That is the question. Sam Dinkin discusses the pros and cons of pilots in space vehicles. He ignores the most critical issues, though--the willingness of passengers to fly, and the FAA to license, an unpiloted space transport. I've discussed this issue in the past.

Years ago before I started teaching I was an aerospace engineer. I was part of the Air Force'a NASP program and attended several of the Aero-Space Plane conferences. At one of the conferences there was a panel discussion. I do not remember the particulars of the discussion but I do remember an answer to a question posed to Scotty Crossfield. Scotty was asked about manned vs unmanned vehicles. His answer: "If there is no man in it, then the hell with it."

By the way in reference to my previous post, Scotty had been a childhood hero of mine and I was able to meet him and get his autograph at that conference

I think the dangers of remote control flight in aviation have been discussed before. If the pilot (human or otherwise) isn't in the craft, then they have less stake in the outcome of the flight. Second, remote control brings in several points of failure. The benefits of removing the human pilot may justify these costs, but i suspect that paying passengers will pay substantially more to have that human pilot.

Guilty as charged! Passengers would probably be unwilling to fly in a plane that does not have a pilot at first. But if we are starting an inherently risky activity that is a lot of fun--i.e. going to orbit, 33-50% off the price may be worth it. Especially if the unmanned rocket has already made a couple hundred delivery trips without a glitch. FAA would certainly take a long time to certify an unpiloted space plane. FAA, however, is allowing UAVs into civilian airspace so it is not impossible.

In a sense you are begging the question with your musing. Why not have teleoperated cars with graceful fail in the event of communication loss? Why not have FedEx replace their co-pilots with teleoperators? That it has not happened yet shows more inertia to me than that it is inherently bad business.

We are welcome to our taste in human pilots. I certainly like human customer service personnel better than VRUs, but I may be alone in my willingness to pay $20/call for the privilege.

As I recall there was similar discussion when BART started operating in SF. The BART trains are designed to run autonomously, and I believe they actually do. Nevertheless when I was there (85-90) there was a human bean up there in the cabin staring out the window. Not that he could have done a damn thing if the computer decided the second train was the square root of -20 seconds away and so directed the first through it at 65 MPH.

Frankly, I think it boils down to event speed. If there's time to think things out long and carefully, prefer humans every time. No computer could have gotten Apollo 13 home, it took human creativity. On the other hand, when the relevant events are occuring in milliseconds, humans are worthless. It's eerie reading the Challenger accident timeline while listening to the ground-to-orbiter communication. The computers are picking up all kinds of indications of disaster, and with long milliseconds to react to it are doing everything they can to avert it, while the humans are still blathering on utterly unaware that things are seriously screwed. Had, for example, there been a rocket-escape system in place on Challenger, and had it been under computer control, they'd be alive today. But if it took a human to pull the big red switch, even with the exact same system, they would not. Human CPUs just don't run at a high enough clock speed.