It seems like it’s “in” right now to express your disdain for “tech bros.” Maybe AI has something to do with it. Maybe it’s because high-profile tech CEOs keep doing increasingly stupid shit while the cults of personality surrounding them raise a firm salute. Maybe it’s because we’re all slowly realizing we’ve had our brains melted by the attention economy, with social media giants selling our attention spans for a couple bucks a pop to unregulated ad firms who productize our PII for anyone else who wants a slice of the brain-rot pie.

Naaaah! That couldn’t be it.

Silicon Valley wasn’t always the progenitor of dystopia that it has become over the last 10-ish years. The persona we’ve named “tech bro,” however, has been around in various forms since time immemorial. He wasn’t as destructive in decades past, but he was just as delusional as he is today! Seriously, take a peek at this 1995 special on Virtual Reality from BBC’s The Net. Droves of the most technically gifted folks on earth convinced themselves that you’d want to check your email by walking around in a virtual 3D world. They’re still chasing this idea.

The self-described visionaries of tech make a habit out of getting high on their own supply, and it’s been that way for longer than I’ve been alive. I get why it happens, though. Try to imagine what it would be like to live in a socioeconomic Simpsons Dome where seemingly infinite sums of money are pouring in, tech stuff goes out, and you and your friends keep getting wealthier and more famous. It’d be easy to conclude that, through technology, all the world’s problems can be solved. You’d reach that conclusion because technology solves all of your world’s problems.

Say hello to the One Laptop Per Child (OLPC) XO. It’s thoughtful, visionary, even inspired… and it solved none of the world’s problems.

One Laptop Per Child was a nonprofit founded in 2005. Their goal was to deliver a laptop to every child in the developing world. The idea was that the laptops would enhance the education of students in developing countries, helping them to lift themselves out of poverty. The XO was their first attempt at fulfilling this vision, and it went poorly. Everything imaginable went wrong; some things I couldn’t have imagined went wrong too.

The XO is not the only computer OLPC ever made, but it’s certainly the best-known of the bunch, at least among folks in developed nations. The XO is the first machine OLPC ever brought to production, and its release marked the peak of OLPC’s cultural relevance. No OLPC device ever had higher expectations placed upon it.

In 2005, this stout, colorful laptop was a symbol of optimism; optimism that we could end poverty, and provide a quality education to every child on earth. By 2008, the XO had evolved into a symbol of OLPC’s failure to fulfill the utopian vision they had set forth. The XO hasn’t been forgotten, either. Here’s one I saw in the Milwaukee Art Museum:

Here’s the text on that card:

The toylike appearance of this small laptop reflects its designers’ goal of creating an educational tool for children in the developing world. Though the project was met with numerous challenges - including how to reliably power the devices in communities lacking consistent electricity - it sparked important debates about the role of design and technology in society.

I appreciate that MAM thought to preserve this weird little piece of computer history, but their description raises more questions than answers:

1. Why did they point out the power situation specifically? Was this really the biggest challenge the XO faced?
2. How did they overcome the power problem, assuming they did at all?
3. What were the other “numerous” problems?
4. Who participated in that debate? Were there any more specific topics of debate?
5. Why did this particular laptop trigger that debate?
6. What’s the XO doing in an art museum? Is this laptop art? Can any laptop be art?

I have spent the last 3 months searching for answers. Thankfully, much has been written about OLPC and its XO. However, some of these questions still required me to lean on primary sources. I am not a historian by any reasonable definition of the term. I’m just a guy with a goofy little laptop, and a whole lot of questions about it.

To answer these questions, I did what any reasonable person would do: I bought myself an XO on eBay. At first, my intent was to offer a formalist look into the XO’s design. In other words, I hoped to separate the art from the artist. By separating my evaluation of the XO from my evaluation of OLPC as a whole, I hoped to give the XO a fair shake. I also felt that this would contrast nicely with the existing historicist corpus which focuses heavily on the history of One Laptop Per Child (OLPC), discussing the XO only as necessary to study OLPC.

I quickly realized that my goal of separating the XO from its creator was impossible - or, at least, the resulting writeup would have been disingenuous. OLPC and its XO are tied together at the hip. Their legacies are inseparable. OLPC’s ideology is reflected in the design of the XO, and the XO’s physical characteristics are essential to any explanation of OLPC’s meteoric rise and catastrophic failure.

The State of Tiny Laptops Circa 2006

I have spent many hours using my XO now, and I can attest that it’s nothing like any other PC I’ve ever used. Usually, PCs vary only in matters of era and scale - older, newer, faster, slower, bigger, smaller, and so on. The XO is different. There are general hints of a normal laptop: the clamshell chassis, a keyboard, a trackpad. But, even those superficial similarities come with obvious deviations from the norm. Using the XO is an alien experience.

Small, affordable, low-power laptops have been commonplace for years, so why is the XO so weird? What’s up with the goofy color scheme and the Shrek ears? Why does it have a D-pad next to the screen? Are those… USB and audio jacks on the top half of the laptop? The weirdness doesn’t stop on the outside, either, but we’ll put a pin in that.

There are technical reasons for many of the XO’s quirks, but much of the XO’s weirdness can be explained with some historical context. When the XO was designed, there was no such thing as a cheap, tiny, ultra-low power laptop.

There were a few tiny, low-power laptops that predated the XO, but they’re all rare birds like the Sony Vaio C1. These primordial netbooks usually cost more than $1000, and they were never high-volume sellers.

The original netbook, the ASUS Eee PC, was released in November 2007. Netbooks were the cheapest mass-market laptops of their time, and they were definitely used as educational devices: I remember using an Eee PC at an IMSA summer camp in the late 2000s. But, the Eee PC doesn’t predate the XO. In fact, the XO and Eee PC were released the very same month.

Nowadays, a Chromebook can easily hit the sub-$200 mark even before you account for 15 years’ worth of inflation. However, Chromebooks wouldn’t appear until the early 2010s, and they didn’t become a ubiquitous educational tool until a decade after the XO’s release. In short, the XO was the first laptop of its kind.

OLPC’s Mission

OLPC’s mission, as it was presented to the public, seems pretty straightforward: enhance the education of kids in developing countries by providing them with affordable, purpose-built laptops. In an incredibly 2007 move, one of OLPC’s first endorsement deals was Heroes star Masi Oka. Here’s what Masi has to tell us:

Imagine that you give a laptop to a child. That one child can come up with these great ideas; you give them the power to learn, the power to connect with ideas all around the world. Imagine the difference that we can make.

This 2008 video presents OLPC’s mission in even simpler terms:

We want to create education opportunities for the world’s poorest children by providing each and every one with a rugged, low cost, low power, connected laptop.

Great! That’s a succinct, well-defined mission statement with an end (education opportunities) and a means (purpose-built laptops). Seymour Papert’s interview for OLPC presents similar sentiments. From their mission statement and most of their advertising, one would think OLPC had a solid, well-defined vision for the XO. All of this is a façade.

