--- schema_version: 1 type: "essay" id: "essay-co-writing-with-language-models" slug: "co-writing-with-language-models" title: "The draft is shared. The point of view is yours." summary: "A model-neutral method for co-writing with language models without flattening your voice or surrendering your claims." dek: "The model can carry the draft. The argument stays mine." status: "published" visibility: "public" indexing: "index" categories: - "creativity" - "ai" - "thinking" tags: - "co-writing" - "writing process" - "voice" - "language models" created: "2026-03-19" updated: "2026-07-25" published: "2026-07-23" canonical_path: "/writing/co-writing-with-language-models" canonical_url: "https://rahilchamola.com/writing/co-writing-with-language-models" preferred_citation: "Chamola, Rahil. “The draft is shared. The point of view is yours.” Curated Randomness, 2026." reading_minutes: 9 authors: - id: "rahil-chamola" name: "Rahil Chamola" role: "author" legacy_slugs: - "rules-for-co-writing-with-claude" - "co-writing-with-claude" series: id: "building-with-models" title: "Building with models" part: 3 related_slugs: - "10-things-ai-has-taught-me" - "the-model-is-not-the-product" citations: [] --- # The draft is shared. The point of view is yours. _The model can carry the draft. The argument stays mine._ The friction was never creation. It was output. I have never been short of ideas. I have ADHD, and the gap between having a thought and publishing it is where most of mine died. Not for lack of quality. For lack of throughput. So the choice in front of me was never AI or authentic. It was AI or silence. I chose output. Most of what is on this site would still be a voice note on my phone. Then removing the friction created a different problem. When you can produce at speed you lose the excuse that you will get to it eventually, and the bottleneck moves from execution to taste. Taste is the harder one. So everything below is friction I put back on purpose: the parts I refuse to hand over, and the checks that stop a fluent draft from quietly becoming someone else's argument. > **The working rule** · insight > > The draft can be shared. What I claim, what I refuse to claim, and the last sentence stay mine. **BEFORE THE DRAFT** ## Give the collaboration a spine ## 1. Write the thesis before opening the chat. Start with one sentence you genuinely believe. The model may challenge it, sharpen it, or generate alternatives. It should not manufacture your conviction after producing a page of plausible prose. **In practice.** “The model is not the product; it is becoming infrastructure” is a thesis. “Write an insightful essay about AI products” is an invitation to average the internet. The thesis does not have to be perfect. It has to be yours. If the work changes your mind, write the new thesis explicitly instead of letting it drift sentence by sentence. ## 2. Build a perspective packet, not a biography dump. Give the model the smallest relevant bundle of audience, evidence, examples, tensions, taste references, constraints, and privacy limits. **In practice.** For a product essay, include the intended reader, three observed failures, the strongest counterexample, two lines whose rhythm you like, and the topics that are out of scope. Do not paste years of private conversation “for context.” *A compact perspective packet* ```text Outcome: What should this piece let the reader understand or do? Audience: Who is it for, and what do they already know? Thesis: What do I currently believe? Evidence: Which sources, experiences, or artifacts may support it? Tension: What is the strongest objection or unresolved question? Taste: Which qualities should the writing preserve or avoid? Boundaries: What is private, uncertain, forbidden, or out of scope? Output: What shape should the next artifact take? Failure policy: What should happen when evidence is missing? ``` ## 3. Authored words are law until explicitly reopened. Use my exact words where they already carry the meaning. Thesis lines, personal experiences, verified claims, and phrases I deliberately wrote should not be silently replaced. The model may flag a conflict and offer alternatives. **In practice.** My instruction is simple: “where possible, copy my exact words whenever we are cowriting.” Mark those words and other authored lines as locked. If an edit needs to change one, show me why instead of smoothing it away. I am not protecting every comma. I am protecting the reason the piece exists from being smoothed away. **DURING THE DRAFT** ## Keep selection visible ## 4. Ask for choices before asking for prose. Structure is already an argument. Request distinct outlines, openings, metaphors, or counterarguments before letting one generated path harden into the piece. **In practice.