Planes are essentially self-driving now, at least runway-to-runway.

No, they are not. That is like saying a company is self-running just because it has an automated production line. Much of the flight is under auto-pilot, but the human pilots are frequently changing the auto-pilot's instructions. There is a lot of training and skill maintaining in being a pilot. They aren't just there to keepen das hander in das pockets und watschen der blinkenlichten.

All take-offs are manual. Nearly all landings are manual. Mostly 'auto land' just takes the plane to just short of the runway, at which point a pilot takes over for the actual touchdown. Full auto land is possible, but with good visibility it is simply less work to manually land than to set up the auto land.

It doesn't work like that, it isn't a chemical fuel you can burn or save. The amount of Pu-238 you need is dictated by your peak power demand. How long it lasts is dictated by nuclear physics (the half life of Pu-238.) You have no control over how fast the plutonium is used up.

You're trying to study a temperature-sensitive environment in its natural state. An RTG produces lots of heat. (They are only about 5% efficient, so they produce twenty times as much heat as electrical power.) The presence of the RTG might perturb or destroy the environment you're there to study. I don't have the detailed knowledge to say if this is the case.

Plus the issues others have raised: mass, scarcity of suitable isotopes, and launching highly radioactive material on top of hundreds of tonnes of potentially explosive fuel is something you'd rather avoid if possible.

You can insure against loss of payload on launch.

I would expect that Orbital Sciences is liable for damage to the payload and had insured themselves, but I have no direct knowledge. Even if they are covered for direct costs, it is still bad for them.

OK, lets rephrase a little, and concentrate on the Challenger failure mode, rather than the actual shuttle.

Imagine a rocket that was compatible with an LES, and also compatible with the Challenger failure mode. (Remove the shuttle, put the liquid fuel engines on the bottom of the external tank, throw a capsule on the top, keep the solid rockets.) Now have the boosters fail in the same way they did with Challenger. Would the LES have sufficient notice to get the capsule to safety?

Systemd does not need to die. All the more power to those who wish to use it.

However, it is undesired by a significantly large portion of users and sysadmins, and it is unsuitable for those who still actually want to run Linux as a Unix-like OS.

For these reasons, in my opinion, it is not (yet) ready to become the init for a number of general-purpose distributions out there. Moreover, it is unacceptable for the udev subsystem to reside in the same source tree as systemd, and it is unacceptable for udev to integrate, except through the use of a stable and init-independent interface, into any particular init implementation or design.