The true cost of the OLPC

A while back I wrote on the real cost of the OLPC laptop for OLPCNews:

At the end of five years of training, continued Internet, and maintenance, the actual cost is USD$972 per laptop, almost quintuple the Libyan estimates, and ten times the original laptop cost. Of course, a more expensive computer system would just drive all of this upwards, so at least we're starting cheap. This all reminds me of Namibia's SchoolNet rejecting Microsoft's "gift" of MS Office (sans operating system!). For the OLPC project to succeed, it needs to accept that it's selling a $100 laptop with an $872 support plan, and find countries that can afford it as such.

On the basis of that, Lisa Hoover interviewed me for an article at Newsforge, ending with what I see as the most important point in all of this:

Camfield says the most realistic way to determine those costs is through thorough field testing. While Negroponte has already supplied school children in a rural Cambodian village with their own laptops and often refers to that effort's success as a motivating factor for the OLPC project, he has declined to provide specific results of the Cambodian field tests.

"[OLPC's] annual projected budget will be $30 billion to push these laptops out, which is more than Intel's annual income and more than the World Bank's entire lending for all of 2005," says Camfield. "And naturally, all the countries will be taking out loans to cover this purchase. It's a huge risk to take without seeing some pilot project evidence. As Negroponte pointed out, the success of the OLPC project may be hard to quantify using testing methods. With a loan, however, there is a simple measure of success -- will the OLPC laptops improve the country's economy enough to begin loan payments in time? Wouldn't a pilot project be a safer way to test this?"

This article got slahsdotted and attacked pretty harshly. Let me state for the record that I don't hate children, and actually do want to see the OLPC laptop succeed -- I just don't think it will if we don't start planning for the implementation side. After such an amazing job of researching and building the CM-1, the "Step 3" problem ( "3 - And then a miracle occurs") approach in the implementation side, and the consistent refusal to explore pilot and test programs, deeply disturbs me.