Logo della Fondazione Giannino Bassetti

FONDAZIONE GIANNINO BASSETTI

Innovation is the ability to achieve the improbable

Intestazione stampa

Focus

Temi in evidenza, a cura della Redazione

Home > Focus > An Interview with Dr.Marc Smith

An Interview with Dr.Marc Smith

by Jeff Ubois [1], 10 July 2007

Dr. Marc Smith [2] is a Senior Research Sociologist leading the Community Technologies Group [3] at Microsoft Research in Redmond, WA. His group focuses on computer-mediated collective action, and he studies and design enhancements for social cyberspaces. In particular, he is interested in the emergence of social organizations like communities in online conversation and annotation environments. The goal is to identify the resources groups need in order to cooperate productively. Marc is the co-editor of Communities in Cyberspace [4], which explores identity, social order and control, community structures, dynamics, and collective action in cyberspace, and has developed software called Netscan [5]that measures and maps social spaces in the Internet, starting with the Usenet. A related effort, Project AURA [6], allows users to associate conversations (and more) with physical objects using mobile wireless devices and web services.

Jeff Ubois:
Could we start by talking a little bit about what you're doing at MSR there, and then move into how you think about the downstream effects of your work?

Marc Smith:
I'm a sociologist and part of a group that studies the ways that groups of people, when they interact through computational media, leave traces; and how those traces in aggregate form patterns, and how those patterns have a lot of utility for both basic research and for various practical purposes that include things like the enhancement of search engines. So, for example, we collect data from social environments on the Internet. Sometimes old school places like Usenet, but also still mod and happening things like web forums and discussion lists and emails and personal emails. And all of these are examples of social computational artifacts. All of them are inscribed with some notion of the social, which is to say a relationship, which is also to say a recipient.
All your email has somebody's name on it. A lot of your documents have somebody's name on it, or have been dropped into an email and sent to a person. And so, these things then describe a variety of data structures that can be looked at over time, and I think most interestingly, they can be looked at the form of the directed graphs or social networks that a lot of these relationships or these artifacts inscribe. So I hand you something, it creates a tie. I send you an email, it creates a tie between you and I, and these things then ramify into large complex networks amongst large groups of people, some connected very tightly with many others, and some others differently, so.

Jeff Ubois:
You had a great presentation I thought on this idea of social science and climate monitoring. Can I get you to recap that a little bit?

Marc Smith:
Oh, sure. We've just started collecting and dredging, collecting a bunch of data from these repositories of threaded conversation, most notably UseNet, but others too, let's just broadly say there's this space of social software and this is what the wikis and the blogs are in, the whole read/write web and other kinds of collectively authored content. Let's just call that a space, and admit that we have never really seen this space with clarity and holistic detachment that, for example, we have when we view the planet as a whole. And I thought it was notable that, when I started just poking around at the history of weather modeling and weather prediction and weather data collection, in part because I live in a place where the weather, while monotonous, is sort of irritating, and you'd like to know why it's cloud cover 268 days a year.

Nonetheless, weather modeling is an interesting topic. We are all connected to the weather, and it's a topic that we have perennially joked about as the weather man is rarely correct; but it is, in fact, the case that the weather man has become more correct, and it is measurable the gap between predicted and observed. And so the predicted / observed gap has closed and so you ask, well, why? Is it the case that meteorologists in the last 40 some years have just gotten smarter? I'm sure it's the case, but there are probably other explanations. And I point mostly to the fact that 47 years ago we got the first satellite photograph, television satellite photographs of weather patterns from space. And it was only 47 years ago. And image doing weather prediction the day before that was available? How good a job could you do?

Jeff Ubois:
Yeah.

Marc Smith:
So my metaphor is that social sciences and the study of this domain - and its many domains, not just the domain of online social media, but just many behavioral sciences lack data. They exist in an era prior to 47 years ago when satellite photographs of the earth became available. They lack these holistic data rich streams of information about their phenomena.
Further, I think it's not just the sensing side of meteorology that got better. It wasn't just that we knew what the temperature actually was, we were able to then compute new models and tune the models and continuously tinker with the models and compete with different models, almost on an evolutionary process - well, in an evolutionary process, whittling away the model until the model has gotten quite good, in step with computations, storage and sensing. And computations, surge and sensing and theory and more graduate students, that's needed to really move the ball. So let's now take the analogy to hey, is there a satellite photograph of new social phenomena? Is there a new widespread data collection technology available? And the analogous models and storage and computation frameworks to do things perhaps like prediction of weather? And this is perhaps where the metaphor breaks down. Does it have to be prediction? I'm not sure it does. But it certainly is the case that there is big bend in the prediction quality of meteorology, as satellite observation and see surface temperatures and all that kind of stuff. You know a lot of that computational power goes up. So the metaphor where it does work best is saying, hey there's this whole new domain of social interaction, and then a whole bunch of existing domains of social interaction.

Jeff Ubois:
Right.

Marc Smith:
And they are about to be subject to machine data collection at a scale and at a diversity in the population never before seen. And already we can see that when people take their social interactions to the keyboard on the desk or on their lap, but soon we are coming to a time when computation comes to your interactions when somebody is sitting on your lap.

