A minimax chess engine in regular expressions

This is from the same gentleman who (among other things) demonstrated that printf() is Turing complete and wrote a first person shooter in 13kB of Javascript.

https://github.com/HexHive/printbf

https://github.com/carlini/js13k2019-yet-another-doom-clone

Previously: Chess written in sed https://news.ycombinator.com/item?id=6261314

Of course, the sed version does make use of control flow commands in sed and only probes 1ply (I think) so this version is significantly different in that regard.

Regex isn't (necessarily) turing complete :)

> Because our program just consists of a sequence of regular expressions, you can't loop at all! That, technically, means we can't actually perform Turing Complete But we can do any bounded computation by just unrolling any loops we may have.

Although some (most?) implementations may be. Though by the above quote, the author didn't make use of that.

I wonder if this is inspired by LLMs which similarly do tons of repetitive processing most of which is unrelated to the final answer.

When I see this sort of thing I just I want to take my hat off and stand in solemn appreciation for the true heroes among men.