...it's ah, frustrating...don't get me started on how much I despise Bluetooth, which I'm supposed to be able to use to hook up a Nokia N810 to the internet if I want, if that on its own won't connect to wi-fi...hate like hell phones aren't wi-fi...sick, just sick of it all. Glad to hear someone else has the similar feelings. Glad to know I'm not alone in all this miasma of frustration. Thanks for writing. P.S. Got to your site via Blackbelt Jones.
Hi, when I talk about the future, I'm not talking about stuff that is just right behind the corner. More like something that is at least 25 years away. The "internet of things" folks are talking about spimes (as articulated by Bruce Sterling) that are made out of yet-to-be-seen renewable and fabbable materials. With the current paradigm of gizmos (modular circuit-based complicated gizmos like Nokias that hardly work as expected and are hard to tear apart and recycle) there is an unsustainable edge in the volume of crap we generate as human species. How many phones are thrown away every year? Chipchases culture of repair is not going to help us much here. How many of you love those faceless products of which you have no idea how it become into being through the production line? We stick our own face into those products by skinning, hacking and pimping them, in order to make them unique.
Rather than anonymous products with barcodes with embedded metadata descriptions what they are supposed to be as-is, spimes are uniquely identifiable and trackable objects that start and end as data. A person in the era of spimes is a wrangler, who is interested in wrangling, interpreting and making use for that massive ammount of data. You share the 3D model of your things, you remix them as if you were part of open sourcing your physical environment, you use fabbers to fab brand new spimes out of those 3D objects, you track the life-cycle of such objects and you capture and interact with the context of spimes with various technologies – that are not necessarily prime examples of singularity or "thinking things" but rather what computers are good for, sorting, linking and visualizing data for your benefit. Some smart people even think the science as we know it, based on atomization and reduction is behind us (e.g. http://www.edge.org/documents/archive/edge248.html).
The future mobile phone is perhaps a spime in itself, a wand for managing and interpreting such physical non-human appliances around you. We have no idea what we do with the information those things would be able to record and communicate, but as with many technologies in the past, the best stuff is not planned through intelligent design beforehand through consortiums of power but are rather found through serendipity and fruitful accidental discoveries as we intertwingle ourselves in the middle of it.
The current generation of phones are doing ancestry features of future spimes, be it upcycling, remolding features or participating in near-field communication. I have nothing against money and keys morphing with your mobile phone. It's all great and frightening at the same time. After my experience in Tokyo regarding such micropayments and near-field functionality I have to say it's the progress we have to take now, be it risky or not. But trying to see behind the next hill and extrapolating backwards our relationship with technology is not supposed to be the end of thought.
Ubicomp is a broken vision, I agree, but for other reasons. It makes us believe technology is going to morph itself into the environment and we become unconsciously aware of it, without necessarily sensing its presence explicitly. Internet of things makes us much more aware of the everyware around us. What western science and production lines did to us was to disconnect ourselves from the environment in which we live in. It's time for progress and envisioning the yet unimaginable.
I think the field of ubicomp has morphed somewhat from Weiser's original vision. In one of his original papers he says:
"Ubiquitous computing is exploring quite different ground from Personal Digital Assistants, or the idea that computers should be autonomous agents that take on our goals. The difference can be characterized as follows. Suppose you want to lift a heavy object. You can call in your strong assistant to lift it for you, or you can be yourself made effortlessly, unconsciously, stronger and just lift it. There are times when both are good. Much of the past and current effort for better computers has been aimed at the former; ubiquitous computing aims at the latter."
I think this makes it pretty clear that ubicomp artifacts were never intended to possess "intelligence". As such, what you define to be the traditional vision of ubicomp, is not, in fact, the traditional one. Rather, it is the modern one, which is an ugly descendant of the original.
However, I do agree with Teemu in that the real problem is this seamlessness for which much ubicomp research strives. I think seams and tangibility are important properties: sometimes the computer should be exposed.
You may be correct, but then again, so did Weiser's vision change - this paper has quite a clear vision of interconnected agents http://www.ubiq.com/hypertext/weiser/SciAmDraft3.html with plenty of built-in intelligence.
Well, that's debatable, and entirely depends on your definition of "agent". Maybe what you're talking about is adaptation rather than intelligence (the one does not require the other). If you have an issue with any kind of adaptation at all, then your position is much clearer. In the Scientific American paper you link to, Weiser states:
"If a computer merely knows what room it is in, it can adapt its behavior in significant ways *without requiring even a hint of artificial intelligence*."
So what he's saying is that some kinds of adaptation are useful, but we don't want to go down the path of AI. Think of a venus fly trap: it adapts to the presence of a fly (or other body), but you'd hardly call it intelligent.
Yeah, I guess that really comes down to what was meant by AI when Weiser wrote that, and what is meant by AI now. These days, when we say that we put intelligence in the door, or that it's a "smart toaster" it means that it can make more complicated decisions based on some logic. It's hardly AI in the old sense, though some developers of expert systems do define that as AI.
It seems to me, based on this discussion, that the big difference is really in the sense that the old vision imagined old things being re-equipped with intelligence (walls, floors, showers). But the technology is really going into the new things (DVD players, laptops, cell phones). Which in one way completely fulfils the ubicomp promise (just go count all the microprocessors in your apartment), but on the other hand, since very few of these are interconnected, invalidates it too. And there is little point for them to be interconnected.
Many people already feel quite overwhelmed by technology already. They feel stupid, because the VCR flashing 12:00 feels smarter than them.