OLPC’s Actual Mission

The concept for the XO originated from constructionism, a theory of education pioneered by the aforementioned Seymour Papert. Constructionism is a theory of learning wherein students work on projects that require them to apply information they already know in order to learn new things. It’s learning by doing, through trial and error.

I had never heard of constructionism until I embarked on this research project, and I was shocked to discover that it has deep ties to many of my fondest childhood memories. If you participated in FIRST Robotics like I did, you’ve done constructionism. I had a LEGO Mindstorms robotics set, too; those are named after Mindstorms, Papert’s constructionist manifesto.

Papert frequently voiced disdain for the status quo of education in the United States and abroad. Nicholas Negroponte, OLPC’s founder and mouthpiece echoed Papert’s criticism, but also dove into insults and condescension. At various points throughout the 2000s, Negroponte accused teachers in developing countries of being lazy, incompetent, and drunk (5). In this YouTube video, Negroponte lays out his vision:

Why would a kid in the developing world need a laptop of all things?… Take the word “laptop” and substitute it with “education” and nobody would say that. This is probably the only hope - I don’t want to place too much on OLPC - but if I really have to look at how to eliminate poverty and create peace, and work on the environment, I can’t think of a better way to do it.

Okay, so the laptops aren’t just providing education opportunities like the previous video indicated. Apparently, they’re going to eliminate poverty, create peace, and save the environment. Oh, and they aren’t just providing education, they are interchangeable with education.

This is all pie-in-the-sky nonsense, obviously, but Negroponte was a culture setter. He was an early investor in Wired, and the founder-turned-director-turned-chairman of MIT Media Lab. His words, no matter how outlandish, were backed by serious cultural caché. If Negroponte said computers would transform and supplant many aspects of education, who among us would have argued with him?

In his holy war for constructionist philosophy, Negroponte saw the XO as his greatest weapon. On multiple occasions, Negroponte referred to the XO as a “Trojan horse,” a device that would allow this hot new educational philosophy to infiltrate schools that adopted the XO (5).

The Idea of the XO

In his early presentations of a mockup XO, Negroponte promised that it could be charged with a simple hand crank, and it would cost only $100. The price point was so essential to Negroponte’s pitch that, before it was named XO, OLPC’s laptop was simply called “The $100 Laptop.”

Negroponte’s habit of breathless promises was bad enough that it didn’t go unnoticed even at the time. After that 2005 presentation, David Kirkpatrick wrote in Fortune Magazine about Negroponte’s “infectious bluster,” saying that “the impediments [to Negroponte’s goals], needless to say, are numerous and daunting.”

According to Kirkpatrick, Negroponte didn’t just promise a hand crank and a $100 price tag. He promised a laptop that would last for days on a single charge. Negroponte envisioned a computer that an illiterate kid could learn to use, and then use the computer to learn to read.

Throughout the years, Negroponte made a variety of other promises of the XO. He promised that it could be dropped out of a helicopter without breaking. He promised mesh networking, which would allow the laptops to communicate with each other without Wi-Fi or any other networking infrastructure. He promised that, if one XO had an internet connection, nearby XOs could get online by using the mesh network to piggyback off that one internet connection.

All of this was tantalizing to contemporary nerds, and the idea of the XO was very well-received by the public. In December 2006, Popular Science handed XO the highest award in their “Best of What’s New” edition, about 11 months before XOs starting getting delivered to customers. In typical PopSci fashion, the entire article is two paragraphs of rote description of some of the XO’s unique hardware features, with only a single sentence providing a half-baked summary of OLPC’s mission.

Part of me wonders if the empty-calories writing style of popular tech media in the 2000s helped to facilitate OLPC’s various screwups. I imagine that the public’s perception of OLPC would have been more critical if PopSci had mentioned that OLPC’s founder had some pretty radical ideas about the future of education. Regardless, public perception of OLPC was overwhelmingly positive long before a single XO had made it into a child’s hands.

OLPC was so confident in the XO’s abilities that they were allergic to any sort of field testing. In a 2006 TED Talk, Negroponte said this (5):

The days of pilot projects are over. When people say “Well, we’d like to do three or four thousand in our country to see how it works.” Well, screw you. Go to the back of the line and someone else will do it, and then when you figure out that this works, you can join as well.

Negroponte’s vision was ambitious in just about every imaginable way. Not only did he want to stuff tiny laptops with novel technology, he wanted to sell them for a quarter of the price of any other laptop on the market, and he wanted to make boatloads of them - hundreds of millions, apparently. He also wanted to reinvent the very practice of education in developing countries, where both physical and political infrastructure to facilitate education is most fragile, and he wanted to do it with no small-scale testing ahead of time.

This went about as well as you’d expect.

Floundering

In April 2008, a year and a half after PopSci showered the XO with praise, they featured it again. This time, the title is “The $100 Headache” and the accompanying illustration is hilarious:

OLPC simply couldn’t make laptops fast enough or cheaply enough to keep up with Negroponte’s promises. According to PopSci, they had only shipped 200,000 laptops abroad, with another 80,000 laptops going to folks in developed countries as part of OLPC’s “Give One, Get One” campaign.

OLPC had promised a $100 laptop, but the XO ended up being priced around $190. PopSci, among others, was critical of OLPC for this. In OLPC’s defense, though, the XO was still priced incredibly low.

OLPC blew a hole in the price floor for laptops. Exact accounts of the XO’s original selling price vary a bit, but all agree that it was between $180 and $200. That’s less than half the price of any contemporary Eee PC, and it’s also about $50 less than Intel’s Classmate PC, which was released a few months later. The XO’s price point was incredibly low compared to contemporary laptops - the only problem with the XO’s price is that it didn’t match Negroponte’s promises.

There were bigger problems than the price, though. Key features of the XO demo never made it to market. The hand crank that purported to solve the power problem? Nowhere to be found. The lack of a hand crank wasn’t a shortcoming on OLPC’s engineers’ part, though. It was physically impossible - it had never been possible.

Let’s Do Some Math

The XO is a remarkably low-power device, even compared to contemporary netbooks. But, it still uses enough power that a hand crank is infeasible as a charging method. To prove that, all we need is some high-school physics.

My XO seems to charge reasonably quickly from a 12-Watt supply, so let’s try to generate 12 Watts using a hand crank. Electrical stuff is never perfectly efficient, so let’s say that 80% of power actually makes it to the laptop. With 80% efficiency, we actually need to generate 15W, not 12, because 12 / 0.8 = 15. This is very generous, by the way - a more realistic efficiency number would probably be about 50%.

Cool, so we need 15 watts. Let’s make the crank 10 centimeters long, and let’s say that a kid can reasonably turn the crank at one full rotation per second.

