Thoughts on the “One Laptop Per Child” project

Thoughts on the “One Laptop Per Child” project


Sometimes first impressions are totally wrong. Other times, they turn out to be right—usually by complete coincidence. My first impression of the “$100 Laptop” idea developed and promoted by Nicholas Negroponte and colleagues at the MIT Media Lab is that it’s brilliant. Since then, I’ve heard a lot of criticism, and I think some of it is justified. In the end though, I still think it’s brilliant. Maybe it isn’t the best plan imaginable and maybe the agents making it happen aren’t doing it “all for the right reasons”, but in the end, those are trivialities. The Cold War was a stupid reason to go to the Moon, but it was getting there that mattered.

One of the biggest obstacles preventing computers (and through them, the internet) becoming the “great levellers” of society is that not everybody has one. In fact, only a fairly small minority have them, and the ones who do are overwhelmingly the middle-class inhabitants of the “first world”. Ah yes, the “first world”—another holdover from the Cold War, when we (in America at least) classified the world as “Us” (1st World), “Them” (2nd World—a.k.a. “them Commies”), and “Everybody Else” (3rd World). Maybe someday I’ll spend some time taking that idea apart, but not today.

The poor may always be with us, but we keep coming up with new euphemisms for them (why do we need euphemisms one may well ask—what are we afraid of?). "Third World", “Developing World”, or "The South" have been very popular with different folks. I’ll use “developing” here, hopefully this is nicely inoffensive.

Having been conditioned by an increasingly strong conservative faction in the “1st World” which likes to claim some kind of genetic birthright to success, we’ve long since internalized the idea that we’re somehow smarter, faster, and ultimately more deserving of wealth, computers, and the political power that comes with them. But what we really are, here in the “developed world”, is luckier, better educated, and better capitalized. Any other advantages that you may happen to possess have nothing to do with what country you live in, what social class your parents came from, or what you write on your tax forms, and are as likely to be found scattered thoughout the Brazilian rainforest or the deserts of Somalia as in suburbs in Illinois. Human potential does not breed especially true, possesses an extraordinarily complex intersection with the environment, and really doesn’t vary as much as advertised.

Why am I talking about this kind of stuff? Because there is, in some of the criticism I’ve been reading about the “One Laptop Per Child” project, an insidious thread of elitism, which needs to be outed. We are told, for example, that people will not know what to do with the devices, that they will sell them for cash or trade them for supplies, that they will simply not have the foresight to invest in the self-education opportunity the laptops would provide. The promoters of this argument are assuming that we will quietly acquiesce to this idea because we all know what poor people “are like”.

The fact is, of course, that given any random sample of human beings, some of them will act stupidly. So, in fact, I guarantee you that some of that argument is true—some people will pass up the opportunity in exchange for short-term gains. But not everyone will! Because it is also true, that given any random sample of human beings, some of them will do the smart thing.

If we were purely economic machines, as Western economists so love to model us, we might not put sufficient value on a means of education and communication versus simple animal needs like food, clothing, and shelter. But poetry predates electricity and running water both.

And just as people will cling to music and art no matter how poor they are, they will cling to the power to write and be heard, to read and understand, and to organize and motivate.

Speaking from a purely economic point of view, the “massive sale” model of delivering these devices works against such forms of loss, anyway: if they are given to students in massive quantities, they will effectively flood the market, making it fairly hard to sell them, even if you wanted to. If Quanta makes good on its idea of creating commercial versions of the units, they will also compete with the government-sale units, and with conventional laptops. So, there won’t be as much pressure from wealthier buyers either. (The objection here was that somehow the $100 laptops would get purchased at slightly more than $100 and compete effectively with $1000 units—my point is, “so what?”. If the market wants cheap laptops, then that’s an even better reason to make them, isn’t it? Just make more. This would be a different picture if we were talking about government-subsidized pricing. But we aren’t, we’re just talking about bulk manufacturing).

There’s another saying, which is that "the freedom of the press is for those who have one". Well, every internet-connected computer is a press. And a free press is one of the most important tools for preserving and promoting freedom and democracy, which are in turn, the most important tools for promoting social justice. While most of us in the developed world don’t leverage that power as we might, the people who aren’t getting a lot of social justice are very likely to start talking about it.

Not at first, of course. First, they have to get educated and learn the rules of the game. Some of them will actually have to learn to read and write, some will just have to learn to read and write English (or French or Chinese or Swahili or another regional lingua franca). Some will just have to learn netiquette, and where to get the latest Inkscape or gVim package.

