What else could be the point to static analysis, if not completely ruling out bugs?
> vs. type systems.
Type-checking is a static analysis.
> We're talking about the relative merits of rewriting every line of C code out there in Rust vs. using verification tools,
Where exactly is the merit in deliberately choosing a more complicated approach to ruling out bugs?
> in order to increase our confidence in the safety of the code.
The conceptually simpler the approach, the more confidence one can have in the result.
> The argument isn't even about the quality of the result but the cost/benefit of the approach.
It's precisely in systems software where the cost of bugs is higher than everywhere else. Anything (types or otherwise) that simplifies the reasoning by which bugs can be ruled out in systems software is a strict improvement on the state of the art.