TL;DR: the author found a very, very specific bias that is prevalent in both humans and LLMs. That is it.
TL;DR: the author found a very, very specific bias that is prevalent in both humans and LLMs. That is it.