Inaccurate predictions about educational technology are discussed, and the tendency for them to go unchallenged. Their global generalizability is questionable when the expertise underlying them is restricted to local contexts or is unknown. The limitations of futuristic studies need to be stressed to prevent their conclusions from being applied out of context. Problems arise when studies inaccurately claiming to represent developing-world needs and conditions are used for educational planning in those nations. The article analyzes the predictions made by the annual Horizon Reports and the One Laptop Per Child project. It suggests that these projects are currently responding to criticism in a manner similar to that described in the classic study "When Prophecy Fails". (Contains 1 table and 2 notes.)