Jeff Ubois:
Yeah. So when you think creating these new models and collecting this data on a massive scale, what do you think might be some of the social problems that could be observed or studied or solved with these models and with this data?

Marc Smith:
There are some technologies that, at the very least, early on are dangerous in proportion to the power that they offer, without regard to the costs they impose.

So, you know steam engines, very useful, burn all coal available, level all forests. Choke skies, cause lung cancer. Oh well, oh okay, long term costs, even impinges on me, you know, eventually my ox gets gored, too, I can't move away far away from it. All right, so we mitigate, mitigate, mitigate. But the history of the technology is, you know, seize new power - then mitigate, mitigate, mitigate all the pathologies. How long was it from Model T Ford to safety belts?

Or for that matter, to cup holder, rear window defogger, you know, all wheel suspension, air bags. The whole idea of make them safe, oh, how long did it take?

Jeff Ubois:
Safety comes later. You don't even know the threat in advance.

Marc Smith:
You don't even know the threat in advance. Or, even if you do, you discount it. Happens to other people, doesn't happen to me. Happens to those crazy kids that drive like nuts, you know, so eventually the insurance companies get involved.
And that's when pathologies get socialized and regulated. It's when somebody's pocketbook can't wiggle away from the cost, then it gets resolved. You know, hotels grow big enough to have hundreds of people because of steel, which is great, but unintended consequence, no fire suppression and cheap mattresses, and everybody dies from smoke inhalation fires. The fact that you could build the building big enough to kill that many people depended on one innovation, steel; but needed another innovation, fire suppression, to catch up with it.

Jeff Ubois:
Right.

Marc Smith:
So you're always dealing with this herky-jerky interleaved, interwoven innovation. No one thing innovates all by itself. And few innovations are about the domestication of innovation, until somebody is bearing the cost of that innovation. Now, the great thing about innovation ... well, I'll tell you a story, something I heard at another software company before I came to Microsoft.

This guy early on tells me what's the definition of a successful software package? He says it's one that solves one problem really compelling, while causing ten.

I'll never forget it. And he's right. And I said, why is that? And he said well, why would anyone ever upgrade from the first one. It solved the problem in a compelling way. Did you leave any surface area for embellishment? No. Well, that's no good. So, first you want email. Okay, here, my problem I have to solve is moving a little bucket of bits from here to there in some reliable fashion. Oh, now you have too much email. Now you have spam email. Now I have some room for innovation. But actually giving you email - now the beautiful thing is, few things deliver the first without delivering the second.

Jeff Ubois:
The 'raw new power followed by mitigation' cycle is one that's worth exploring or thinking about. Maybe one question that could be part of that innovation process is what sorts of mitigation technologies will be needed to cope with this?

Marc Smith:
Oh, and by whom? Because a lot of innovation...

Jeff Ubois:
Right, yeah, move the cost to somebody else, if you can, right?

Marc Smith:
That's right, so my Misses was showing me this documentary she got from the library on "the world is our junkyard" or something of the sort. I'll have to check the reference, but they basically track waste streams back to where they go, where they're from. And so until the people in Indian villages or Chinese villages are in a position to say, the lead from the solder from the motherboard from Dell that they built with a design from Singapore in a factory Taiwan is making my children into morons.
And we matter too. But you know and I know that if they matter enough, then the plant will be moving to Nigeria. There's always somebody that doesn't matter.

Jeff Ubois:
That's really dark.

Marc Smith:
And so, really, a lot of costs, if they can be pushed around the planet, probably are more difficult to resolve in terms of the social regulation of them. Because you can always push the cost somewhere else. Some problems, by the nature of the resource, don't have that. So, for example, air pollution. They may be subject to lead up close and personal, but eventually, it actually leaks into everything and it comes to us, too. Now my ox gets gored and I have to start mitigating the pathologies of innovation.

Jeff Ubois:
So this idea of somehow sharing the risks associated with innovation is one that might be interesting to look at. I mean, if the political problem is fairly allocating the risks or the cost of innovation, then somehow you need to be able to look downstream a little bit or have some sort of social process where you...

Marc Smith:
Distinguish innovation from what? Instrumental practice?

Jeff Ubois:
Well, thinking of innovation along this continuum from discovery to application. You have a discovery and then you begin to apply money to that discovery, and that's when - that's when you really have a process of innovation. It's not just merely the equation on the chalkboard, it is the refining of uranium and combining it with dynamite and aviation to deliver something that incinerates 100,000 people, you know. It's not just the theoretical.

Marc Smith:
Right. And innovation, you know, what is it? Necessity is the mother of invention, right. You're always faced with problems and innovation typically comes from the people who are most squeezed out of the sweet solution space. So you don't innovate unless you have to.

Jeff Ubois:
Is that true? I mean, a lot of money is invested deliberately by wealthy people and institutions in innovation.

Marc Smith:
Oh yeah. Look at GM. And look at Bell. You do not innovate, especially if innovate means unexpected new technical practice, or at the moment not fully domesticated, which is also to mean that the profits of which, or the power of which has been fully regulated or privatized, which is another way of saying regulated.

