Sarah Guo: AI still cannot solve the structure and content problems in essay writing
Sarah Guo posted on the X platform on November 9 that the difficulty of writing an essay is that authors think they know clearly what they want to express, but after putting it down on paper, they find that the sequence is wrong, the length is too long, the structure is loose, and the core content is lost. She said that AI still cannot help solve these problems at present.
This statement is Guo’s personal observation and does not provide specific cases, test data, or comparative experiments. The post does not mention any specific AI model or tool by name. Guo is a venture investor, founded Conviction, and focuses on the fields of artificial intelligence and technological progress.
Original source:
Title: the problem with writing an essay Publisher: Sarah Guo (@saranormous) Date of publication: 2026-11-09 URL: https://x.com/saranormous/status/2119668321935323255 Nature of source: original post by the person concerned Verification status: the full original post and publication time were read in this instance; verification date 2026-11-08; publicly accessible Supported content: supports Guo’s statement about the problems of essay-writing order, length, structural coherence, and loss of core content, as well as the judgment that AI still cannot help at present
ABAB AI Insight
Sarah Guo has long invested in artificial intelligence startups and runs a technology-related podcast. She has previously discussed publicly her experience using AI to assist writing, including having models repeatedly criticize and revise drafts, but she still believes AI has limitations in preserving the author’s voice and overall coherence. This post focuses the problem on cognitive and structural reorganization in the writing process rather than on simple text generation.
In terms of product paths, current mainstream AI writing tools mainly optimize continuation, rewriting, summarization, and style adjustment. They still rely on human intervention for steps such as “the sequence is wrong” and “core content is lost,” which require the author to re-evaluate priorities and argumentative logic. Guo’s statement does not claim that AI can never solve this; it only points to the current state. This is consistent with her previous public experience of “using AI to repeatedly criticize drafts but still feeling frustrated.”
A comparable case is feedback on AI writing from other investors or authors. Some users report that AI can speed up first-draft generation, but in long-form writing that needs to maintain the density of personal viewpoints and a tight structure, substantial human rewriting is still needed. Unlike this case, Guo explicitly points to the writing process itself, where structural problems are only discovered after writing, rather than to the quality of model output.
This statement changes Guo’s public boundary judgment on AI’s ability to assist with essay-level writing. She limits “cannot help” to the current stage of structure and content reorganization. The observation metric is whether AI tools later release verifiable functional improvements for “long-form structure diagnosis and core content preservation,” and whether Guo herself updates this judgment in later posts.
ABAB News · Cognition Laws
Structural problems in the writing process appear after putting pen to paper, while current AI tools mainly act on generation rather than reorganization.
“Cannot help at present” is a stage-based judgment and does not rule out future model improvements in long-form logic diagnosis.