The timing has to be right and the hardware has to be possible - if you're too early it won't work.
Flying cars is a bad comparison - they mostly don't exist in widespread use because of reasons not related to computing. Risk, fuel, control, etc. - even then rich people do have helicopters (though that's mostly different).
Pointing out failed predictions does not imply that all predictions are similarly wrong. In computing - the examples I showed (and there are others) are more relevant.