cross-posted from: https://ibbit.at/post/178862
spoiler
Just as the community adopted the term "hallucination" to describe additive errors, we must now codify its far more insidious counterpart: semantic ablation.
Semantic ablation is the algorithmic erosion of high-entropy information. Technically, it is not a "bug" but a structural byproduct of greedy decoding and RLHF (reinforcement learning from human feedback).
During "refinement," the model gravitates toward the center of the Gaussian distribution, discarding "tail" data – the rare, precise, and complex tokens – to maximize statistical probability. Developers have exacerbated this through aggressive "safety" and "helpfulness" tuning, which deliberately penalizes unconventional linguistic friction. It is a silent, unauthorized amputation of intent, where the pursuit of low-perplexity output results in the total destruction of unique signal.
When an author uses AI for "polishing" a draft, they are not seeing improvement; they are witnessing semantic ablation. The AI identifies high-entropy clusters – the precise points where unique insights and "blood" reside – and systematically replaces them with the most probable, generic token sequences. What began as a jagged, precise Romanesque structure of stone is eroded into a polished, Baroque plastic shell: it looks "clean" to the casual eye, but its structural integrity – its "ciccia" – has been ablated to favor a hollow, frictionless aesthetic.
We can measure semantic ablation through entropy decay. By running a text through successive AI "refinement" loops, the vocabulary diversity (type-token ratio) collapses. The process performs a systematic lobotomy across three distinct stages:
Stage 1: Metaphoric cleansing. The AI identifies unconventional metaphors or visceral imagery as "noise" because they deviate from the training set's mean. It replaces them with dead, safe clichés, stripping the text of its emotional and sensory "friction."
Stage 2: Lexical flattening. Domain-specific jargon and high-precision technical terms are sacrificed for "accessibility." The model performs a statistical substitution, replacing a 1-of-10,000 token with a 1-of-100 synonym, effectively diluting the semantic density and specific gravity of the argument.
Stage 3: Structural collapse. The logical flow – originally built on complex, non-linear reasoning – is forced into a predictable, low-perplexity template. Subtext and nuance are ablated to ensure the output satisfies a "standardized" readability score, leaving behind a syntactically perfect but intellectually void shell.
The result is a "JPEG of thought" – visually coherent but stripped of its original data density through semantic ablation.
If "hallucination" describes the AI seeing what isn't there, semantic ablation describes the AI destroying what is. We are witnessing a civilizational "race to the middle," where the complexity of human thought is sacrificed on the altar of algorithmic smoothness. By accepting these ablated outputs, we are not just simplifying communication; we are building a world on a hollowed-out syntax that has suffered semantic ablation. If we don't start naming the rot, we will soon forget what substance even looks like.
I suspect in large part its because using generative tools hits the brain differently and delivers a faster loop for drip feeding of dopamine, compared to creative work which often involves a long delay in ultimate gratification. Our brains optimise for dopamine reward which has been useful for most of our evolution, but we have become very good at hijacking that neurological feature with addictive activities.
I think generative tools might be uniquely sinister because the surrogate activity of prompting and generating still ends with some output that is superficially similar to what you might have aimed towards in starting creative work.
So unlike gambling or binging drugs, using generative tools leaves you with these generated artifacts that feel like creative output. I imagine that if this sufficiently satisfies the other non-dopaminergic rewards intrinsic to creative activity, it is less likely that whatever internal drive compels someone to create (their creativity / spark / soul / whatever the fuck) would object and create the necessary cognitive dissonance to stop using generative tools and return to manual creative work.
In other words they are probably addicted to AI and don't feel any loss from stopping their creative output. Sadly their creative abilities will be atrophying rapidly at the same time and I doubt they'll find much joy in creativity in the future.
They're getting skinner-boxed. AI doesn't always generate what they want, but its success rate is high enough for people who love AI that they want to gamble for the chance of AI generating something they actually want. Literally the same psychology as opening lootboxes and booster packs.