← Back to context

Comment by rezmason

2 months ago

Great post, OWL! I love a good mathematical diversion with cultural context wrapped around it, it's like a beef wellington for the analytical mind.

Codepen is great for discourse like this but I want to put the conventional Genji-mon in a unit test and then go hog wild with rule construction, because I remember running out of steam exploring more robust rules that still satisfy `width` = 5.

I'm bearish on LLMs, for better or worse, but you seem to know your way around them. Do you suppose this problem could be attacked with some sort of trained rigmarole that writes rulesets?