Jeff Ubois:
Well, I guess where I was thinking about the economic rewards of innovation are now perceived to be huge.

Marc Smith:
Oh, I think the economic rewards for selective innovation can and have been huge. I think Mr. Galileo did a lot of defense contracting. So innovation - but, you know, why did he do defense contracting? Well, because they were at war.

If you're in a dominant position, you're not likely to innovate.
Innovation is the behavior, I think, of marginal actors in an ecological landscape. Other things are sort of planned innovation. I don't think that's the same as disruptive innovation.

Jeff Ubois:
Well, planned versus unplanned is maybe a good distinction for us to get into a little bit.

Marc Smith:
Yeah, I mean, you, especially if innovation by its definition, and I'm not sure you would include this, is 'and is disruptive to some existing power'. And if that's the case, then I think it does refine and support my 'innovation is the act of the marginal.' Because why would existing power overturn existing power?

Jeff Ubois:
Well, I mean, I guess to reinforce its own position, right? We have to innovate or other people will do it for us.

Marc Smith:
Well, innovators' dilemma, you know, and so that's not that easy.

Jeff Ubois:
One of our themes is along the line of what would sustainable innovation look like? Are there sustainable patterns of innovation that work around this dilemma?

Marc Smith:
Do you mean innovation of technologies that are themselves sustainable, which is that to mean have minimal pathological externalities? Or is that sustainable innovation a method by which a stream of innovation, whether it's of sustainable things, continues.

Jeff Ubois:
Well, ideal would be a sustainable stream of innovation of sustainable technologies that are socially and technically and environmentally sustainable. One level down would be a sustainable stream of innovation of technologies with mixed effects, that is to say ones that may need a lot of mitigation. And next level down from that, I guess, would be temporary constellations of money and power that can innovate for a little while before they fall over.

Marc Smith:
Remember, you want some negative externalities, right? There has to now some distinction between toxic externalities and merely irritating externalities, because irritating externalities are opportunities for further innovation. So that I think is one thing. But how do you get sustainable innovation, which is to say the kind of continuous process of generation of innovation? I think you sustain marginal populations.

Jeff Ubois:
Okay, well, that's interesting, I mean, it's interesting to put that in the context of MSR.

Marc Smith:
The research labs of incumbents have an interesting history. Westinghouse Labs, Bell Labs, certainly, you know, Park was interesting, in that they are about reinforcing the incumbent.

Jeff Ubois:
Is that a contradiction in the - is that more innovators' dilemma? Or is there more to it than that?

Marc Smith:
No, no, because I think there is vast opportunity to refinements of the system of capital accumulation. There are vast improvement opportunities in making light bulbs, more reliable brighter, whiter, faster, lower cost materials, with greater reliability, able to survive being dropped, blah, blah, blah.
Squeezing all the efficiency out of an existing pattern of capital accumulation is fine, and indeed the incumbent always invests in that. Now, historically, once you got an incumbent position, you weren't necessarily riding a technology that was as fast in its cycle of change. I mean, the massive change - the rate of change of the telephone system, I think, pales by comparison to what microelectronics looks like in the last 30 or 40 years.

Jeff Ubois:
Right.

Marc Smith:
Okay, so often the incumbent had plenty of time to regulate innovation, such that it wasn't really innovation, I mean, it was novelty, it was refinement, but the point of coming up with something that breaks all this and makes our capital investments moot? No, that's not what we are going to think about here. And if, and it does happen, those things emerge and they become unavoidable, then you have a tectonic-like experience in which we deny it, we deny it, we deny it - then say 'okay, we have to accept it.'
For example, the acceptance of Linux, the acceptance of Apache, the acceptance of Firefox, the fact that you really can no longer ignore Firefox, means, you know, another diffusion of innovation. But we've got a - innovation just is a way of labeling the power of the adoption of technical instruments.

Jeff Ubois:
We were talking about instrumentation earlier. Are there some ways of observing innovation if you're thinking about tracking human behavior in aggregate, maybe part of that behavior is innovation or the application of new technology? Are there some observational methods of innovation that you can point to, particularly if they allow us to foresee effects or act in a more responsible way?

Marc Smith:
The possibility of having people out around the city with 2-way radios, or I guess ultimately cell phones, but initially I think some form of 2-way radio, to report on traffic conditions. This was an innovation. And note that it was a response to an irritation, which was a response to an innovation. So you get the car, everybody gets cars, they move to the suburbs, they try to get to the jobs, there's traffic, they listen to the radio, an innovation, and what do they hear? A traffic report, an innovation.
And a market is created with value in which all sorts of companies are formed that provide you with traffic reports, and they themselves innovate largely - you know, they adopt cell phones, and they adopt some kind of way of taking a feed from the traffic reporting system, which is itself, you know, the highway department innovates. It puts up cameras. But innovation is always a response to irritation. If cars never congested, and no one ever had an accident, why would you put a camera up?

Jeff Ubois:
Right, right.

