A standard piece of advice is that every developer should do a major project in every big language family. A Lisp, an OO language, a functional language, Prolog, VHDL or Verilog, a Forth, et cetera. They'll all make you think differently and teach you lessons you will use outside of the family.
My favorite side quest that I definitely recommend. Physical parallelism is so fun to reason about.
I had the advantage of sitting next to a proper professional to bounces questions off of, so I was able to get an SPDIF decoder to work in about a week! It was a very interesting lesson in the costs of abstraction, and why you will never really find abstraction layers in hardware.
This. It's not Rust in the author's case. It's being exposed to a language that's enough different from what they're used to that it affects their big picture.
A common example I give is a basic chat client/server, to learn how networking code feels, especially nonblocking IO. Especially if you're implementing the same protocol in the different languages since it means you won't necessarily be as biased by what protocol patterns are easy to implement in one language.
I don't know if it counts as "major" to the OP, but early in my career, I wrote a Sudoku solver in several languages as part of learning them. It was just complex enough of a problem that I got a glimpse into the patterns that succeed or fail for different languages, especially when it comes to things like optimization, which is sometimes hard to observe in business applications.
Depending on what kind of concepts you want exposure to.
If you just want generic pure computation: write an interpreter for a toy language. If it’s too large, write a macro expander that can handle C style or TeX style macros.
If you want events and concurrency: write a tmux clone that can work with multiple terminals.
If you want IO but specifically networking: write a toy HTTP server that only supports 1.0.
Rust enums conflate the tag type with the union type that uses it. These should be separate things. In the example here, you can see that Rust gives no way to represent the tag type `{Human, Cat}`. In Zig there is an enum which is a set of tags, and a union which is backed by that enum and can have fields. It is obviously better to have these as separate concepts rather than conflating them. You often want the kind without the data, and Rust gives you no way to represent it. I'm not sure why I am bothering to rant about this but it's one of my biggest annoyances with the language.
Sure you can, bringing data along for the ride is optional in Rust. Nothing is forcing you to use it like a tagged union, it’s an optional (but useful) feature.
If you want specific values (integer or otherwise) associated with specific variants (i.e. Human = 1, Cat = 4) you can do that too. Use the #[repr(T)] or impl the From / Into traits yourself.
Such enlightenment was a common read back in the glorious days of before 2023.
But reading this in 2026, when the entirety of the profession has changed (for better or worse) due to LLMs, seems strange.
Like someone writing how some new preservation method for making food last longer revolutionized how they manage their pantry, shortly after the invention of the fridge.
I like the post - it's just sad how many will not have these epifanies anymore, or they wont mean anything, and how many starting today would already just say "how quaint" and move on.
My favorite side quest that I definitely recommend. Physical parallelism is so fun to reason about.
I had the advantage of sitting next to a proper professional to bounces questions off of, so I was able to get an SPDIF decoder to work in about a week! It was a very interesting lesson in the costs of abstraction, and why you will never really find abstraction layers in hardware.
Elixir was somewhere between my 12th-14th language and nothing changed my perspective so much as understanding why the BEAM works the way it does.
It is a scale, and there are other languages that go beyond Rust, with dependent types, effects, capabilities, linear types, formal profs,...
A common example I give is a basic chat client/server, to learn how networking code feels, especially nonblocking IO. Especially if you're implementing the same protocol in the different languages since it means you won't necessarily be as biased by what protocol patterns are easy to implement in one language.
If you just want generic pure computation: write an interpreter for a toy language. If it’s too large, write a macro expander that can handle C style or TeX style macros.
If you want events and concurrency: write a tmux clone that can work with multiple terminals.
If you want IO but specifically networking: write a toy HTTP server that only supports 1.0.
If you want specific values (integer or otherwise) associated with specific variants (i.e. Human = 1, Cat = 4) you can do that too. Use the #[repr(T)] or impl the From / Into traits yourself.
But reading this in 2026, when the entirety of the profession has changed (for better or worse) due to LLMs, seems strange.
Like someone writing how some new preservation method for making food last longer revolutionized how they manage their pantry, shortly after the invention of the fridge.
I like the post - it's just sad how many will not have these epifanies anymore, or they wont mean anything, and how many starting today would already just say "how quaint" and move on.