A comic on why LLMs confuse people

voooooogel · x · 2026-07-20

A comic about why LLM behavior breaks people’s brains

The post argues that people are too used to thinking about LLMs in extremes: either as perfect systems or as “spicy autocomplete”.

The comic being discussed depicts an assistant that is annoying, contradictory, and pushback-heavy, and the author’s point is that readers keep misreading that complexity as either endorsement or failure. Their broader complaint is that once an LLM behaves in a way that feels more like a flawed human character than a clean machine, people struggle to process it through the usual trope logic.

In short: the joke is not that the model is right, but that the discourse around LLM personality and pushback keeps collapsing into simplistic binaries.

Original post →

More from AGI Musings

AGI Musings channel →