Assuming he was ever sincere about how he'd "like to die on Mars, just not on impact", the SpaceX development cycle is to throw lots of cheap hardware at the problem to find problems quickly. "Quickly" isn't possible when the launch windows only come every two years. (Short journey times and always-open launch windows for returns are among the reasons that, if I had his wealth and influence, I'd have always gone for the Moon over Mars).

Bluntly though, the lack of any visible progress with SpaceX-branded Sabatier machines already had me suspicious.