With a bit of plugging and chugging, that’s… 2.4 kg (5.4 lb) of force to turn the crank. That seems plausible, even if it might be tough for a little kid to do that for more than a few minutes at a time. But, the XO’s battery holds 20.2 Watt-Hours. In order to fully charge its battery, you would have to turn the crank 6000 times, working continuously for an hour and 40 minutes. Obviously, this is ridiculous.

Anybody who remembers their high school-level physics can prove that the hand crank was preposterous. Negroponte ran MIT Media Lab! He’s a brainiac! He was surrounded by other brainiacs! I’m floored that OLPC pushed this as hard as they did.

Unmet Promises, Continued

OLPC had promised that the XO would support mesh networking, and to my surprise, my XO’s “neighborhood” menu shows options for connecting to mesh networks. Although you can buy cheap devices that support Wi-Fi mesh networking nowadays, you still won’t find it in most folks’ homes in 2024. Out of all the promises OLPC made, I was shocked to see that they managed to ship one of their most technically ambitious features.

Remember that thing about piggybacking off one XO’s internet connection? The technical term for that capability is “backhaul,” and it still isn’t a totally solved problem in 2024. I was ready to shower OLPC with praise for their astonishing achievement …

…until I did some research and found out that it doesn’t work. The mesh networking feature never worked properly and OLPC released an update in 2010 that dropped support for it entirely (5).

Battery life didn’t meet expectations either, although it was very impressive for its time. My XO holds a charge for 2 hours and 40 minutes while I’m connected to Wi-Fi and browsing Wikipedia with the screen brightness at roughly 50%. That’s phenomenal performance when you consider that my XO’s battery is probably about 15 years old. I imagine that, when new, my XO’s battery would have lasted 5-6 hours under the same conditions. That’s enough to get through the school day, although it would have been much shorter if kids were watching videos or playing games. Even if the XO’s battery life was excellent by contemporary standards, it never matched Negroponte’s promises.

The XO still has many special features optimized for its target use case. Many of those features are utterly singular, even 15 years later. The only problem is that it didn’t have all the features Negroponte promised. I hope you’re noticing a pattern.

Deployment Woes

I can’t personally speak for the way the XO held up in developing countries, but Morgan G. Ames certainly can. In her book The Charisma Machine: The Life, Death, and Legacy of One Laptop per Child, Ames spends most of a chapter describing her firsthand observations of a deployment of XOs in Paraguay. The picture she paints isn’t pretty. Here’s a quick summary:

The hand crank was sorely missed. Classrooms often lacked power outlets, and you couldn’t assume an outlet would actually work when you plugged something into it. Unfortunately, the promise of hand cranks and tremendous battery life had informed the government’s decision to not install more power outlets ahead of the XO’s arrival.

Durability was an even bigger problem. Negroponte had promised an extraordinary level of durability, going so far as to chuck an XO across a stage during a demonstration. Because of OLPC’s durability claims, early adopters like Paraguay had not foreseen any need to educate teachers and students on proper treatment of the XO. The results were disastrous.

Note the little green “antennae” on the top of the laptop. Those don’t contain an actual wireless antenna, but they are the latching mechanism that holds the screen closed. If you try to grab an XO by its antennae, they break off pretty easily. The XO has a proper built-in handle, of course, but I can imagine why a child, with child-sized hands, would see the antennae as an equally appealing grip. Since the antennae serve as the latching mechanism that holds the laptop closed, an XO with broken antennae becomes much more susceptible to screen damage.

According to Ames, the epidemic of broken XOs was so severe that governments responded by developing direct relationships with Quanta, the company that built the XO for OLPC. By working directly with the manufacturer, they could get parts quickly and cheaply. This sounds like a good solution, but remember that these repair programs were developed after XOs had already started breaking down at a rate that justified a national response. These repair programs would inherently have started off on the back foot - their backlog must have been nightmarish. By Ames’ account, these national repair efforts never caught up, and broken XOs usually outnumbered working ones in any given deployment.

Defects and Quality Problems

Power, durability, and lack of support weren’t the only problems Ames observed. The XO was also outright buggy. Ames mentions that the trackpads were flaky. They’d inexplicably stop working for a minute or two, then go right back to normal. My XO exhibits a behavior that matches her description. Every once in a while, the cursor will shoot off to some random spot on the screen, and then when you try to start using the trackpad again, the cursor moves around seemingly randomly. My wife had an XO when she was a kid thanks to OLPC’s “Give One Get One” program, and she remembers her XO doing the same thing.

The software had issues, too. A couple of times, while launching a program, my XO has shown me a cryptic error suggesting that it was out of storage space:

I checked, and the 1 GB of flash storage was never more than halfway full. After I dismiss the error message, the program appears to be loaded correctly, but it doesn’t respond to any user inputs. The only solution is to close that activity and try opening it again.

My XO appears to have its stock software loadout, and it came with a bevy of programming environments. One of those is Pippy, a stripped-down Python coding environment with a bunch of demo programs. One of these programs, named “slideshow,” simply doesn’t launch on the XO at all. Instead, it hangs for a moment, then dumps the following error:

assertion error 361:oil_test_check_impl() illegal instruction in mmxCombineAdd()

The CPU supports MMX, so I have no idea why this error is being thrown. Weird.

I also noticed that you have to wait for the machine to fully shut off before closing the lid. If you tell the laptop to shut down, then immediately close the lid, the shutdown routine pauses until you open the screen again. The next time you open the XO, it will resume shutting down, and you’ll have to wait until it’s done before you can boot it up again. I faintly recall my Linux laptops doing this at some point, but not in the last 10 years. In any case, I imagine this issue caused many unexpected dead batteries.

By mid-to-late 2008, OLPC was floundering. Laptops weren’t being built as quickly as they promised, uptake among developing countries was slow, and quality problems ran rampant among the handful of countries that had signed up. Worst of all, Intel’s Classmate PC hit the market, and it competed for the same market OLPC envisioned for the XO.

The remaining history of OLPC only gets sadder and sadder. If you’re interested in that, you should buy a copy of Morgan G. Ames’ book (ISBN 0262537443, 9780262537445). The Charisma Machine is a comprehensive exploration of constructionism, utopianism, and “tech bro” vanity, all of which are important to understanding how OLPC got into a position where they could fail so spectacularly.

Catastrophic Failure

Some sources, like this World Bank blog post, plainly state that the XO failed to improve educational outcomes in developing countries, although it incidentally succeeded in teaching computer literacy. To support that argument, World Bank cites this paper (Malamud and Pop-Eleches, NBER Working Paper 15814, Home Computer Use and the Development of Human Capital).

Here’s the problem: Malamud and Pop-Eleches state in the above paper that access to a computer doesn’t correlate with improved educational outcomes. It makes no such correlation with the XO in particular. No matter what you think of OLPC or the XO, it’s incorrect to use this paper as evidence that the XO specifically didn’t improve educational outcomes. This distinction is important because the XO is fundamentally different from virtually every other mass-market PC ever made.

I maintain that, if the XO hadn’t been smothered by Negroponte’s impossible promises and OLPC’s mismanagement, it had the potential to affect academic performance in a way that diverged substantially from a normal PC.

