They should take on a huge project, like say, converting the linux source code into Rust.
The sheer bulk of the code will effectively "force" them to make Rust ergonomic. They might even end up with annoying things, like different sized ints on different CPUs, or... (horror of horrors)... running Rust through a pre-processor as part of its compilation.
The Rust compiler is written in Rust, so rest assured the Rust core developers spend a lot of time dealing with the language itself. Additionally Mozilla is spending quite a bit of developer time on Servo, so we have quite a few people actively writing a lot of Rust code.
Well both of those things basically already exist; you can do conditional compilation based on architecture, etc. and macros can do some of the less hacky things a preprocesser can (as well as many things it can't).
The biggest Rust project (which already has many of the core contributors working on it) is Servo, and it's probably about as ambitious as you can get while still having a nonzero chance at completion. I don't think there's much additional value in taking on a completely unfinishable project to shape the Rust language so it's good for a kind of programming (i.e. Kernel programming) that is far removed from what 99% of potential users will use it for.
Also, Dennis Ritchie developed Unix while simultaneously developing C, so it's not hard to imagine that he added features to the language to simplify his Unix code. At one point he added 100,000 lines to Unix within a year, so he had reasons to make the language ergonomic.
"To encourage people to pay more attention to the official language rules, to detect legal but suspicious constructions, and to help find interface mismatches undetectable with simple mechanisms for separate compilation, Steve Johnson adapted his pcc compiler to produce lint [Johnson 79b], which scanned a set of files and remarked on dubious constructions."
So although C designers saw the dangers of C and provided static analysis from the early days, many C developers keep ignoring them.
I, for one, see C as both unsafe and unergonomic. The many features of C++ can be seen as various attempts to make some or another thing expressible in C -- parameterized datatypes, namespacing, encapsulated resource management -- that wasn't before.
The sheer bulk of the code will effectively "force" them to make Rust ergonomic. They might even end up with annoying things, like different sized ints on different CPUs, or... (horror of horrors)... running Rust through a pre-processor as part of its compilation.