Well, it wasn't made to trick LLMs. It was made so that the user could write a blog post about a new benchmark. Benchmarks get a lot of clout (and even money) these days.

▲

dom96 a day ago | parent [-]

I wrote the blog post so that more people would see the language and benchmark. The goal was to create a language that would trick LLMs and I was genuinely curious how models would fare (and I continue to be curious so I will run it on new models).

Why do I want more people to see this? Because I want ideas for how to make the language even trickier for LLMs.

▲

serf 17 hours ago | parent [-]

>Why do I want more people to see this? Because I want ideas for how to make the language even trickier for LLMs.

unless coupled with tools and paradigms thatd make it a joy for a human to use, why?

seems like it's an effort to make a language that's unusable by anything or anyone.

i've witnessed abliterated local models start taking apart compilers with ghidra and friends to better understand the underlying language when denied net access or source base for the lang, I would imagine the confusion this thing would cause would prompt most abliterated and able models to pretty quickly start reverse engineering.

▲

dom96 15 hours ago | parent [-]

Yep, most models do indeed start reverse engineering the binary. But even slowing the models down is useful.

> unless coupled with tools and paradigms thatd make it a joy for a human to use, why?

The goal is to make a language that is easy for humans, while being hard for LLMs/coding agents. It's currently on the harder side for humans too. But I'd love ideas of how to better balance this.

	▲	stasomatic 9 hours ago \| parent [-]
		Aren't you making them even smarter once they crack your PL?