Marc Smith:
Let's look at a picture of cars moving smoothly. Okay. You know, so innovation and irritation may - necessity is the mother of invention. Necessity means gotta have it, and so, you know, cliché but true. But that just means you're in pain. In some way you need to resolve your problem, and it doesn't mean you succeed. It's just that why would you bother looking for a solution to a problem you don't have?

Jeff Ubois:
One form of innovation seems to be to build something and then create or excite a need around it. You know, foment some crisis around a problem that might not exist. Another form of innovation might involve addressing something really basic like clean water for everybody. Maybe there's some way of categorizing - I mean it seems like we're coming up with categories of innovation, whether it's initiation versus mitigation...

Marc Smith:
Sounds like you're trying to categorize just action, I mean, is there good action and bad action? Well, yes. And do good actions have bad consequences? Yes. Do those - does that make the action bad? Well, now we have to deal with knowledge and intent, and things get murky. But, you know, innovation in abstract is hard to really say good innovation, bad innovation. You know, the automobile was an interesting innovation. An unintended consequence was an increase in teenage pregnancy. An unintended consequence is the formation of municipalities, that had to get their act together in order to create a tax base, in order to field an automobile mounted law enforcement official, in order to manage these death on wheels speeders coming through town, which were then also seen as a potential revenue source to feed the municipality. But it was definitely a driver towards municipalitization, which was an irritant.

Jeff Ubois:
So you might argue that since the effects of innovation can't be foreseen and go far beyond the will of the innovator, maybe the innovator doesn't really bear much personal responsibility ...

Marc Smith:
Well, wait a second. So I made the distinction based on the intent and knowledge and capacity for knowledge. If you knew that you were about to create the technology of maximum human immiseration, and you do it, well this is a different thing than you were gene slicing your way to the cure for the common cold, but whoops... You know, that's a different thing. Now, we're angry at you, if we have any opportunity to talk to you about this again, we might be terse, but there is a difference between 'oops' and 'I'm going to get you all, bwa-ha-ha.'

Jeff Ubois:
Yeah.

Marc Smith:
Okay, so if you make that distinction, and I think we all need to, then you accept that the world's gone all grey and foggy at all sorts of choice points.

Jeff Ubois:
Right.

Marc Smith:
Where, I don't know, I mean we could argue. But the idea that there's good innovation and bad innovation, well, you know, there are good people and bad people. There's good action and bad action. It's complicated. But yeah, I think you could say, hey the innovation that makes it so that we exploit these folks better. Well, I guess the question is what, which folks? You know, if I say that in the form of I can increase click through on your website by 15%. Could those two things be synonyms and be okay? You know, is that an increase of exploitation [of the visitors]? Well, you know, if Amazon can increase sales of this book, and it's your book, you [may] think you are delivering value, but Amazon's taking more of your paycheck out, burning up all that fuel with UPS trucks, servers are up all the time, and chewing up the landscape.

Jeff Ubois:
So what if we bump it up a level and think about institutional design. Is it a question of having accounting systems to take note of externalities in certain ways?

Marc Smith:
Yeah, absolutely. And I think it takes a while for people to even bother to collect statistics on automobile fatalities. It takes a while to perceive all of the negative externalities.

Jeff Ubois:
So how can we encourage people? How can we encourage people to take a look at those negative externalities?

Marc Smith:
Well, you know, there are a lot of folks that are doing things like the true cost movement. And I would even argue that my own little AURA project has good intent. That follows very much along this very line, which is when you look at any packaged product, are you, in fact, looking at the full spectrum of information associated with that product, that you might really need to know in order to make the decisions about your consumption of it. And that doesn't just mean get it for the best price, although that's part of it. It doesn't just mean getting it fixed when it's broken, that's part of it. It might mean things like which country did it really come from and do I like their labor practices. Or does it contain the pesticides or the heavy metals that I don't want to have recycled.

Jeff Ubois:
In fact, you know, we should have done AURA up high in this interview. That's a really interesting application -let's go down the AURA path here for a minute because it is so relevant...

Marc Smith:
Right. So there's innovation with a big, warm, fuzzy big heart, I mean, you want it to be good. You also want it to be an interesting business, right? So we observe, hey, here's a development in technology. Lots of people see this. I mean, innovation is rarely, truly unique inspiration. Many people see the mobile devices changing and getting sensors and connected in a pervasive way and the world has machine readable objects in it, and hey, the computational capacity of these devices is actually growing quite rapidly, and so you're now toting around a non-trivial package of computation, storage, bandwidth, display, and audio. It's a non-trivial device, and looking at the world and saying, hey, couldn't you know more about the world? Well that's not the innovation. I think the innovation is, hey, you could put all those pieces together and actually (a) make it work now. So there's the reduce to practice. Hey, you know, we can do that now. Well, let's.

Jeff Ubois:
Yeah.

Marc Smith:
And I think a lot of the wave front of innovation, and let's just put a place mark here for later, because I want to come back to the idea of hitting a vein, and I mean that in a mining and not a medical or sinewy reference. Because I think that things like silicon microprocessors and the whole Moore's Law 30-40 year race was hitting a vein.
And suddenly, you can just go and go and go. And yeah, you know, Moore's Law only goes - and you could draw this other line along with Moore's Line, and that is capital expenditure invested in manufacturing plants...That innovation did not happen by itself, but sometimes it's sort of obvious. You throw money here, engineers there and more money comes out here. And so you hit a vein, and you can go for a long while on it. And we're still mining several, and perhaps have discovered several more, veins.

