Anthropic's Claude family writes differently from ChatGPT. Where ChatGPT tends toward short bullets and structured listicles, Claude tends toward longer flowing prose with carefully balanced parallel structure. The signature is more sophisticated, but still recognizable.
How Claude writes differently
Claude is trained to produce prose that reads as thoughtful and considered. The output usually has longer sentences, more clauses per sentence, and a strong instinct for parallel structure: "X is one thing. Y is another. Z, however, is something else entirely." The rhythm is more even than ChatGPT's bulleted output, but in a way that produces low burstiness rather than high.
Claude also tends to qualify and contextualize. Where ChatGPT will state things directly, Claude will often hedge: "It is worth noting that...", "While it is true that..., it is also the case that...", "In a sense, this is both X and Y."
Claude's signature phrases
| Claude phrase | Pattern | How to fix |
|---|---|---|
| "It is worth noting that..." | Soft opener | Just state the next point |
| "That said, it is also the case that..." | Both-and hedge | Pick one side |
| "In a meaningful sense..." | Reflective qualifier | Delete or replace with a specific claim |
| "This is both X and Y..." | Both-and again | Replace with a simple statement |
| "On one level... on another level..." | Two-axis qualifier | Drop one axis |
| "What is striking here is..." | Reflective opener | Open with the concrete observation |
A workflow for humanizing Claude
Related guides
Run Claude output through the humanizer
Cuts the qualifiers, breaks the parallel structure, keeps the substance.
Open the humanizer