The Case of Uruguay

Uruguay was one of OLPC’s earliest adopters, placing their first order in October 2007. The country also ended up being OLPC’s largest single customer for the XO, ordering 300,000 units before moving on to other equipment.

Uruguay’s massive deployment of the XO makes them an excellent case study in its effects on educational outcomes. Fortunately, the Institute of Economics at Uruguay’s Universidad Nacional de Catamarca researched this matter exhaustively, publishing their findings in this paper (De Melo, Machado, Miranda and Viera, August 2013, Profundizando en los efectos del Plan Ceibal).

I had to rely on machine translation to read this paper. If there are any Spanish speakers out there who’d be willing to fact-check me, I would be very grateful. You can shoot me an email here.

De Melo et al. correlate schools’ test scores with the number of days students had been exposed to XO laptops. The study follows the same set of students over a 6-year period, sampling their test scores in 2006, 2009, and 2012. They compare that data to control groups (schools with no XO laptops), and they consider other factors known to influence academic performance, like homeschooling, access to potable water, and access to books in the home.

Unfortunately for the XO, the conclusion is bleak. No matter how they cut the data, researchers found no correlation between use of the XO, and literacy/mathematics outcomes. Other partner countries reached the same conclusion (7). There’s more to the story, though.

The researchers ran a survey to compare the use of XOs in the classroom, and saw a precipitous decline from 2009 to 2012. In 2009, 41.5% of respondents reported that they used the XO “every day or almost every day” in their language classes. In 2012, that number dropped to 4.1%. Likewise, 31.8% of students reported that they used the XO “every day or almost very day” for mathematics in 2009, but that number dropped to 6.1% in 2012. Meanwhile, the number of students reporting using XOs “one to three times a week” remained relatively stable in both categories, going from 35% to 32% for language and 24.9% to 28.4% for mathematics.

Since this is a longitudinal study, the researchers frame this survey’s results as a decline in usage as the students progressed from primary to secondary school. This is a very diligent, academic way of looking at things. But, I’m not an academic, so here’s some speculation.

My Achey Breaky Laptop

Ames told us that XOs quickly began breaking down after they were delivered to students. Over time, national repair efforts proved insufficient. As a teacher, the longer you watch this debacle play out, the less you would trust the XO as a classroom tool. Electronic teaching tools must work properly, 100% of the time, the moment you switch them on. Anything that falls short of this requirement is a waste of the limited hours in a school day, and a waste of the limited attention span of the students. My hypothesis is that the decline in XO usage had two key causes:

1. A lot of XOs were broken, so those couldn’t be used at all.
2. Teachers lost confidence in the XOs that weren’t broken.

Figure 8 of De Melo et al. shows that, in 2009, XOs were used less frequently in schools with “very favorable” sociocultural environments (14.04% daily use for language, 5.08% for math) compared to the overall population (41.5% daily use for language, 31.8% for math).

There are many reasons why this could occur. De Melo et al. attribute this to an age difference between teachers, explaining that older teachers in more favorable environments would be more apprehensive to incorporating technology into teaching. To me, this doesn’t pass the smell test.

The researchers show that “very favorable” sociocultural environments used the XO 2.5-6x less than average. The researchers provide no evidence of a gap in teachers’ ages, let alone a gap so massive that it would explain this vast of a difference in the way the XO was used. Every other sociocultural tier used the XO much more often, and all of those less-privileged groups were much closer to the mean.

The epidemic of broken XOs spread fast, limiting its feasibility as a classroom teaching tool. By 2009, the first of Uruguay’s XOs had been in kids’ hands for 2 years, so the plague of broken XOs had certainly taken off by then. Since the XOs were constantly broken, it stands to reason that teachers with access to quality education tools would more readily fall back on alternatives, while teachers without a superior alternative would be stuck with their half-broken fleet of XOs.

There are reasons to doubt my hypothesis. Uruguay had substantially better educational infrastructure than Paraguay, so Ames’ account of Paraguay’s repair issue might illustrate a worse situation than Uruguay would have experienced. De Melo et al. show the 2006/2009 student questionnaire at the end of their paper, and it includes questions that would allow us to see if broken XOs were as common in Uruguay as Ames says they were in Paraguay. For whatever reason, the researchers don’t include the results from these questions in the paper.

Ames also cites training problems: OLPC barely provided any training programs, so a lot of teachers didn’t go to the trouble of incorporating an unfamiliar device into their classes, and many of them assumed that the XO was a toy. That’s understandable, given its appearance. But, in my opinion, the wild anomaly in XO adoption rates among relatively-privileged schools suggests that the repair problem played an outsized role in the XO’s failure.

If my hypothesis is true, and declining utilization of the XO was caused by repair and maintenance problems, then we’ll never know what impact the XO could have had. Negroponte’s misleading messaging and OLPC’s lack of field support made the XO impossible for teachers to trust. Environments with the necessary cultural and economic tools to maximize the XO’s potential were necessarily the same environments that would be inclined to abandon their XOs when they started breaking down.

The Tragedy of the XO

This could have been prevented. The XO is a well-made device. It’s much more rugged than a normal laptop. It feels more robust than plenty of contemporary laptops did. Plenty of XOs have survived to the present day even when used pretty heavily - just look at my XO. Keynote speech showboating and marketing hype drowned out any sensible voices in the room who could have said “hey, maybe we should teach kids how to care for these things.” OLPC left the developing world with a boatload of broken laptops, and not much to show for it.

It’s definitely not Uruguay’s fault, either. Their Plan Ceibal delivered broadband internet access to nearly every student in the country in only a couple years. They had proven their ability to execute on a project of this scale, but OLPC set them up for failure.

Like my Silicon Valley strawman from earlier, Nicholas Negroponte was high on his own supply. OLPC over-promised and under-delivered to an extent that would offend even the most zealous Web 3.0 blockchain-something-or-other startup. The XO was not doomed by any particular shortcoming in its design, but by the outlandish promises OLPC made of it.

The XO was delivered late, cost more than originally promised, and lacks essential features from the demo. The XO never met its sales goals, and XO owners weren’t very happy with them. The largest adopter of the XO, Uruguay, didn’t see measurable improvements in key educational outcomes.

The XO is, in no uncertain terms, a failed product. But, its greatest sin was not its own. The XO’s greatest barrier to success was the organization that created it.

The Computer Itself

I’m not sure where my XO came from. At least one batch went to New York City public schools, and I can’t imagine they’re the only American school district who had fleets of XOs. Another guess is that my XO was part of OLPC’s “Give One, Get One” program. G1G1 was a promotional campaign where you could pay for two XOs, receive one, and another would be sent to a developing country. At least 80,000 XOs were delivered to folks in developed nations through the G1G1 program, so we can be sure that a lot of the XOs floating around on the American used market were sold directly to end users instead of institutions.

