Compile-time reflection is great in principle. It gets us quite close to the aim of compile-time dependent types, which would be a great feature for Rust and bring it close to parity with newer languages such as Zig, as well as lay the groundwork for further important developments such as proof-carrying code. It would be quite important to know why this Rust team member felt "uncomfortable" with the technical contents of this talk which were clearly described by the author as highly speculative, and thus inherently open to discussion.