Bryan Cantrill, a prominent technology figure, has published a detailed critique of the widespread practice of using large language models to generate LinkedIn content, arguing that AI-authored posts are readily identifiable and counterproductive.
According to Cantrill, LLM-generated writing exhibits consistent stylistic markers that are “impossible to ignore” to anyone familiar with AI output. These tells include excessive emojis, single-sentence paragraphs, repetitive constructions like “it’s not just… but also,” and overuse of em-dashes. The result, Cantrill argues, is that readers quickly recognize the content as machine-generated rather than authored by the person posting it.
The problem extends beyond mere stylistic annoyance. When readers encounter LLM-generated posts, Cantrill contends, they lose confidence in the authenticity of the content itself. “When you — person whose perspective I want to hear! — are obviously using an LLM to write posts for you, I don’t know what’s real and what is in fact generated fanfic,” Cantrill writes. This uncertainty causes readers to disengage, he says, with many stopping reading entirely.
Cantrill acknowledges that LinkedIn itself contributes to the problem by actively encouraging users to “rewrite it with AI,” making the temptation to use AI assistance difficult to resist. However, he argues that convenience is not sufficient justification for outsourcing authorship.
While Cantrill emphasizes that LLMs have legitimate applications, he draws a sharp distinction between helpful and harmful uses. According to his view, LLMs excel at brainstorming, text comprehension, and editing—roles in which they serve as tools under human direction. But using them as primary authors is problematic because it replaces authentic voice with generic output.
Cantrill’s core recommendation is direct: writers should author their own content and trust their own voice. He argues this serves both personal credibility and the quality of the communication itself. The post was originally published as a LinkedIn post on November 11, 2025, and has generated significant discussion around authenticity and AI use in professional social media.
Key facts
- LLM-generated posts contain consistent stylistic tells including excessive emojis, single-sentence paragraphs, and repetitive constructions that are recognizable to readers familiar with AI output
- When readers recognize content as AI-generated, they lose confidence in its authenticity and often stop reading entirely
- LinkedIn actively encourages users to rewrite content with AI assistance, making the practice more tempting
- LLMs are useful for brainstorming, text comprehension, and editing but are described as poor primary writers
- The critique argues that using AI to author posts undermines the writer’s authentic voice and credibility