I have reason to doubt that my XO was a G1G1 device, though. The eBay seller had over 40 of them for sale, all shipping from the same location. I asked the eBay seller where they got their XOs, but they wouldn’t tell me. I have a feeling my XO was part of a fleet that was used by a school, then sold at a public auction.

Mechanical Design

If I’m right that this XO was used in a school, then it held up well to the rigors of the American classroom - far better than the horror story Ames describes in Paraguay. It has definitely taken a beating: the bottom is scuffed up, and the plastic shell has a crack in one corner. It still looks much healthier than some of the Chromebooks I’ve seen at public auctions.

Would I want to try dropping my XO on the concrete floor of a classroom in a developing nation? Hell no! It’s chunky and ruggedized, but it’s not that ruggedized. The handle is sturdy enough to make the XO feel eminently chuckable, but not sturdy enough that I’d actually try.

The closer you look at the XO, the weirder it gets. Notice how the top part of the clamshell is really chunky, and the bottom is super thin? The entire computer is in the upper part of the clamshell, and the bottom contains only the battery, handle, and keyboard. As a result, the entire machine is rather back-heavy when open. Thankfully, the XO’s handle sticks out the back of the laptop, keeping it stable. When using the XO on my lap, however, I feel the need to lift my knees a bit to make sure it doesn’t tilt back.

The XO’s “antennae” don’t contain any electronics. They’re just a whimsically-designed locking mechanism that holds the screen closed during transport. They also serve as a cover, protecting the XO’s USB and audio ports when the laptop is closed. As mentioned before, though, the antennae were a critical point of failure, despite their cute looks and clever port-protecting abilities. The latching mechanisms on contemporary laptops were relatively boring, but far less likely to break.

The XO’s screen also swivels like a tablet PC. I’m not sure why OLPC bothered with this, since the XO doesn’t have a touchscreen. The screen is totally exposed to the elements when the screen is flipped around, and I imagine a lot of screen breakage happened in this configuration. It also seems like an awfully complex (read: expensive) mechanism to include in a device that was to be priced as cheaply as the XO.

The buttons flanking the screen prevent this feature from being entirely pointless, though. The directional buttons to the left of the screen are mapped to the arrow keys, and the buttons to the right of the screen are mapped as follows:

• Circle: Page Up
• Check Mark: End
• X: Page Down
• Square: Home

This is a sensible combination of keys if OLPC assumed the XO’s “tablet mode” would be used to read books. It just isn’t enough to justify the cost and complexity of the big rotating hinge.

PC-ish things

The XO is, fundamentally, a PC. It has an x86 CPU, some RAM, some storage, and some ports. Boring. It does get more interesting when you look closer, though.

Inside the XO you’ll find an AMD Geode processor, a totally forgotten line of ultra-low power CPUs that AMD acquired from Cyrix, and then discontinued a couple years later. As a 433MHz 32-bit CPU, the XO’s Geode processor is less like an Intel Atom CPU, and more like a Celeron… from 1998. The XO is equipped with a paltry 256 MB of RAM, which would also have been normal for a computer from 1998. The 1 GB of storage on the XO, though? That’s more like 1995.

In the XO, OLPC created a laptop with performance that was a decade out of step with contemporary consumer PCs. Nowadays, a 10-year-old laptop can run everyday tasks without breaking a sweat. It wasn’t always like this, though. In a world of Core 2 Duos and Athlon 64s, the XO’s Geode was an anchor. The XO is glacial compared to contemporary laptops. But, man, it sure was cheap!

Despite its slowness, the XO is totally adequate for a variety of education-adjacent tasks. For example, it’s a decent word processor. It shipped with a heavily modified version of AbiWord with all advanced functionality stripped out. You get basic text formatting, pictures, tables, and that’s about it. My XO did not come with a spreadsheet program, though my research indicates that a compatible spreadsheet program must be out there, somewhere.

Ames and De Melo et al. agree that the most common use of the XO was web browsing. Predictably, my XO’s preinstalled web browser is a lightly-modified copy of Firefox. Firefox has been losing market share for most of the last decade, but in the late 2000s it was the browser of choice among folks with enough computer literacy to realize that Internet Explorer was terrible. Firefox is a boring choice, and that’s exactly why it was the right choice.

The XO’s selection of IO ports is better than I expected. It has a headphone jack, a microphone jack, and three USB 2.0 ports, equalling business laptops of the era. Of course, contemporary business laptops would also have a VGA port, an Ethernet port, a built-in modem, plus a dock that would have a second VGA port, even more USB ports, parallel and serial ports, and maybe even a DVI port… but let’s not dwell on any of that.

The XO doesn’t have any ports where you could connect an external monitor, so I can’t attach a screen-capture device to it. Its built-in keyboard doesn’t have a Print Screen key, and it doesn’t seem to respond to the Print Screen key on an external keyboard. Now you know why this blog post is full of crappy pictures of the XO’s screen - sorry.

The Super Secret SD Card Slot

It took me a while to realize it, but the XO has an SD card slot! It’s not on the side of the laptop with the other ports. Instead, it’s buried under the screen. See, it’s right there:

The SD card slot should have served an extremely important purpose on the XO. Ames speaks at length about the problems caused by the XO’s extremely limited 1 GB of storage. Some kids were constantly deleting stuff, including classwork, off their XOs to make room for new stuff - particularly games and TV shows. The handful of kids who showed an affinity for OLPC’s constructionist vision would have quickly lost interest when they had to delete their projects. The SD card slot should have solved this problem.

After a little digging through PC Mag‘s public archives, I get the impression that 2 GB SD cards were becoming commonplace in the mid 2000s, and smaller cards were widely available. Since the OS and preinstalled programs take up about 400 MB, a cheap 512 MB card would have almost doubled the usable storage space on the XO. Yet, I didn’t find a single mention of the SD card slot in Ames’ entire book. I can only conclude that she never encountered a single teacher or student who used it.

Given how OLPC provided very little (if any) training on how to use the XO, it’s possible that nobody even realized the SD card slot was there. Seriously, if you had a laptop with an SD card slot buried in the seam between the screen and the base, would you have noticed it? I bought an XO for the sole purpose of studying it, and I didn’t notice it for over a month!

The XO is chock-full of peculiar design choices, but the decision to hide the SD Card slot defies explanation. It’s profoundly stupid. OLPC could have taken feedback from teachers and students, and shipped out SD cards. They were funded by Google and AMD - they definitely had the money to do this. Even if they somehow didn’t, they could have talked to the national programs overseeing the XO’s deployment, and suggested they purchase some SD cards.

Add “limited storage” to the list of problems that were utterly preventable, but kneecapped the XO anyway.

Battery and Charger

As I mentioned before, my XO gets about 2 hours and 40 minutes of battery life under a light Wikipedia-surfing load. This certainly comes nowhere close to the multi-day life promised by Negroponte in 2005, but it’s double what I’ve gotten out of the original battery on any other 15+ year-old laptop.

