June 2, 2026
Ctrl+Alt+Delete This Prose
Can A.I. Produce Writing That We Want to Read?
Readers Are Dragging Robot Prose as Soulless, Sus, and Weirdly Too Perfect
TLDR: The article asks whether artificial intelligence can produce writing people genuinely enjoy, as schools and publishers struggle to tell machine text from human work. Commenters mostly say no: they’re roasting it as too smooth, too fake, and missing the messy human spark that makes writing worth reading.
The big question in Jay Caspian Kang’s piece is simple enough for anyone who’s ever side-eyed a suspiciously polished email: can artificial intelligence write anything people actually want to read? His answer is basically: maybe someday, but right now the robot still has a vibe problem. Teachers say they can often spot fake student writing, and readers recently went into full detective mode after a Granta story sparked accusations that it felt machine-made. That turned into a mini literary meltdown, with outrage, denial, and the deliciously awkward publisher response that basically shrugged, “maybe we’ll never know.”
But the real fireworks are in the community reaction, where people are sounding less curious than personally offended. One commenter says the real issue isn’t whether A.I. can write stuff we want to read, but whether it can write stuff we should want to read — a line that feels like a philosophy class and a subtweet at the same time. Another says the dead giveaway is the creepy, too-smooth consistency: human writing wanders, slips, and changes tone, while robot writing feels like it was ironed flat. And the harshest drag? One reader said they want to read A.I. words about as much as they want to stare at Midjourney “art,” calling both shallow and lacking intent. Ouch.
Still, not everyone is slamming the door. One more playful take suggests machine writing could work as bonus material for fictional worlds and side stories — basically, fanfic fuel, not soul food. The consensus, though, is loud: readers may forgive mistakes, but they still want a real mind on the other side of the page.
Key Points
- •The article examines whether AI can produce writing that readers would genuinely want to read and whether such writing is becoming difficult to distinguish from human prose.
- •It connects AI writing to higher education, arguing that if AI-generated student work becomes undetectable, traditional anti-cheating incentives may weaken.
- •The article says current AI writing still often contains recognizable markers such as formulaic structure, punctuation habits, and strained metaphors.
- •It cites a recent controversy in which *Granta* published a story by Jamir Nazir that many readers suspected was AI-generated, while Nazir denied the allegation.
- •The author reports creating a Claude-powered guessing game using Project Gutenberg texts and AI imitations of major authors, finding the early AI examples still fairly easy to spot.