Jeff Ubois:
Yeah.

Marc Smith:
And so in prior days, you know, learning that, hey, we can make metals strong enough to create boilers, and build boilers. And then actually create the machinery to pull that energy out of that into mechanical force. Well, okay, and you can go a long way with that. And, you know, first you burn all your trees, and then you go for your coal, and if you're lucky to sit or unlucky to sit on coal, then you get to have an industrial revolution. So, that's like the slip strike events in plate tectonics. You go whew! Away you go. And often when you land, you land with a thud, and there's lots of reverberation. I would call the Industrial Revolution some reverberation.

Jeff Ubois:
What did you mean by slip strike? I don't know that term.

Marc Smith:
Oh, in plate tectonics there's different kinds of earthquakes, and one is when two pieces are sort of like when your hands are like if you press them together like you were praying, and then sort of made them different shaped, sort of wiggled them a bit, and then pushed them against each other. They would hold and hold and hold and eventually the energy in it builds to the point where they just like slip past each other in a very explosive way. That's a slip strike, versus some where basically the ground just falls down and wiggles up and down for a while.

Jeff Ubois:
Yeah, okay.

Marc Smith:
There are different kinds. So, you know, sometimes you hit this break through point and you just run, and we've got nano and genetic and electronic and photonic and radio frequency and computational and the, you know, whoa!! We've barely scratched the surface. There's a ton of stuff that we now see we could do, and it's just a matter of figuring our way through it. There are probably 40 year horizons for most of these basic sciences. The chip guys may go, oh, well, golly, we're running out.

Jeff Ubois:
Yeah.

Marc Smith:
But the photonics guys are going, yeah, you know, well we have some pretty big problems, but, you know, we're getting there. And there's no reason in principle why when we get there, we won't be back on that curve going.

Jeff Ubois:
Right, right.

Marc Smith:
Just like we were before. And in the meantime, you guys go sort of like spread out a little bit.

Jeff Ubois:
Yeah, I mean I think that's the whole thing about Moore's Law, it's kind of layered, right? There's lots of different ways of ensuring that the industry as a whole moves ahead at that rate. You know, you can improve the masking or your ability to parallelize things, or ....

Marc Smith:
Right, right, right. So it's the push the parallelization. The minute you hit an obstacle of scaling, you push out.

Jeff Ubois:
So we were talking about AURA, really, and wanting to be good and wanting it to be a business and the world being filled with machine readable objects.

Marc Smith:
Labels aren't big enough. You want more information. And handhelds are there. Well, let's actually reduce it to practice. You "innovate". You try something new. We intend that people will get a better deal on goods, or that they'll be able to strike up a conversation and annotate goods, such that things that I think you're saying, that negative externalities, the total cost, becomes evident.
And that may not just be total cost, but total benefit, as in 'Hey, I didn't know my computer could do that. Hey, I didn't know that that author also did write another book.' Well, in the world of Amazon, that's sort of a lame example these days, but. So the idea of cultivating a conversation around every physical object in the world, at least on the surface, seems like a perfectly innocuous and no doubt likely to be salubrious kind of intervention. Shouldn't it be that you could find out more? But well, you know, there's several people who find it to be a hostile concept. There are various people who would rather see you not have all the information that is available, available to you. There's a reason why they made the label exactly the way it is. Why would they want you to have access to all that other information?

Jeff Ubois:
Well, preserving information asymmetry in the context of trading is a tricky thing. I suppose it's an innovation to have some new way of doing off market transactions that affect share prices.

Marc Smith:
Yeah. Or it's an innovation to be able to play video cassettes in your home. Eventually you go, okay, that's gonna happen. You give in and you go take several billion dollars out of the marketplace, because it turns out not to have been a bad thing for you. It was a good thing. I mean, was that [the VCR] a good innovation or a bad innovation? You know, the movie industry, it was a very good innovation that they thought was going to be very bad, and they made a boatload of money, which may or may not have been good, because now you're doing utilitarian calculus, and you know, there was a reason they don't offer that in college anymore. Was it better that because of video cassettes more people cocooned at home, leading to the decline of neighborhood theaters, where people used to go out and be together and have more experiences of shared culture, which is to say mono-culture, which is to say factory-generated Hollywood culture, which was that bad? But, well, you know, on the way to and from the neighborhood theater that you walked to - but wait a second, we don't walk anymore. So that's more car trips. And that, you know, you could go nuts trying to figure out where to bound what is good and what is bad, and where the end of the ramifications of the technology are.

Jeff Ubois:
That's right. Well, I mean, maybe there's some way of defining the stakeholders in a particular innovation in some way ...

Marc Smith:
[Its easy to] identify them. They're at the top of the legal documents that are served to you. That's how you identify the stakeholders.
The ones that actually get the job done, make themselves known, tell you that you're creating negative externalities for them, and insist that you remedy. Everybody else just suffers.

