I agree that humans have a degree of flexibility that computers will not have for a very long time. I'd say, however, that this flexibility is not of extra use more often than it is. What was the last time we took scientific data that absolutely required a human to be along for the ride? Tack on to this (1) that spacecraft can be remote controlled safely from Earth, (2) autonomous control systems are getting smarter and more powerful each year, and you have even less practical incentive to put a human on board. Obviously if the distance from earth is significant, the degree of control is lessened because of the transmission time, but that's another discussion.
I am a staunch supporter of a continued human presence in space exploration, but I can't deny that the current state of technology and funding make it utterly impractical. The benefits to having humans along are almost always outweighed by the drawbacks, especially in an era that struggles with heavy lift technology. Perhaps the solution is to simply wait until propulsion, life support, and launch technologies mature.