The basic issue with this criticism of Cargo is that the fact that updating dependencies can cause breakage never actually goes away. The package management tool either makes this problem easy, or it makes it hard. The approach Cargo chooses is to make "cargo update" more likely to work by allowing library authors to specify incompatibilities. This can go wrong in edge cases. But that's inherent to the problem itself. Minimum version selection only seems to work because it doesn't actually solve the problem of keeping dependencies up to date. The moment you want to do that, you're stuck with "go get -u", which is just a worse version of "SAT solving" because it doesn't know how to detect and avoid incompatibilities.

(As an aside, I don't like the term "SAT solving" for the problem of package version selection. By focusing on an implementation detail, it makes the problem seem scarier than it actually is. Register allocation is NP hard too, but nobody calls the register allocator "the SAT solver".)