The challenge in a traditional MSR design is that you have half the periodic table swimming around in your primary loop at high temperature. From a corrosion perspective that's pretty challenging.

Correct. But there's some fancy high-nickel and other stuff that can likely handle it if you control the chemistry carefully.

Salty water is a problem, but pure chloride salt is less corrosive to stainless steel than pure water, according to a presentation I saw by someone from another chloride reactor company.