But if that can be done, then there is real power. Unlike the “simputer”, for example, the OLPC $100 laptop has a keyboard. That’s good for more than just typing URLs, folks. That’s the ability to write a message and pass it on.

Now, of course, for any of this to work, the devices have to pass the technical hurdles. This is, fortunately, what we pay engineers for. And in the developed world we’re pretty good at engineering. We did, after all, go to the moon, and that was a long time ago. You know, a really long time ago, before we invented VLSI microchips.

A lot of previous, small projects aiming at getting computers into the developing world focused on community facilities, and generally used second-hand or surplus equipment that had been designed for use in the developed world. Many problems ensued because the machines simply weren’t built to handle what life in the developing world was going to throw at them. Most machines would choke if the power spikes too often, or if the room isn’t climate-controlled well enough.

So, one thing the $100 laptops have going for them is that they are actually engineered for the places they are going. Now, of course, the engineers can make mistakes, and the application is revolutionary enough that it’s bound to have unforeseen consequences. Perhaps the mesh networking won’t work as well as hoped in practice, or maybe the power supplies will wear out sooner than hoped, or the flash memory might saturate. Nobody can really expect to know that for sure until the units are in the field. But that’s true of any device—the only reason it’s such a big deal here, is because of the huge size of the orders.

One seriously misplaced criticism is that the laptops will wind up in black market or gray market channels being sold to wealthier users. This is actually kind of silly. It’s based on the erroneous idea that the $100 price tag is an artificial subsidy. There’s room to be skeptical about whether it can actually be done, but the point here is to make such a massive order of units that that price tag can be achieve through sheer economies of scale.

It’s also based on design. Computers don’t have to be expensive. Manufacturers have just figured out about what we (meaning we in the first world) are willing to pay for say, “a new laptop” computer, and they design the best thing they can make that has that price tag. That’s why, for example, laptops have cost around $1000-$2000 for the last ten years. Surely you can see that if they only had to produce a unit to the specs of a 1990s laptop in 2005, they could slash that price tag instead of trying to up the features or performance, right?

Consider for example, what cell phones cost. Another complaint about the $100 laptop that I’ve seen is “why is it based on a laptop instead of a cell phone?” This is seriously misinformed in my opinion. The thing is pretty close to a cell phone or palmtop device in terms of its internal technology. “Laptop” is a form-factor and a description of use, not a technological category. The $100 laptop is not a derivative of any existing series of laptops—it just looks and works as a laptop. Internally, you’ll find the same extremely rugged, low power, inexpensive design that you would find in a phone.

The fact is that marginal costs on electronics manufacturing can be made very small. What kills that in the developed world marketplace is that you never get sufficient economy of scale, because the orders simply aren’t large enough. People want different features in a laptop, there are many different brands, and lots of people will want something different next year.

But what if you found a very large number of people who are more interested in stability and a low price tag than in getting the latest and greatest? And I mean a really large number of them. The OLPC $100 laptop project is talking about delivering a minimum of 100 million units. That’s a truly enormous number. Or put another way, it’s $10 billion in transactions.

Quanta, the company who has been selected to make the units, quotes something very close to that for its present annual revenues. So this project is going to double their 2006 or 2007 revenues, despite the tiny per-unit price tag. True, they probably aren’t going to be making a lot of margin on this job. But the volume is nice. Also, by negotiating sales through large government buyers, they are setting up guaranteed sales, not just a potential market.

Which brings us to another criticism, which is that it’s all very “top down” “command economy” stuff. Some people are bothered by the idea of doing this in such large orders, negotiated by big government with big business.

Well, normally, I’d tend to side with those folks. But the fact is, that big orders are the only way to get small unit costs, and that means producing an awful lot of identical units, so somebody does have to do a “big design up front”. It’s regrettable, because it probably will introduce some inefficiency and market mismatch. No doubt many people will still need to buy conventional laptops in the developing world (I’m missing the part where that’s a bad thing, though).

This is just a form of collective bargaining, which has been an important skill for poor people for a long time. No I don’t mean “since 19th century Marxism or trade unions”, I mean at least since 500-250 BC when Roman “plebeians” started using military strikes to force government reforms. No doubt someone can think of an earlier example. Certainly I am sensitive to this point; I owe my university degree to the collective bargaining power of American manufacturing trade unions.

The technical criticisms are, in my opinion, the ones with the most merit.

Laptops by themselves would be useful, of course, but the killer app for these machines is clearly to get people connected to the internet. The problem being, of course, that network support is poor in many of the places where these laptops will be deployed. Mind you, I said “poor”, not “non-existent”.

