That is, the idea is that you build a high-speed airline with spaceports all over the world, refueled by huge solar arrays refining Ox and CH4 out of the air. Then it turns out your airliners, by virtue of the fact they are orbiters, can also launch satellites and go to Mars and other interplanetary destinations (where similar spaceport situations can be built).
The key to the whole thing is to make the second stage an orbiter big enough to do re-entry so the whole system is re-usable. Re-usable second stages are a giant pain, but if the secret is to get big, and that means you're kind of "accidentally" running an airliner, then perhaps SpaceX has found the right business model for commercial space flight in such high volume that things like going to Mars and asteroids and so forth are sort of on the cost order of chartering a private flight -- well within the reach of many organizations.
I'm not sure if the BFR ship would have any emergency survivability. It looks to me like the sort of thing where the flight goes well or everybody dies. In which case, yes, no need for cabin crew!
(And to be clear, I have nothing against "goes well or everybody dies" as long as the probability of "goes well" is sufficiently high.)
1) Problem during launch. There's been one successful use of a launch escape system in rocketry history, and such a system has never been developed for a shuttle type craft.
2) Problem after entering orbit. Cutting your orbital flight short isn't going to change the fact that you still have to survive re-entry stresses and have a working vertical landing system.
I'm not sure if you're aware, but this is a feature of the latest version of SpaceX's capsules. I'm sure it would exist for their larger vehicles.
edit: although I'm just not sure what you'd do during an abort with all that fuel that larger vehicle holds...
The fact that the BFR ship serves as both second stage and capsule combined, whereas the Dragon 2 is just a capsule with a separate second stage to put it into orbit, makes the problem a lot harder. I bet that the system for the BFR will consist of "don't explode."
Oh, that's interesting. What makes you say so?
The reason seems to be a combination of NASA not being interested in propulsive landing (and they're funding most of Dragon 2) and deciding to go all-out on BFR instead of refining Dragon 2 further.
I think if you take off and you immediately find that one of the rocket motors isn't gimballing correctly, or something else looks off with the engines, it might be prudent to abort before building up a lot of velocity. (not to mention aborting before something explodes) It's certainly easier.
I imagine the most problematic abort in something like a NY to London flight would be if multiple engines fail several hundred miles out. All kinds of velocity in the wrong direction if you want to fly back, not enough thrust to maintain the proper trajectory towards the target. I think there are bound to be some scenarios where you can do a limping, powered landing (the vehicle probably only needs one or two working engines) but you have to do it downrange.
Aborts also leave you with too much fuel on board, probably way too much fuel. You could burn it off with the engines, although you'd want to actually use it to reach your origin or destination if possible.
If you abort early, you can still reach the origin. If you abort late, you can still reach the destination on the remaining engines. So the pertinent question is: do these two regimes overlap, or is there a gap, and how big? I don't have the answers, and you may be right that there is one.