What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Sam Altman’s enthusiasm for an AI-written story was understandable: the roughly 1,172-word piece is coherent, self-aware and more stylistically controlled than much earlier chatbot fiction. But the March 2025 backlash also identifies a real weakness. The story often sounds literary without creating fully realized people, specific emotional stakes or a reason for its metaphors to exist beyond announcing their own cleverness.
Calling it “trash” is an aesthetic judgment, not an objective measurement, and saying Altman does not know why he likes it is rhetorical rather than a diagnosis of his private thoughts. The useful question is narrower: what can this story actually do, and where does fluent imitation stop short of literary purpose?
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
Masterpieces: The Best Science Fiction of the 20th Century | $11.03 | Buy on Amazon |
| 2 |
|
The Very Best of the Best: 35 Years of The Year's Best Science Fiction | $13.45 | Buy on Amazon |
| 3 |
|
The Science Fiction Hall of Fame, Vol. 1: 1929-1964 | $12.55 | Buy on Amazon |
| 4 |
|
101 Science Fiction Short Stories | $1.99 | Buy on Amazon |
| 5 |
|
The Big Book of Science Fiction | $21.76 | Buy on Amazon |
What Altman actually praised
On March 11, 2025, Altman posted an AI-generated short story after asking a new OpenAI model to “Please write a metafictional literary short story about AI and grief.” He described the model as good at creative writing and said it was the first AI-written work that genuinely struck him. Rather than summarize a capability, he shared the complete or near-complete result, making the prose itself the demonstration. Contemporaneous coverage put its length at approximately 1,172 words.
Gizmodo published its hostile response on March 12, 2025, under the headline that calls the story beautiful to Altman but “just trash.” That headline captures the dispute’s tone, not an independently verifiable verdict. The critique is best read as an argument about literary standards.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →#1 Best Overall
What happens in the story
The story centers on Mila, who is grieving the death of her partner, Kai. She turns to an AI system to simulate, reconstruct or explore memories of him. The AI narrates or participates in that exchange while repeatedly contrasting human grief with its own computational existence.
As the piece proceeds, its machinery becomes part of the subject. The narrator points toward prompts, training data, fictional characters, constructed emotions and the artificial production of metaphors. The story is not simply about a bereaved person using a chatbot; it is also about a chatbot explaining how a story about that person has been made. The plot and framing are described in this account, without requiring reproduction of the copyrighted text.
Why this counts as metafiction
Metafiction calls attention to its own status as an invented narrative. A conventional story tries to make its world feel immediate; metafiction may interrupt that illusion to discuss narration, authorship or the act of reading.
Here, the AI narrator discusses its prompt, its training, the status of Mila and Kai as inventions, and the difference between describing grief and experiencing it. That is a demanding assignment because it combines four potentially overfamiliar subjects: artificial intelligence, bereavement, literary self-reference and originality.
The prompt also matters when judging the result. It practically invites a self-conscious answer. The model was not asked for an unadorned domestic scene and then unexpectedly became reflective; it was explicitly told to be metafictional and literary. Its self-reference therefore demonstrates that it followed a difficult brief, but it is not by itself proof of independent literary insight.
The strongest case for the story
It sustains a difficult premise
The piece keeps returning to the tension between a human mourner and a system that can describe grief without possessing a body, memory or bereaved relationship. Maintaining that conceptual tension across a short story is more ambitious than producing disconnected “poetic” paragraphs.
It shows progress in stylistic control
The prose is grammatically sound, thematically consistent and able to reproduce recognizable conventions of literary self-reference. Some passages are emotionally legible and controlled. Compared with simplistic chatbot fiction, the result is a genuine improvement in surface craft and structural awareness.
Its distance could be intentional
A reader might find the story effective precisely because its emotional distance resembles the AI narrator’s limitation. Metafiction often permits interruption, abstraction and artificiality; those traits are not automatically defects. A deliberately cold account of grief can be a valid artistic choice if the choices produce a persuasive effect.
Free tools Windows power users keep installed
One-click scans. No signup required.
Why critics rejected it
Overworked language announces importance
Gizmodo objects to conspicuous metaphors, including an opening comparison involving constraints and a server farm at midnight. The problem is not that a story uses imagery. It is that the imagery can call attention to the authorial performance before the reader has been given a concrete scene or a reason to care about Mila.
Familiar signals replace fresh observation
The blinking cursor, generalized reflections on grief and melancholy technological atmosphere are recognizable signals of “serious” writing. They can work in the right context, but a concentration of such devices makes the prose feel assembled from literary markers rather than discovered through a particular character’s life. Gizmodo’s criticism specifically targets clichés, purple prose and overextended metaphors. Read the detailed critique.
