Are AI labs "pelicanmaxxing" — optimising metrics while missing the point?

A blog post by Dylan Castillo asks whether AI laboratories are "pelicanmaxxing" — obsessively optimising measurable metrics at the expense of genuine usefulness or meaning. The piece sparked discussion on Hacker News, earning 46 points and 12 comments from the tech community.