Rellic[0] is a decompiler that uses a similar, more modern approach to turn LLVM IR into goto-free C code

[0]: https://github.com/lifting-bits/rellic

Maybe I'm just in the wrong field, but I wish papers were still written like this.

A lot of the foundational papers in NLP were written like this too. Clear and concise without grandiose overstatements. Dana Angluin comes to mind.

Things like this make me wonder what could be done differently by adding Z to the X and Y used here and reconstructing the intent in 3D.

Walter Pitts was working on something like that, but he burned his work.

https://en.wikipedia.org/wiki/Walter_Pitts