Jeff Ubois:
Maybe there's some way of identifying those people a little bit in advance. A little bit earlier on.

Marc Smith:
So innovation, you know, is a lot like revolution. It's not usually a smooth or peaceful process, even if it doesn't necessarily take place with violence.

Jeff Ubois:
So, you know, one of the jokes about Silicon Valley is that it's a single company town, and there's lots of little divisions of it. Or it's kind of like Hollywood, where these studios - where these movie projects are pulled together and pulled apart, you know, in the scale of hundreds of millions of dollars. Is there a way of thinking about governance of innovation as an emerging property of certain types of networks? Or a self steering, kind of self governing phenomenon?

Marc Smith:
If there are solutions to problems that people have, if they have anyway to find out about it, they will. And if they have anyway to implement it, they will. In terms of governance, that's a different thing, because that starts again to say, if you knew what the right thing was, could you legislate it, could you enforce it? If you mean the efficient diffusion of innovative practices, I think that's pretty efficient. You know, you have highly motivated information seekers, and now a very rich mechanism for information exchange. If you need to know, and you speak any of the major languages, and you have an Ethernet cable, you can probably can find out. The main thing is, do you know that you should know? Do you know that you want to know?

Jeff Ubois:
One way of thinking about what the Bassetti Foundation is doing is studying approaches to mitigation.

Marc Smith:
May I suggest that your best method for mitigation is documentation of negative externalities?
So, we live in a world where if you get any positive returns, you get to keep them.

Jeff Ubois:
Yeah.

Marc Smith:
And it has been said that the genius capitalism is the privatization of reward and the public - or the socialization of cost. It makes some sense, though, because all organisms consume resources and excrete waste products into an environment - some of that environment reuses waste products where it actually sees input, but the idea that the upper classes are telling us to eat shit is, well, perhaps appropriate, but there's always going to be this process of consumption of resources and the creation of negative externality.

Jeff Ubois:
Yes.

Marc Smith:
And in fact, a thriving society is probably the one that effectively organizes itself so as to prey upon some set of other societies in an effective manner and distributes its resources such that it can perpetuate this fitness for out-exploitation. You know, the Italian city of Milan would be delighted if in some way the Chinese leather goods factories could be thwarted. The Milanese fine goods, the clothing manufacturers, the designers have been brutally beaten by the Chinese. They want to "innovate", which is to say beat back the threat of. And so they would be delighted if they could create negative externalities. That would be innovation.

Jeff Ubois:
We have good labor practices and you don't. And people...

Marc Smith:
Our shoes are made in a way that are now demonstrably better in some way and you can't duplicate it.
I mean grapes have to be grown in Champagne, if you want to call it Champagne.
Negative externalities are very important.

Jeff Ubois:
And that's not just a label, but I think that gets back to the personal practice of innovators and the institutional practices. So documenting these negative externalities...

Marc Smith:
And do you want to? In some cases you have no desire whatsoever to document negative externalities.

Jeff Ubois:
Well, when I'm the innovator, maybe. When I'm the innovator trying to make a pile of money.

Marc Smith:
Well, yes. Yeah, I mean it depends on which end of the innovation you're on, and which end of the externality you're on. Now if you're on the receiving end of a positive externality, you're delighted.

Jeff Ubois:
One of the things that we've been - one of the bits of advice that's floated back from people I've spoken to is that the public dialogue around innovation needs imrovement. My favorite example of that is around stem cells, where you had reasonable scientists talking about possible benefits, losing the argument to anti-abortionists really badly, and then rather than managing to educate the media or the public or legislative bodies, they got people in wheel chairs with placards to holler back. Maybe not the best style of dialogue, and yet without the good bumper sticker, you tend to lose political arguments. Are there some - you know, the other side of your role there I think at Microsoft is improving collective dialogue or collective deliberation. Are there some ideas about that, that we might want to explore?

Marc Smith:
I think that when you try to bring these adjudication decisions into mainstream national decision-making, that the level of mainstream national decision-making is such that you will never do a very good job of it. And so, the idea that all of America should be sitting around talking about stem cells, I don't know. I don't really - I don't agree.

Jeff Ubois:
Well maybe then it is to push those kinds of deliberative decisions back into professional bodies of a certain kind, or make them reference standards of like that UN declaration that I was talking about earlier.

Marc Smith:
Well, I think it's just very hard to do prior regulation. I mean, should we say, hey, you physicists, you cut that out. Not over there. That black hole stuff, that's gonna create the Q-bomb.
Don't do that. Well, you know, this is the whole Terminator I or Terminator II movie thing, right? Can you go back and stop it before it happens? Can you decide not to play with matches? Can you not get consumed by the fire you start? No really ahead of time, no. No, you don't know what's dangerous. And if it's dangerous, then you have a game of, well, I better learn it before he does.
And so probably everything that can be done will be done. Anything that can be done that might work, will be tried. Anything that works, even marginally, will be exploited, as soon as possible by anybody who has the opportunity.
And just think about "innovation" in genetic populations. You know, the minute it confers an advantage and it gets conferred to other generations, then it gets, you know, away it goes.