** Ask for three four-part outlines: one practical, one argumentative, one narrative. Require each to name the claim it delays or foregrounds. Then choose rather than asking the model to choose for you. Selection is authorship. A model can create breadth cheaply; that makes the choice between options more important, not less. ## 5. Draft in stages. Move through thesis, structure, section drafts, evidence review, critique, and rewrite. Keep one bounded artifact and one active writing job at a time. **In practice.** Draft the opening and first section, then check whether the voice and evidence contract hold before generating the remaining 1,200 words. Early correction is cheaper than humanising an entire generic draft. 1. Human thesis 2. Scoped perspective packet 3. Model-generated options 4. Human selection 5. Bounded section draft 6. Evidence pass 7. Independent critique 8. Human final edit ## 6. Separate the writer from the critic. The critique should inspect a stable draft against explicit criteria. For consequential work, use fresh context or an independent model rather than letting the draft approve itself. **In practice.** Give the critic the essay, thesis, evidence list, and a rubric: find unsupported claims, generic transitions, missing counterarguments, and any place the voice becomes prescriptive. Do not ask whether the essay is “good.” Disagreement is useful evidence. Compare the claim, source, edit, and uncertainty. Do not settle a conflict because one model sounds more certain. ## 7. Facts are not style material. Never allow invented numbers, quotations, experiences, or citations to make the paragraph flow. Unsupported gaps stay visible until they are verified or removed. **In practice.** Use [SOURCE NEEDED] for a missing statistic and [EXAMPLE NEEDED] for a thin claim. A visible gap is less elegant than a fabricated bridge. It is also far more useful. I ask the model to separate what is observed, inferred, and unknown. That one distinction prevents a surprising amount of confident fiction. **BEFORE PUBLICATION** ## Take the work back ## 8. Restore human rhythm. Read aloud. Vary sentence length. Replace generic abstractions with concrete nouns and verbs. Remove anything that sounds competent but carries no lived observation. **In practice.** A model may write four balanced sentences explaining that context is important. I would rather keep one blunt line: “Capability gets copied. Context compounds.” Then make the next paragraph earn it. The goal is not to trick an AI detector. The goal is to sound like a particular person who noticed a particular thing. Mechanical “humanisation” tricks create another kind of generic prose. ## 9. Inspect the diff, not just the polish. Ask which claims changed, which authored lines moved, what evidence appeared, and what remains uncertain. Smooth prose can conceal a semantic rewrite. **In practice.** After a rewrite, request four lists: preserved claims, changed claims, removed evidence, and unresolved gaps. Review those before reading the polished version end to end. This matters most when the draft sounds better. Surface improvement makes it easier to miss that the argument became safer, broader, or less true. ## 10. Own the final sentence. Verify citations, privacy, implications, and disclosure needs. Shared labour does not transfer responsibility for publication. **In practice.** Before publishing, I should be able to explain why every section exists and defend every factual claim without saying “the model suggested it.” ## A complete starting packet *A first request for an essay about AI products* ```text Thesis: The model is not the product. It is becoming infrastructure. Audience: Product builders who confuse model capability with product experience. Evidence: Three observed product failures and one credible counterexample. Voice: First-person, discovery-first, punch / explain / punch / ground. No expert-at-a-podium posture. Locked lines: The model is not the product. Speed is the tool. Judgment is the craft. Next output: Three distinct four-part outlines. Include the strongest counterargument. Do not draft prose yet. Failure policy: Mark unsupported claims [SOURCE NEEDED]. Do not invent examples, numbers, or quotations. ``` ## Signs the model has taken over ### Stop and recover the point of view if… - [ ] The thesis appeared only after the prose. - [ ] Every sentence is equally smooth. - [ ] The evidence could support almost any argument. - [ ] No authored line survived unchanged. - [ ] The same pass drafted and approved the work. - [ ] You cannot explain why a section exists. - [ ] You would not defend a claim without mentioning the model. Writing has always involved editors, conversations, references, and tools. What matters is knowing where the judgment lives. Let the draft be shared. Keep the point of view yours. --- Creator: Rahil Chamola Canonical source: https://rahilchamola.com/writing/co-writing-with-language-models Preferred citation: Chamola, Rahil. “The draft is shared. The point of view is yours.” Curated Randomness, 2026.