> a keynote is an unusual place to talk about experimental feature a language is considering (unless it's a small component of a bigger update/forecast).
Quite the opposite, I believe. Since a keynote is seen by a dramatically larger slice of the audience, it's a good slot for talks meant to provoke thoughts and discussion, even if that discussion isn't necessarily one you agree with.
Rust is a great language with a good macro system that gets used more than it needs to than if it had a stronger type system.
To succeed at improving the language, contributors must strike a balance between approaching the PL research frontier while remaining a practical language.
ThePHD’s talk is a take on that trajectory that appears Pareto-optimal. It’s a really good subject for a keynote, I think.
Quite the opposite, I believe. Since a keynote is seen by a dramatically larger slice of the audience, it's a good slot for talks meant to provoke thoughts and discussion, even if that discussion isn't necessarily one you agree with.