That's akin to saying that it seems fundamentally impossible to make landing rockets safe which, in fact, is exactly what Boeing/Lockheed were saying when SpaceX was first revolutionizing that space as well.
I’m not aware of any rocket landing safe enough for human use. NASA nixed the idea of propulsive landing for Dragon 2 for this reason. It’s extremely difficult to make safe, since just about any reasonable engine configuration means guaranteed death if a single engine fails at a critical moment. Compare with modern airliners where an engine can fail at any point in flight and the plane can land safely.
You're right; I meant to refute the following point:
NASA nixed the idea of propulsive landing for Dragon 2 for this reason (safety)
It wasn't because of safety, but because it would have needed tests, development and certification (for a new type of landing) while already having an established method (splashing into the sea).
> NASA nixed the idea of propulsive landing for Dragon 2 for this reason.
That is completely false. First of all, NASA didn't nix it, they just didn't make it a priority as it had little value from their perspective.
The reason it was not done is that para-shouts have to be in the design anyway for abort situations, so that was fixed.
So for SpaceX, the question was to likely delay the program, and take on a whole lot of extra engineering work that they were not actually getting paid for, remember fixed price contract.
They were only going to work on it if they really thought they needed it for something like Red Dragon. And then they could still add it later.
And one of the primary reason SpaceX thought that its to hard, is that they landing feet would have to have gone threw the heat shield. That would have made the whole heat-shield design massively more complex.
He said it was the difficulty in proving the safety. There's an informative article here. [1]
NASA likes parachutes because they've always used parachutes. SpaceX likes retropulsive landings because Mars is their goal, and Mars' atmosphere isn't dense enough for parachutes. It's also safer for the crew in nominal operation and enables a much higher degree of rapid reuse, relative to NASA's traditional operation of taking a salt water bath in the ocean.
So they could go through the [very reasonable] extensive costs and testing involved in proving the safety of the retropulsive landings, or just go old school, strap a few parachutes on and work on getting crew to the ISS (which was the goal at the time). They chose the latter and with the plan of getting back to retropulsive landings later, which they also did. Parachutes remain the main landing mechanism for the Crew Dragon, but it now also has retropulsive landing capabilities to be used in case of a chute failure.