Bio
Sophie Alpert is an engineer at Clay and former React core team lead at Meta. She writes at sophiebits.com on software engineering, AI workflows, and the practical limits of language model tools. Her post on lossless text transformations was adopted as company writing policy at Clay.
Operating themes
- Operating thesis: language models are useful but structurally unable to preserve author intent through text transformation; teams that treat AI edits as equivalent to human edits are accepting meaning loss they have not accounted for.
- AI writing limits
- Author intent and meaning
- AI workflow design
- Software engineering
Cards
- Every LLM rewrite changes meaning because the model lacks the author's mental model of what the words are meant to communicate, Every LLM rewrite changes meaning because the model lacks the author's mental model of what the words are meant to communicate [Tier B]
Sources captured
- 2026-06-25, There are no lossless transformations of natural-language text, sophiebits.com (https://sophiebits.com/2026/06/25/there-are-no-lossless-transformations-of-natural-language-text)