Unlike other airframes, the U-2 seemed to have an inordinate number of cases of "this part isn't working today" (which means a collection fail or a mission scrub) and "we have to RTB, that's gone out again" (yet another mission scrub) during routine missions.
The statement above appears to be a casual observation rather than fact.
I flew the J-75 U-2 for 5 years and it was rare that a planned mission failed to produce results. Sometimes cloud cover intervened, or contrails were being made which is about the last thing you wish to happen. But the cameras, ELINT, deceptive jammers and sensor kit, which were all state of the art, usually worked as advertised.
The biggest problems in pre INS/GPS days were in not launching to the minute planned. This then threw off all pre-computed navigation sextant shots: delays of 10 mins or so were no problem, but the occasional long tedious wait for a go/nogo authorisation from above was not much fun sitting in a small cockpit in a pressure suit. All in all, Kelly Johnson produced a fantastic aircraft so good at high altitude long range missions that it is still in service more than 50 years after it first flew missions in earnest.
If only the current crop of UAVs could do as well, we could indeed replace the pilot who is both a real-time systems & aircraft manager with an electronic package!
It will happen, but it seems not as soon as we would wish.