I didn’t get this laptop with a charger, and I didn’t need to. That jack on the side? That’s a garden-variety center-positive 0.21” DC barrel jack, the exact same kind used by millions of common consumer devices. OLPC’s wiki claims that the XO will charge from any 11-18 Volt supply. They don’t specify any particular wattage, which is silly. I charge my XO with a 12V, 1A supply that I think was intended for a contemporary Linksys router. The battery charges quickly even while the machine is being used.

OLPC didn’t solve the “no electrical service” problem the way they said they would. However, the XO can still be charged relatively cheaply and easily compared to other laptops. It’s not fair to expect that an average Joe in a developing country has a random wall wart laying around, but this is a much better solution than a proprietary power jack. 10 years later, with the rise of USB-C chargers, mainstream laptops finally began to catch up with the XO’s common-sense approach to charging.

Keyboard

This is, beyond any shadow of doubt, the worst keyboard I’ve ever used in my life. I’m not mad at it, though.

The XO’s keyboard consists of a rubber membrane with little bumps for the keys. This keyboard feels like the rubber buttons on a TV remote, but softer and less distinct.

Doing any significant amount of typing on the XO is a chore. There is no feedback to separate a good keystroke the XO will register, from a bad keystroke that it won’t. Smaller keys can be triggered by bumping them on their edges, but larger keys need a downward press from the top of the key.

The keys are tiny, too. The alphanumeric keys have less than a quarter of the surface area of the keys on the ThinkPad keyboard I’m using to write this post. Granted, the XO is small enough that full-size keys were out of the question. But, there is a lot of space between the keys, and that isn’t being used for anything - I’m guessing this is an innate limitation of the membrane design.

In fairness to OLPC, this keyboard wasn’t meant for me and my big fat stupid adult hands. A child’s smaller hands might find it much easier to hit only one key at a time. The membrane has no gaps where dirt or moisture could get into the machine, which is important for a device that will likely spend a lot of time outside.

Most importantly, this design is probably a lot cheaper to manufacture than a normal laptop keyboard. Given OLPC’s price target, it’s possible there simply wasn’t another design that fit their cost requirements. There’s also no imaginable way that this keyboard could ever break, short of outright abuse. Early XOs had issues with the membrane breaking down, but OLPC fixed it on future models. None of my research suggests that the keyboard was ever a source of widespread maintenance issues.

After several hours using the XO, I’ve gotten more comfortable with its keyboard than I expected to. I still type slowly, but I’m reasonably accurate and I don’t have to look at the keyboard while I type. For as dreadful as the keyboard is when compared to a conventional laptop keyboard, there are a lot of good reasons why it was designed the way it was.

Display

Take a peek at this photo of the XO with its word processor loaded:

Take a close look at that screen - I encourage you to click on the photo to view it at full resolution! Note how the text is crisp, and the contrast is pretty good. Did you notice that I pointed the screen into the sunlight? If the XO were a normal laptop, you wouldn’t be able to see a single word on that screen. Thankfully, the XO is not a normal laptop.

This is not an e-ink display. Yet, there’s the Bee Movie script, plain as day. My choice of camera angle was adversarial, too.

This is called a Transflective Liquid-Crystal Display, and you’ve seen them before - LCD watches use them, among other things. However, unlike a watch, the XO’s screen has a white backlight and a full-color panel.

Crank up the XO’s display brightness, and you’ll get a bright full-color image with quality comparable to other cheap mid-2000s laptops. Carry the XO out into sunlight, though, and without adjusting the display brightness, the very same image shows up in grayscale with phenomenal contrast. I actually enjoyed writing that Bee Movie crap outside in direct sunlight, despite the horrendous keyboard.

I have no idea why this technology didn’t end up in every laptop. The Wikipedia page for Pixel Qi, the company that created this display, indicates that these screens weren’t any more expensive to make than any other LCD. This is supported by the fact that its first mainstream appearance was in the cheapest laptop ever built up to that point.

Mary Lou Jepsen, John Ryan, or whoever the owns the rights to this tech (assuming anyone owns it): please, PLEASE bring this back from the dead. If I could buy a new T-series ThinkPad with this display technology, I’d happily shell out an extra hundred bucks for it.

The Heart Of The XO

In 2008, Intel released their Classmate PC, another laptop designed for students in developing countries. Unlike the XO, the Classmate ran Windows. Naturally, the Classmate appealed to countries that wanted their students to learn how to use a practical PC operating system they could use in a future career.

In an apparent act of desperation, OLPC announced that they would make Windows XP an option for high-volume buyers in mid-2008 (Microsoft had flirted with OLPC as far back as 2006). This never came to fruition, though. I found no evidence that a single XO ever shipped with Windows.

If you ask me, it’s a blessing that XP never shipped on the XO. The XO is at its best when acting as an educational device with built-in computer features, not the other way around. I’m grateful that Windows never made it to the XO, because Windows would have broken the mirage.

The Classmate sold well, but Windows innately forced it into a subtly different market than the XO. Sure, it might be valuable to teach computer literacy to youth, but trying to teach computer literacy without internet access was almost as pointless in 2008 as it is today. Internet access was scarce in schools in many developing countries. This ITU report from 2009 shows that, in many countries in South America and Africa, fewer than half of schools had any internet access at all, let alone broadband. By the way, home internet access is still scarce in developing nations in 2024.

In a Classmate, an Eee PC, or a Chromebook, the educational tools are contained within the PC. On the XO, the “PC stuff” is contained within the educational tool. OLPC reimagined fundamentals of the PC user interface, up to and including the elimination of the very concept of files and folders.

Side Note: What is a Desktop Environment?

Note to nerds: You can skip to the next section if you already know what a desktop environment is. I make some generalizations in this section that might make nerds upset.

A “desktop environment” is an OS’s user interface. The Windows desktop environment is the thing with the Start menu, desktop icons, a little area in the bottom right corner with a clock and some status icons, and so on. MacOS’s desktop environment is the thing with the program tray at the bottom, the bar at the top, desktop icons, and whatnot.

People usually think of their desktop environment as an intrinsic part of the operating system. That’s understandable - what would Windows be without the Start menu and icons? Here’s the thing, the desktop environment is just a program running on top of Windows (or MacOS, or …). It’s just another app. The only thing that makes the desktop environment different from a normal program is that your OS launches it automatically when you log in.

Your smartphone has a desktop environment, too - when you go to your phone’s home screen, or you go to the screen where you can switch between open apps, you’re looking at your phone’s desktop environment.

Your smartphone’s environment is what we’ll call a “single-tasking” interface. In other words, when you open an app, it takes up the whole screen - you can’t have two things at the same time. On the other hand, Windows and MacOS have a “windowed” interface, where multiple apps share space on the screen simultaneously, and can be moved around freely.