Jeff Ubois:Yeah.

Marc Smith:
Nobody in the population goes, now wait, whoa, whoa, whoa. Before I breed with you, and I understand that you have genetic advantage 23, but what are the long term global consequences of us actually successfully avoiding that disease? Don't you understand that we're just going to over-graze this area, and that's the greater harm?

Jeff Ubois:
Strong precautionary principle stuff tends to break.

Marc Smith:
Nobody is going to forego, as Axelrod would put it, you've got the shadow of the future, w. It's the discount parameter. How important is the future to me? And if I'm only looking one or two turns into the future, then smashing you over the head and grabbing your wallet makes sort of sense. Whereas, you know, inviting you out dinner and chatting with you and seeing that you safely get home makes perhaps more sense, if I think that over time, this relationship might yield me more than whatever's in your wallet right this second.

So certainly the discount parameter matters. And are we talking about this week, this month, my kids, their kids, their kids' kids, you know, who knows? None of us are very good at long discount parameters.

Jeff Ubois:
No, that's right.

Marc Smith:
We're not even very good at like the millisecond discount parameter.

So I think the issue of governing innovation is a question that when cast at that resolution, has a limited useful answer. Now, if you're saying how do we stop stuff that we see is going in a bad direction? Well, now you have a technology regulation and negative externality problem. And that I think is one that is more tractable. Which is, well, you know, I'm not so sure about technology regulation, myself, but I think maybe nuclear nonproliferation is a good technology regulation. Gun control. I'm actually for it, you know, call me kooky. That's technology regulation. Clipper chip cipher keys controlled by the government. I'm not for that one. Well, wait a second, why? So there's no clear - I mean just because I happen to, well, those are my flavors.

So, I think that you can "govern" innovation when the language and the data to document negative consequences are more available, freely available, easily used. So, for example, I think that a la AURA, one other interesting issue will be with the mobile device. I predict that the mobile device will contain the equivalent of a genetic sequencer and a gas chromatograph, a spectral analyzer.

Jeff Ubois:
Okay.

Marc Smith:
And I believe that to some extent, the entire chemical spectra will be sensible by a mobile device. And so let's just start with, I can detect carbon monoxide. I can detect sarin. I can detect, you know, just arsenic in the water. So what if when you pick up your phone, there's now a button and you look at it, and it's the equivalent of sort of a digital dosimeter, only it's an environmental dosimeter. It's how much mercury is in your environment, so it sort of like slurps the air and every now and then you dunk it in the water or the food you're eating.

Jeff Ubois:
Yeah.

Marc Smith:
Oh, okay. What will happen when the real-time map of that data shows high correlation with socioeconomic status and racial qualities, which is to say that poorer people live near heavy metal? What happens when we make visible the erosion of the environment and the lived environment even more important, I mean, we're humans, for God's sake. Well, you know, that data, no doubt, will be subverted, it will be blocked, but it's now a force in the universe that didn't exist before.

Jeff Ubois:
Right. Let's talk about encouraging innovations like that...

Marc Smith:
If somebody came to me saying, you know, I want to cultivate innovation, I could give them all sorts of answers. You know, oh, well, you want to cultivate some kind of fluidity or liquidity of ideas, give people opportunities to mix and mingle, but you know, ultimately, if there's a stress on the system - and why would Milan care if there wasn't a stress on the system? Chinese are competing with us, population's growing older, we're barely reproducing the native population, there's a lot of issues.

Jeff Ubois:
Yeah.

Marc Smith:
Well, okay, giving people opportunities to get together and try to solve their problems, is pretty much how you do it.

Jeff Ubois:
So that's kind of the more the private conversation. The semi-private conversation or the not mass scale conversation.

Marc Smith:
Well, this is like, what are the obstacles of innovation? Is there enough capital in the market? Can they get server space? Do you need a license to innovate? You know, then it's a matter of the observations of all those things that are related to, for example, open source software projects or volunteer driven end user content creation projects. There's something called the dung beetle effect.

Jeff Ubois:
I don't know this one, but I suppose somebody's rolling up a ball...

Marc Smith:
Scientists have observed dung beetles and they make the following observations: 1) dung beetles don't actually communicate with each other except through the motion of the ball; 2) dung beetles will attempt to roll a dung ball themselves, but if they bump into somebody who is rolling a bigger ball, they may join that ball; 3) beetles that meet bigger balls, attract even more beetles. Dung beetle scientists have figured that beetles working together can roll bigger balls than lots of dung beetles rolling smaller balls all by themselves.

Jeff Ubois:
Okay.

Marc Smith:
The interdependence requirement. The key thing is that no dung beetle has to do anything to join a ball, they just join a ball. And if they only push it even for a minute, that's okay, that's fine, even if they go back to their own ball, that's okay, that's fine. No obstacle to contribution. So are there obstacles to innovation? Sure.

Jeff Ubois:
Yeah.

Marc Smith:
Innovation means solving my problem. Well, where do you have a problem? The problem you have is somebody else's innovation. Is there a net gain globally? Is it better for the Italians that shoes are now cheaper? Even if their uncle is out of work? I don't know, go ask an economist. It doesn't feel that way, and if you tell the Italians that, they're going to say go to hell. I'm not the globe, I'm me.