Part of the problem with criticisms of the laptops’ design seems to rest with the fact that the laptops are designed with “worst case” conditions in mind. Theoretically, if you spread these laptops out among people distributed more or less evenly throughout a vast swath of desert or rainforest, running solely off of hand-cranked power, they can still form a “mesh network”, in which each working laptop is both a server and a client. This is a beautiful theoretical extreme, but of course, the laptops probably wouldn’t work very well in practice in such an application. Given the need to use cranked power, users would probably only start up the machines when they want to use them (not when others need them for relay). Given such a thinly distributed group, it would be necessary for nearly every laptop to be on all the time in order for the mesh to work.

Fortunately, however, that situation would be pretty rare. Even in the developing world, people mostly live in fairly tight clusters, there will often be some form of electricity available to at least some of the users, and fixed network nodes will exist. The ability to cope with the deprivation of any or all three of these factors is built-in robustness of design—design for the worst case, not the most likely case. The factors cover for each other too: public network nodes would eliminate much of the dependence on the net, AC power would mean more laptops could be left on more of the time, and clustered users increases the chances of the mesh holding up, even if only a small percentage of machines are on.

So, the presence of any one factor serves to mitigate the problem. I can also imagine behaviorial fixes—such as students getting into the habit of working at the same time of day (likely to happen naturally, anyway).

Critics responding from developing countries have scoffed at the “cute” product design of the OLPC, citing greater cultural value in utilitarianism. That sounds to me like a bit of bitterness, though. I think that, especially as far as kids are concerned, the possibility of owning something that even kids in the “developed world” would think is “cool” has got to be a positive experience, even if it has always seemed to be such a remote possibility that they have trained themselves not to expect it.

It’s also true that such features do not come at the expense of utility. I’m pretty sure that it does not cost any more for the computers to be a shiny lime green than for them to be gunmetal gray, olive drab, or some other boring color. There’s little point in deprivation for deprivation’s sake, no matter how ascetic you are when you need to be.

Other critics have argued that phones or radios would be more useful. I disagree. First of all, the laptops are communications devices, so they could be used to solve many of the same problems that phones or radios could. Secondly, they have much broader uses in education: not only can students write assignments on them, but they can also be used to access web resources or resources that their teachers make available to them.

But the real kicker is the third use: they can write and publish their work on them. There are, of course, difficulties with that—someone is going to have to make webhosting available to them, or the information will only be passed laptop to laptop. But I predict that someone will make the hosting available.

What will they publish?

I have to confess to a selfish motivation for wanting to see the $100 laptop project succeed: I’m looking forward to reaping the positive externalities of a 100 million new GNU/Linux users suddenly being able to publish. Oh, sure, I imagine that 99.9% of them won’t do anything with that power, but that still leaves 100,000 people! All of them GNU/Linux users, I might add. Who knows what itches they might choose to scratch or what insights they will have.

The biggest horror of poverty for me is not the personal deprivation itself, but the horrific sense of wasted human potential. Building these computers and putting them in the hands of people who have had so little power will be a remarkable step forward. It’s not about some kind of weird idea of charity or some kind of magnanimus gesture—it’s about putting tools in the hands of people who can use them. It’s about taking a gamble on the ingenuity of people all over the world by making a real investment in them.

It’s not going to happen overnight of course. The OLPC project is mostly about getting the technology to kids. We won’t see the real benefits until they start to grow up. Investments are like that. But delay or no, this is not a one-way transaction: we in the developed world will get something back for this effort. I think it’s important to realize that this is not some hand-out carrying an implicit obligation, this is a case of developing countries helping themselves: they are paying the bill to make something that their people need. The fact that the contractors and engineers are in “developed nations” is just a sign that we actually have something they need, which shouldn’t be a surprise to anyone (one might also regard Taiwan as still being pretty close to a “developing nation” itself, though I’m not sure who gets to decide when a nation is finally “developed”).

We in the free software community really ought to be more excited about this. Because if it does work, it will work partly because of the success of free software, and the people using it will know that (or learn it, or even take it for granted). Most of them, just like most people everywhere else, will not do anything about that; most won’t give back to the commons—but I think some of them will. And with the enormous numbers of people that this project will effect, even a very tiny fraction could have world-changing effects in our tiny global community. I don’t have any real idea what those effects will be, but I’m looking forward to finding out.

Category: 

Author information

Terry Hancock's picture

Biography

Terry Hancock is co-owner and technical officer of Anansi Spaceworks. Currently he is working on a free-culture animated series project about space development, called Lunatics as well helping out with the Morevna Project.