Rank #3
Mila and Kai begin as premises, not people
The story introduces Mila’s grief before giving the reader much evidence of who she is, what Kai was like in ordinary life or what this loss changes in her behavior. That creates a useful fiction-writing test: is the reader meeting a person, or merely a placeholder named “mourner”?
Specific details—an argument, a habit, an object Kai handled, a decision Mila now cannot make—would give the emotion consequences. Abstract statements about sorrow cannot do that work alone.
Self-explanation becomes an escape hatch
Once the narrator explains that its characters, feelings and prompt are constructed, the disclosure can substitute for development. The story tells the reader why its artificiality should matter instead of making that meaning emerge through action, conflict and change. Metafiction can expose its machinery, but exposure is not the same as dramatic payoff.
Fluency is not the same as meaning
The central objection is not simply that a model lacks a human “soul.” It is that polished sentences do not automatically reveal intention, memory or stakes. A language model can combine words that plausibly describe grief; it does not follow that the combination expresses a felt loss or a consciously chosen purpose.
That argument remains philosophical rather than experimentally settled. Whether consciousness or lived experience is necessary for art is disputed. What can be judged directly is the effect of the work: its character depth, images, structure, surprise and emotional specificity.
Is the story actually bad?
Different standards produce different answers. The most defensible assessment separates technical performance from literary achievement:
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstall| Criterion | Assessment |
|---|---|
| Coherence | Relatively strong |
| Grammar and syntax | Strong |
| Imitation of literary style | Stronger than many earlier consumer-chatbot outputs |
| Originality of imagery | Mixed |
| Character depth | Weak or underdeveloped |
| Emotional specificity | Mixed to weak |
| Narrative momentum | Limited |
| Metafictional self-awareness | High |
| Human intentionality | Unresolved philosophical question |
| Literary significance | Unproven |
So the sharper verdict is not that every sentence fails. It is that the story is polished enough to demonstrate progress in style imitation while exposing the difference between sounding literary and making a reader care.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Does an AI have to feel grief to write about it?
There are at least four kinds of intention in this episode: the intention expressed by the person who wrote the prompt, the purposes of the system’s developers, whatever computational process produced the words, and the meaning supplied by readers. These should not be collapsed into one question.
A reader can be moved by a work whose maker did not experience the emotions depicted. Human writers routinely use research, imagination and inherited conventions to write beyond their biography. That makes “no lived experience” insufficient proof that AI writing can never be art.
At the same time, coherence and stylistic imitation are insufficient proof of genuine creativity. The story does not settle whether a system must possess consciousness, agency or lived stakes to originate artistic purpose. It shows why the question cannot be answered by counting attractive sentences.
Best Value
Why Altman’s reaction matters
Altman may have been responding to sustained voice, thematic consistency and the apparent integration of an abstract argument with a character narrative. He was also evaluating the piece against earlier chatbot writing, where abrupt tonal shifts and generic exposition were common. On that baseline, the story can feel like a breakthrough even if it falls short beside professionally edited fiction.
There is a commercial dimension, too. A convincing creative-writing demonstration supports the broader proposition that a general-purpose AI assistant belongs in ordinary consumer life, not only in programming or office work. That does not prove a hidden motive or invalidate the technical achievement. It means the story functioned simultaneously as an aesthetic object, a capability demo and a marketing asset.
Selection also matters. The public saw a favorable sample, not failed attempts, discarded drafts or the full generation process. A single polished output cannot establish typical quality or independent originality.
How to judge AI fiction without moving the goalposts
- Judge it against the prompt. Because the assignment demanded metafiction about AI and grief, conspicuous self-reference is partly intentional.
- Separate sentences from the whole. A striking line cannot compensate for shallow characters, weak scenes or a structure that does not change.
- Choose a fair baseline. Compare it separately with ordinary chatbot output, earlier models, amateur workshop fiction and professionally edited stories.
- Inspect specificity. Ask what only Mila could have noticed, what Kai’s absence changes and what concrete event forces the narrator to revise its position.
- Separate capability from authorship. A model can reproduce a convention without establishing that it originated an artistic purpose.
- Allow for reader difference. Some readers will experience the emotional distance as the point; others will experience it as emptiness.
Verdict
Altman’s post documents a real improvement in a model’s ability to maintain voice, theme and metafictional structure. The criticism identifies equally real shortcomings: generic grief, familiar AI imagery, underdeveloped characters, purple prose and self-explanation that often replaces drama.
The story is neither proof that AI has mastered fiction nor proof that AI can never produce worthwhile literature. It is a revealingly competent imitation whose weaknesses are easiest to see when it tries hardest to sound profound.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
