← back

An OpenAI Engineer and His Friends Debate the Future

Brent Katz, New Yorker · Other · September 28, 2026

This one’s actually a short documentary; I’d highly recommend it! It’s very story-like and just under 30 minutes long.

I’m not going to summarize the documentary here. I have opinions about the presentation of the content—especially the sort of cliché “computer science people are weird and out of touch with the world” undertones such media tends to possess—but this isn’t what I want to spend time writing about.

There are several points in the documentary (it might actually just be one, but I remember at least one time) where Simon states his concerns about AI surpassing his own quality, skill, and originality. I interpreted this statement as in regards to his artistic ability to write good comedy.

I’ve thought a lot about AI’s impact on art. I’ve generally adopted the stance that much art should be defined by its humanity. When I look at paintings or read poems, I try to put myself in the author’s shoes.

A not-very-short-slightly-related-slightly-unrelated aside: I was recently looking for a painting of a woman knitting, preferably by a window facing the outdoors or outside near a tree. I go to Brown, and the notorious New England winter weather has felt soon-to-come, like an old friend just leaving his house to visit mine. I wanted the painting’s outdoor locale to capture this feeling.

The knitting was because I wanted to use the art as cover-art for a writeup pertaining to protein folding.

While searching, I came across this painting I thought was perfect. Or, almost perfect. The aspect ratio wasn’t horizontal enough for the writeup, so I (my friend Carolyn) had to crop it. The painting was by Carl Holsøe, and it piqued my curiosity. I started perusing his other works, noticing how he tended to create scenes with windows, typically with white mullions. There are women wearing older gowns and aprons in these frames. The women are in the left or right thirds of the frames, rarely the middle.

The point I’m trying to make is that the value the painting had was twofold. It was pleasing to look at—”nice on the eyes”—but it also gave me a glimpse into the life of another human. I genuinely view the latter as essential to my experience of appreciating art. The humanity is thus implicit.

I wonder, though, if this is less true for other art forms. Perhaps more so ones where people care about the creators less? I think TV shows are good examples of this, and this is probably because of how artists diversify in different spaces. A painter might focus on one medium, maybe even a few types of subjects (e.g. landscapes, people, etc.) for a large chunk of their portfolio. Singers innately stamp their songs with their voices. Their “trademarks” are easy to identify quickly.

In contrast, directors of movies and TV shows (I feel) can’t do this in the same way. There are exceptions. Like Quentin Tarantino and feet. But much of the identity of these mediums comes from other things: plot, genre, setting, actors, etc. which aren’t unique to the director. When people want to watch a Tarantino film, I would bet it’s more of an ethos/credibility thing: we know Tarantino directs good stuff, so this next movie will be good. And his style contributes to the movies being good, but we probably don’t watch him just for his style.

So, I’m mildly concerned that these sorts of artforms will seriously suffer as LLMs improve. Specifically, ones where the identity of the piece is heavily distributed.

edit 9/29: a friend told me i didn’t articulate the movie director point well. i basically mean i think more people might listen to a song or look at paintings because of the artist’s style than for a movie. for movies i think it’s quality indication more than style preference.