In Windows and MacOS, the desktop environment is shipped with the OS, and you can’t get rid of it or replace it with something else. With Linux, there are several different desktop environments, and you can pick whichever one you like. They have funny names like GNOME and Cinnamon. Some of them are similar to MacOS, others are very similar to Windows, and others are nothing like Windows or MacOS.

Sugar In a Nutshell

OLPC developed a bespoke desktop environment for the XO. Its name is Sugar, and it’s my favorite part of the XO. Sugar completely reimagines the desktop computing experience, and the results are fascinating.

On its surface, Sugar seems more like a smartphone UI than Windows/MacOS. It’s a single-tasking user interface where apps take up the whole screen. Just like a smartphone, Sugar lets you have multiple apps open at the same time, but you can only have one app on the screen at a time.

Like a smartphone, Sugar has a dedicated app-switching menu where you switch between the apps you currently have open. Just like a smartphone, Sugar has a home screen where you can find all the apps installed on your device, and a little status bar where you can see the device’s battery level and network status. Here’s what all of that looks like - tap the button at the top-right of the keyboard, and a menu appears, surrounding the edges of the screen:

The app switching part is at the top, and you’ll notice the battery, audio, and Wi-Fi info in the bottom right.

The XO’s designers didn’t invent these UI concepts. After all, the first iPhone came out at the beginning of 2007, almost a year before the first XOs left the factory. Single-tasking UIs go back to the 90s with PDAs.

However, that “app switching” function is a real innovation. Early mobile OSes, like Symbian and early verisons of iOS, didn’t have a dedicated menu for this purpose. Instead, you had to return to your home screen and chose whatever app you wanted to use. Sugar was among the first single-tasking UIs to support a dedicated app switching feature.

This isn’t an earth-shattering advancement in UI design. But, it’s clear that Sugar’s designers were thinking outside the box.

Activity-Oriented UI

Unlike any other PC operating system at the time, Sugar doesn’t distinguish between “files” and “programs.” The general user story in Windows/MacOS is that you open programs and use them to create and modify files. Those desktop environments imply a duality between program and file, between code and data. That duality is fundamental to all widespread human-computer interfaces going back decades.

Sugar throws that out the window.

I was fibbing when I said Sugar had an “app switching” interface, and a home screen where you launch “apps.” Sugar doesn’t switch between apps. Sugar switches between “activities.”

In Sugar, an “activity” is both a file, and the program that’s operating on that file. It’s both “Untitled Document 1.doc,” and Microsoft Word. It’s both Firefox, and “https://www.google.com.” Sugar’s user interface gives no separation between files and programs.

In Android, if you are looking at a photo in your gallery app and you want to look at a different photo, you navigate through the gallery app to find the different photo. In Sugar, you close the photo viewer and you load a different activity that has the photo you want to see.

Sugar’s “Journal” maintains a chronological log of activities. There are no folders, just activities. You can give an activity a descriptive name, but that’s it. It’s just a list of stuff you’ve been doing recently.

Take Sugar’s word processor, for example. There is no “File” menu. You won’t find buttons for “Open,” “Save,” or “Save As.” Instead, if you’re working on a new document and you want to save it, you go the “Activity” menu, you click “Keep,” and that creates a new entry in your journal. Of course, Sugar actually auto-saves your activities, so you don’t need to manually save activities if you don’t want to.

If you closed the activity and you want to return to it later on, you go your journal, and click on the activity you kept. Your document opens in the word processor. Thus, in Sugar, the two faces of an activity are inseparable. You cannot open a file in the wrong program. You can delete activities from your journal, but you cannot overwrite one activity with another one.

Coming from a normal PC desktop environment, Sugar feels very counterintuitive at first. But, it only took me about 20 minutes to fully understand it. It’s really, really easy to understand. Because Sugar auto-saves activities, you can’t lose your progress if you forget to save keep them.

Maybe you think this sounds stupid. Maybe you think Sugar puts a ceiling on the user by collapsing files, folders, and programs into a single chronological log of activities. And you know what, you’re right! It absolutely would limit what you can do with a normal mid-2000s PC. But, again, the XO isn’t normal. You’ll run into the limits of the XO’s 1 GB hard drive long before you’ll run into the limits of Sugar’s user interface. In fact, that’s exactly how it played out in practice - Ames says that kids spent a lot of time deleting old activities not in an effort to keep the Journal nicely organized, but because they needed to make room on the XO to download the new episode of Naruto.

I must confess some bias. I am a sicko for unconventional PC user interfaces. Sam was my primary text editor for most of my time in college, and I run tiling window managers on all of my computers. Sugar is badass, high-key. It has virtually no learning curve despite being completely foreign. It throws convention to the wind in order to create an utterly original experience that lowers the PC’s barrier to entry practically to the floor.

Indeed, this low barrier to entry is intentional. Sugar was designed to be pictorial, so that you could figure it out even if you couldn’t read the words on the screen. Were any kids ever going to figure out how to use the XO, then teach themselves to read with it? I doubt it, but they’d probably have a better shot of doing that with an XO than a Classmate.

Trojan Horses

Negroponte’s Trojan Horse of constructionism is actually four horsemen: their names are Turtle, Etoys, Scratch, and Pippy.

Turtle, Etoys, and Scratch are all basically the same thing. You drag and drop code blocks together to make a guy move around on the screen. They’re all redundant with each other. Scratch, the only one you might have heard of, is the best by far.

Here’s what Turtle looks like. You get the idea:

Pippy is the fourth horseman of the Constructionist invasion. Although I mentioned it earlier when taking about bugs, I’m actually pretty happy with Pippy. It’s a barebones Python coding environment. You can write Python code, and run it with a simple click of a button. It comes with some nifty libraries for working with graphics and sound. I can imagine a budding programmer writing a simple game in Pippy. I could see this being a perfectly adequate tool for teaching middle and high-schoolers the basics of coding.

My XO also came with a terminal app, which, predictably, drops you into a Bash shell. I was shocked that they included this - it seems like a massive footgun for a target user who had probably never seen a Linux terminal before. Then again, Constructionist thinking demands a high level of technical openness, so absolute control of the system had to be put in kids’ hands.

Roots of Adolescent Hacking

In the spirit of openness, the XO offers other places where a nerd can get their fingers into its guts. If you hit Ctrl+Alt+F1 - or, rather, Ctrl+Alt+Neighborhood - boom, you drop into a root shell. There’s no jailbreaking to be done, no hacking to get absolute access to the machine. It’s all at your fingertips, and that’s completely intentional. It’s a core principle of constructionism. Alt+Right gets you back into the desktop, by the way.

So many electronic devices don’t let their owners actually own them in full. Imagine if your car’s engine was locked away in a safe, and only the manufacturer had the keys to the safe. That’s basically how your phone works. But, with the XO, you just pop the hood.

As a nerd and a writer, I love this. It took no effort at all to study the nuts and bolts of the XO. Even better, Ames observed kids actually taking advantage of the XO’s hackability. They installed alternative desktop environments on the XO. They also installed Wine, a Windows emulator which they used to play video games. As a kid who spent considerable effort circumventing the restrictions imposed by my school’s IT department, I vibe with these kids’ spirit of subversive computer use.

