sorbet-rails VS sorbet

There are solutions like sorbet[^1] and sorbet-rails[^2]. Have you tried them?

[^1]: https://sorbet.org/

[^2]: https://github.com/chanzuckerberg/sorbet-rails

> I guess what I'm asking is: do you see a future where there is more explicit control afforded to people who want to pick their own tradeoffs without resorting to writing everything performance-sensitive in extensions written in C/Rust/whatever?

An approach exists already in the present, and it's Stripe's Sorbet AOT compiler (https://github.com/sorbet/sorbet/tree/master/compiler).

That's incorrect. You may be thinking of Stripe, and AFAICT it's not very actively developed anymore: https://github.com/sorbet/sorbet/commits/master/compiler

Sorbet can do this, as long as you have type signatures for your code. Given Ruby's highly dynamic nature that's where tools like Tapioca come in to generate these, for example for Active Record models where instance methods are generated based on the database schema. But the moment when something returns T.untyped you're back where you were before - it helps but isn't perfect.

Sorbet[1] is currently the most robust tooling available. It has a few rough edges and limitations but it otherwise works great for 90% of use cases.

[1]: https://sorbet.org

There are solutions like sorbet[^1] and sorbet-rails[^2]. Have you tried them?

[^1]: https://sorbet.org/

[^2]: https://github.com/chanzuckerberg/sorbet-rails

Since I tried Sorbet, a Ruby type-checker, the idea of building an ahead-of-time compiler for Ruby is a cool side project that has been in my mind for some time now. To my surprise, during Hacktoberfest, I got to know Natalie, this AOT compiler written in C++ by Tim Morgan.

Rails has something similar called https://sorbet.org/ Haven't used it since I personally like the flexibility. But it exists if you're into it :)

Ever heard of Sorbet? https://sorbet.org/