It's not that hard to predict the future of technology if you follow the relevant fields of science/engineering. There are usually a set of core problems that e

It's not that hard to predict the future of technology if you follow the relevant fields of science/engineering. There are usually a set of core problems that everyone in the field is aware of and has been working on for decades. What is difficult, is predicting the timing of mass market viability of those technologies. (So in this respect it's like how many people had identified a housing bubble in 2004, but anyone who'd bet heavily on its imminent collapse then would've lost out.) Microsoft in particular actually have a track record of entering a market prematurely with an unpolished entry (often because the hardware resources/architecture/maturity of the underlying academic field aren't yet there to support a polished entry), failing, shelving the idea, and then being caught flatfooted N years later when someone else enters with something more polished.