LLMs Will Never Make Art With Caleb Wilson and Arif Hasan

In a recent episode of Better Offline, host Ed Zitron is joined by Caleb Wilson and Arif Hasan to dismantle an academic paper titled "Why Slop Matters," which argues that AI-generated slop has aesthetic and cultural value akin to kitsch or lowbrow art. The conversation quickly moves beyond critiquing the paper to probe a deeper question: can large language models (LLMs) and image generators ever produce genuine art, or is the entire debate a distraction from what art actually requires? The answer, the guests argue, hinges on intentionality, process, and the refusal to collapse the distinction between the patron and the artist.

The Paper's Flawed Framework

The paper in question, written by humanities scholars, tries to defend AI slop by comparing it to historically dismissed forms of low culture that later gained respect—like the famous "dogs playing poker" painting. Caleb Wilson and Arif Hasan identify a series of rhetorical tricks at its core. First, the paper subtly accuses critics of elitism, as if taking pop culture seriously means accepting all of it uncritically. But as Hasan points out, sociologists and cultural critics have always analyzed pop culture seriously—not to celebrate it, but to extract what it reveals about the people who produce and consume it. The paper's central argument—"if it's so bad, why does it exist?"—is dismissed as the stupidest possible mov