I always thought the most unlikely technological development in my lifetime was the handheld GPS device. It would be "most unlikely" because it required tremendous, simultaneous, and largely unforeseen advances in several different technologies, each of which was hard to predict in 1981. The list is at least:
1. Low power, low voltage, low noise L-band receivers, sensitive enough to be compatible with the weak signal coming from the internal antenna of a handheld device;
2. Stupendous amounts of digital signal processing, also at low power and low voltage;
3. Digital map databases of (substantially) every road in the world, accurate to a few meters;
4. A substantially world-wide, wideband wireless data link to get the digital map into the handheld device in the first place;
5. Low power, low voltage, high resolution, multicolor flat panel displays;
6. Gigabytes of low power, low voltage data storage memory; and
7. High energy density, high power density batteries capable of supplying the whole thing.
And, perhaps most impressive of all, the manufacturing technology to make all of the above small enough to fit in a handheld device, at a price low enough to sell by the zillions.
Of the list above, probably only #2 could have been predicted, and then only if one were willing to extrapolate the then-relatively-new Moore's Law by a very large amount. (Recall that Mead and Conway had only written their Introduction to VLSI systems the previous year; until then it was not clear that such complex chips could even be designed on human time scales, let alone built for a profit.)
The fact that a handheld GPS device is now an anachronism, since the technology is now small enough and low-power enough to be integrated into other handheld devices, like smart phones, pleases me no end.