Just tossing laptops at kids won't make them learn more or help them get better grades, but that doesn't mean that they don't also have extraordinary educational potential. Allowing, of course, that you accept the tenets of Constructionist learning theory (which sounds valid to me, but having no formal education credentials or experience, I'm hesitant to sign on as a believer just yet). The important things are to A) design the computer and its software as dedicated learning devices (i.e. no solitaire, no HALO, or other time-wasters, unless they are games that are designed for learning purposes) and B) design an entire learning system and infrastructure that maximizes the OLPC program's strengths while mitigating its weaknesses. Perhaps a relatively decentralized network of students and teachers, whereas a child learns concepts and skills via interactive programs (which would perhaps utilize visual data and animation along with text, using the apparent tendency for humans to arrange information in a visual-spacial manner as leverage for greater learning) combined with regularly scheduled communication with a teacher who would assist the student in placing those skills and concepts in context, as far as how it fits in to the world and life as a whole. The computer could provide knowledge and skill sets, while the teacher could provide guidance and experience. That's just my idea of course. The point is that computers are tools, and while that have great educational potential, they must be designed and utilized as part of an effective system, otherwise they're just toys.