This. Even if it's just an experimental project, at some point down the line if there's a possibility that it's going to have to integrate with the rest of the system - you've just introduced significant expectations for the existing and new developers on the team. While I don't think that rust has a huge learning curve, it's different enough that you can't just easily transition from a different high-level language.

I would never introduce a new language to an existing code base without an exceptionally good reason.

That's a catch 22 though. If you never introduce it no one will learn it?

The people that mattered were on board with the experiment though. It was a healthy enough environment that the project failing if it did wasn't going to be terrible for my career

There are an infinite number of things I could learn. There still needs to be a business rationale as to why a new technology stack should be introduced which outweigh the costs. Speed, memory safety, rate of errors, ability to hire talent, ease of deployment, whatever. Using it because one team member thinks the language is cool does not meet the bar.