[th0ma5]

Since there's no way to truly objectively tell if LLM output is correct, this seems like it would have its limits, even if it seems subjectively good, but I have that problem with all of the LLM stuff I guess.