At the exact same time that kids in Paraguay were using clever tricks to install DOOM on their school computers, I was using proxies and batch files to install Halo CE on my school computers. That’s pretty cool, but it also wasn’t making any of us smarter. Shooting games weren’t turning us into sociopaths, but they also weren’t making any of us better at our three R’s.

Most of us didn’t learn much about computers by doing this, either. Most kids were really just going through the motions, following the steps someone else had scribbled down on some wide rule notebook paper. We weren’t discovering the inner machinations of the computer, we were installing video games.

Herein lies the problem with a laptop designed around constructionist principles: constructionism only works for nerds. Most of us never learned how those batch files worked because we didn’t care! Hacking was a means to an end, not a topic that captured our imaginations.

A Laptop Can Be Art

In Season 21, Episode 4 of Top Gear, Jeremy Clarkson poses a philosophical question: “Can cars be art?”

One argument states that art necessarily has no purpose except itself - ergo, a car cannot be art because it was made for another purpose: it’s a transportation device. Clarkson argues that the Alfa Romeo Disco Volante is art even within those parameters. The Alfa is an engine’s melodic howl echoing across the switchbacks of the Stelvio Pass. The Alfa is the sensation of lateral momentum as you are thrust forward through a meandering Alpine ravine. Its “being a car” is merely a medium to those sensory and emotional ends. The Alfa is a car in the same sense that the Mona Lisa is oil and pigments on a wooden plank.

The XO was conceived not as a computer, but as a manifestation of constructionist philosophy. The XO’s vision is passionate, romantic, and detached from the realities of everyday life. In other words, the XO is art.

Is the XO a masterpiece? I don’t think so. I’m not even sure if the XO is good art. The execution is hit-or-miss, and the underlying concept is troubled. The XO is trying to be an educational device for the masses. But, it’s also trying to manifest an educational pedagogy that’s inherently exclusive. OLPC’s idealistic pursuits contradict each other, compromising the XO’s ability to fulfill any single aspect of their vision.

Legacy of the XO

As much as I harp on Nicholas Negroponte and OLPC, I really don’t get the impression that anyone involved was acting in bad faith. They wanted to make a great thing that would help kids learn, they just failed in spectacular fashion. It’s Hanlon’s Razor incarnate.

OLPC began to see slightly better results later in the XO’s life (6). But, by the time those results were coming around, the writing was on the wall. Today, Chromebooks dominate the ed-tech scene, and OLPC no longer produces their own hardware. Instead, OLPC provides support to developing nations seeking to integrate technology into their education systems. Here’s a graphic from their website:

Ah, so OLPC doesn’t stand for “One Laptop Per Child” anymore. It stands for “Operation Learning Project Computer.”

… Right.

All my research suggests that its effect on education in the developing world was net-zero. Neutral. Nada. Neither good nor bad in any significant measure - okay, maybe it was a waste of scarce public funding, but I’m really trying to avoid politics here. The XO’s legacy isn’t plainly negative, though. In fact, almost 20 years after Negroponte’s first presentation, the XO’s aggregate effect on the world might even be slightly positive. Although the XO failed to improve the world on its own, it inadvertently acted as a canary in a coal mine.

In 2010, the Journal of Technology, Learning, and Assessment (JTLA) published its “1:1 Special Edition.” This collection of seven papers, all publicly available on ERIC, study the impact of so-called “1:1” educational technology solutions. A 1:1 ed-tech system is one wherein each student is assigned their own device for doing their classwork. In other words, it’s a system where there’s one laptop per child.

Sound familiar?

Several of the papers assume that the reader knows what “1:1” means in this context (8, for example), so clearly the term had been coined long before JTLA’s special edition. However, the papers also make it clear that 1:1 was a novel research topic, not a prevailing educational practice. Opinions on the feasibility and effects of 1:1 ed-tech vary between the papers, but I’d describe the overall tone as “cautiously optimistic.”

Although the XO was never sold in the quantities OLPC initially promised, the scale of its deployment blew the doors off all the studies in the JTLA special edition. As far as I can tell, the XO was the world’s first attempt at putting 1:1 educational technology into practice at a large scale.

OLPC, with virtually no preparation and no prior experience, set the XO on a course straight into uncharted waters. Predictably, the ship sank. But, we can learn a lot from the XO’s failure. For all its faults, the laptop itself was not the primary reason OLPC failed to improve educational outcomes in the developing world. Instead, it was OLPC’s hyper-idealistic worldview that fueled its failure.

Subsequent 1:1 ed-tech solutions learned from OLPC’s mistakes. As of 2024, 1:1 systems still produce some mixed results (4), but most signs point towards their ability to affect positive change on educational outcomes (1, 2, 3). Unlike OLPC, these modern programs provide technical support and training programs that help teachers incorporate the laptops into their curricula. Modern 1:1 ed-tech solutions also integrate a Learning Management System (LMS), a software tool that helps teachers and schools manage laptops and course materials. OLPC’s “School Server” software was kind of like an LMS, but it missed fundamental features, like the ability to push course materials to students’ devices remotely (5).

Despite the XO’s apparent place as the genesis of large-scale 1:1 ed-tech, I can’t find any sources that describe the XO in those terms. I’m not sure why, but I suspect it’s because OLPC was such a disaster that proponents of 1:1 ed-tech would rather not associate themselves with OLPC.

To be fair, this probably also isn’t a lens through with OLPC would like to be seen. From this standpoint, we see that OLPC basically used a bunch of poor kids in South America as cheaply-available guinea pigs in an experiment that would directly affect their livelihood by influencing their learning. This definitely isn’t the most flattering way of looking at their mission. Then again, Negroponte never shied away from condescending the education system.

Like most art, the XO isn’t a terribly practical tool, but we can still learn a lot from it. The XO solved none of the world’s problems, but it suggests that we can.

A few footnotes…

Software: The XO came with a bunch of other software that I didn’t talk about, including a couple basic games, some music programs, and the funniest text-to-speech engine I’ve ever encountered. They aren’t overly relevant to the XO’s legacy and this post is long enough as it is, so I decided to skip over those. TamTamJam is pretty cool, though.

High Quality Photos: Besides the image from PopSci and the graphic from OLPC’s website, all photos in this blog post were taken by me. I hereby release my photos into the public domain. I encourage you to use them as you please. Attribution and/or a link back to this blog would be appreciated, but it isn’t necessary. A ZIP file containing the highest quality JPEGs (better quality than you’ll find on this site) can be downloaded here via HTTP, or here via IPFS.

Corrections: Much has been written about the XO’s saga, but I still had to rely on primary sources (like the laptop itself) to assemble a complete picture of its history. Although I have spent many hours making sure this post is factually accurate, I’m only human. If you see any errors, please reach out to me at this email address.