Jeff Ubois:
How local is your self interest? Which is again another way of saying, what are the negative externalities?

Marc Smith:
Exactly.

Jeff Ubois:
So...

Marc Smith:
Where you draw the line really matters.

Jeff Ubois:
Yeah. So we didn't really answer, and you know, you're smack there in the middle of one of the largest institutions related to innovation in the computer industry.

Marc Smith:
Oh no, this is research, not innovation.

Jeff Ubois:
Okay.

Marc Smith:
What about it?

Jeff Ubois:
I'm thinking of the cultural policies of those institutions.

Marc Smith:
What do you mean by cultural policy?

Jeff Ubois:
Well, how do you decide that it is better to have a product that solves one problem in a compelling way and creates ten others? Or do you decide that you want to own those ten others, or let other people build on top?

Marc Smith:
Usually you can't stop what other people do.

Jeff Ubois:
You can make it easier or harder. You can document the APIs or not, right?

Marc Smith:
Right, right, so do I build software that embodies my values? Yes, sir. AURA has an API. It's freely available. It's released under a license term that allows you to do all sorts of innovation with it. We don't own your innovation. We don't charge you for it. We've opened it up so that it's easily reconfigurable. When we publish our data - when we collect data about Usenet, for example, we publish it to universities and share it. We publish it to the world, and they can see it. You know, we do a lot of things that embrace some of my ideas about information sharing and openness and being a contributor to a collective project. My organization doesn't oppose that, and often requires some aspects of it, at the very least, the peer review and publication, validation and certification process.

Jeff Ubois:
Have you run into conflicts? Like...

Marc Smith:
No, only in that you can want innovation and not get it. You can have innovation and not be able to actually deploy it. That's the conflict. There's no real conflict. Everybody's like, hey, this is great, but, you know, I don't have the resources to invest in it.
Therein lies the real conflict, which is, I don't have the resources to invest in it, because either it's perceived that there's a better investment opportunity, or this investment opportunity looks like it's disruptive in some way. Which is to say the other one is a better investment opportunity.

Jeff Ubois:
Well, also, in a sense your job is to create disruptive innovation, right?

Marc Smith:
No. I don't think our job is that. I think our job is refinement. And our job is to make writing the next desktop operating system easier. Or to make what gets written better, but not to say no more desktop operating systems. There's plenty of research to be done in incumbent research labs, where it's not really about coming up with a new thing that undercuts the old thing.
Coming back to my slip/strike metaphor, sometimes you give in. When IBM actually got into and then out of the desktop computer business, but never got out of the desktop client business, or at least being involved in that business, it capitulated. It was like the slip/strike. Pressure builds up, it builds up, it builds up, it releases. Suddenly you go, okay, Linux exists. All right, Firefox exists. Back in the Netscape versus IE, giving in to and supporting some level of parity, if not a second position to something like Firefox would have been unthinkable, but slip strike bang, innovation.

Jeff Ubois:
We will cooperate with the inevitable.

Marc Smith:
That's right, and that's what great leadership is.

Jeff Ubois:
The AURA stuff is very interesting. I want to push on that one a little bit before you go.

Marc Smith:
My thought with AURA, and with Netscan , and I guess my equivalent good intention with Netscan is, do you have any idea how much information there is about you out there? And by looking at Netscan, you come face to face with that information. And I distinguish ourselves from others who do this, because no doubt, others have, and I would distinguish ourselves in the following way: We let you know what we know about you, and we let you scrub your name off of it. So you know that we know, whereas the others don't let you know they know, and you can see what we see. And that's different from what you get when somebody looks, but does not let you see them looking. So it is the reciprocal...

Jeff Ubois:
Reciprocal transparency, in a sense, yeah....

Marc Smith:
...Rather than the un-reciprocal that is more typical these days. So I think that points to our larger goal of designing tools that are digital mirrors for digital lifestyles, where you're trying to convey just how easy it is to leave an imprint in computation.

Jeff Ubois:
Thanks very much.

Show/Hide links in this document

Links in this document:

  1. 1] /schedabiografica/Jeff Ubois
  2. 2] http://research.microsoft.com/~masmith/
  3. 3] http://research.microsoft.com/community/
  4. 4] http://www.amazon.com/exec/obidos/ASIN/0415191408/investiofsocialc
  5. 5] http://netscan.research.microsoft.com/
  6. 6] http://research.microsoft.com/~masmith/2003%20-%20Ubicomp%20-%20AURA%20Demo%20-%20Smith%20et%20al.pdf
CC Creative Commons - some rights reserved.
Marc Smith
See also: Microspheres by Andrea Pitasi
Articles by:  Jeff Ubois
Articoli
Search by:
Search video by:

- Mailing list Subscription - Cookies Policy -

RSS Feed  Valid XHTML  Diritti d'autore - Creative Commons Gruppo Fondazione Giannino Bassetti in Facebook Gruppo Fondazione Giannino Bassetti in Linkedin Segui la Fondazione Giannino Bassetti in twitter

p.i. 12520270153