Grunewald: Don't use AI to write substantive work
Writer argues the thinking happens during writing itself, making AI composition counterproductive for serious intellectual work.
What to know
- Grunewald argues the writing process itself is where thinking happens, making AI generation of prose counterproductive for developing and refining ideas.
- He identifies three problems: writing is inseparable from thinking, AI output contains subtle errors and vagueness, and unlabeled AI writing misleads readers.
- The essay carves out legitimate AI uses—transcription, data analysis, brainstorming, and editing—while drawing a line at full composition.
- Grunewald qualifies his stance as specific to current and near-term AI models, leaving open the possibility future systems might be capable enough to justify delegation.
“Writing about something, even something you know well, usually shows you that you didn't know it as well as you thought. Putting ideas into words is a severe test.”
Erich Grunewald · Substack essay
Erich Grunewald Writer and essayist
How it unfolded 1 development · click the chart to see its coverage articlesposts
-
1
Grunewald argues writing itself is a thinking tool
The essay explains that the act of converting outlines into prose forces writers to test and refine ideas, catching logical gaps and generating new insights that would not emerge from AI generation.
“Often I'll have an experience where I'm trying to turn an outline into a finished product, and I'm playing with a transition, and it's not working, and I realize, oh, the reason this transition isn't working is because actually these two points should not be juxtaposed. The thing that I'm trying to do here is wrong. And if I feed the outline into an LLM, it is not going to stop and consider maybe the outline is bad.”
— Erich Grunewald -
I don't know what everyone else is doing but I've certainly been writing a lot more often because of my interactions with AI. Maybe some people used to write more but have been leaning on AI so they can write less but I don't think that's universally the case.I agree that being lazy about writing and simply using a short prompt or list is…
2 more of the top 3 · 53 posts in this stretch
-
R
Very glad to see this article from fellow # JavaChampion Paul Bakker: https:// paulbakker.io/writing/no-ai-fo r-writing/ > Writing forces you to deeply think about the problem you’re solving, and what you’re trying to communicate to your readers. This is why writing is hard, but also why writing is such a useful tool to help you think. If you…
-
This advice is too coarse. I believe LLMs have made me a better writer of technical reports. I have a keen ear for language, and I can usually tell when a sentence or a paragraph is orbiting around a point rather than landing on it firmly. Some writers insist that unclear prose is dispositive proof of unclear ideas -- a sign the writer has not…
-
-
background
Grunewald defines acceptable AI uses in research and writing — The author clarifies that he supports AI use for transcription, data analysis, information search, brainstorming, and line editing—as long as humans deliberately accept or reject each edit.
-
background
Grunewald cites three core objections to AI writing — The essay identifies three main reasons to avoid AI for substantive writing: the writing process is essential to thinking, AI output is subtly inaccurate, and unlabeled AI writing deceives readers.
-
background
Grunewald publishes essay against AI writing — Erich Grunewald posted an essay on Substack arguing that people should almost never use AI to write substantive work such as blog posts, research reports, memos, or novels.
Also covered reported alongside — the timeline has no entry for these yet
-
first by HN Best, 8d ago · also HN Frontpage
2 more headlines
- Almost Never Use AI to Write Anything Substantive HN Frontpage · 8d ago
- Don't Use AI to Write HN Frontpage · 6d ago
What people are saying 21 voices from 2 sites · best of 54 · verbatim
- Sep 22
-
My take is slightly different.- Brainstorming with AI is fine, but it takes skill to use it as food for thought, rather than as a source of answers.- Do NOT structure your blog post with AI. Once the structure is there, it is almost impossible to humanize it.- Do NOT use it for writing a draft. Same as above. A few times I did it, and ended up…
- Sep 21
-
I'm the author of the article.There are a number of comments that indicate that I'm using coding agents, and wonder if coding doesn't require the same behavior as writing.Coding agents are absolutely a useful tool - however, that doesn't mean that pure "Vibe Coding" is the best approach for most scenarios. Using coding agents correctly requires a…
-
The statement should be: never use AI to draft text that is to be presented as your own authorial voice.Never, never, do that!I draft all my own text that anyone will ever read as “that’s James Bach talking to me.” I want no one to be able to credibly question that.But of course, writing is more than merely drafting the text. As I draft, I write…
-
Giving the compiling evidence of how addictive generative AI is actually, it actually makes perfect sense that people are looking for strategies to excuse their AI-usage. I think no matter the strategy (and given AI is actually as addictive as evidence are suggesting), chances are the usage patterns will all converge to “Just use AI dummy“.This…
-
Just this morning I used Claude to build an "omit needless words" app, which lets me paste in text and then marks up that text with strikes through the words that I can leave out.I used it to edit a document I'm working on, where I accepted about a third of its suggestions. I feel pretty good about it.If you have a Claude account you should be…
-
Reading through docs and doing research: of course. They're like much better search engines in that regard. It would be silly not to use them. Obviously, you need to review their search results against their sources.Structuring arguments: bad idea. Here you're having the model do your thinking for you. The problems that occur with having it write…
-
> Writing requires thinking, don't let an AI do the thinking for you.Okay, but reading also requires substantial thinking. Moreover, we can say the same about, say, coding. And I am sure there are people who will say, yes, don't use AI to code. But I do, and the results are amazing, and I still think very deeply: in fact, I spend more time with AI…
-
So let me get this straight:- You can use AI to look through docs and do research.- You can use AI to structure your argument.- You can NOT use AI to actually write the prose.- You can use AI to proofread, especially for how effectively you've communicated.I don't know, man, that sounds a lot like using AI to write. Your writing probably says…
-
Once upon a time the word processor was invented, and it made it so much easier to arrange your words on paper. iIf bits were wrong, you could just delete them, or copy and paste them across pages. It was actually pretty amazing what you could do; especially if you didn't mind dot-matrix type on fan-fold paper. Of course if you wanted to make a…
-
I totally agree! I think AI could be used as a discussion partner or a proofreader. Otherwise it tends to extrapolate (unnecessarily enlarge) parts that are of low value and omit or not stress enough parts that are of most importance. It does not understand or know what's inside my brain and won't take it out from there by itself. It has problems…
- Sep 19
-
> First, once the text is on the page, it’s easy to passively let the approximate word suffice, rather than thinking about word choice in the same effortful, active way we do when generating prose de novo.Meanwhile, me, as an English non-native speaker, ended up discussing two sentences I want send to HR for ten minutes while applying to a job.I…
-
This is just some sort of Luddism all over again. Sorry to be blunt, but nobody cares about your proofreading, because your text will most likely be read by another AI anyway. Your boss doesn't care about what's more convenient for you either. If you generate X amount of value for the company and your colleague generates 3X, you'll soon find…
-
I recently heard the best noun for someone who let's AI write for them: a meat proxy.Apart from the extremely good points made in the article; there's also whether you care that someone reads what you (or rather, the AI) has written. People have gone from reading it, to ignoring it (as soon as they realize it's AI), to using whatever moderation…
-
It's probably annoying the way I oversimplify things, but you reminded me of something I read somewhere.Some Jazz musician was asked to define Jazz.He couldn't really, his answer was something like: 'I don't know. But I'll know when I hear it'.To add something to the discussion directly: thinking requires vocabulary. Vocabulary is the currency of…
-
Is editing really any less 'thinking' than writing the original copy? I don't use LLMs to generate prose but I can compare to writing code. Let's say I use an LLM to write code to some specification. After it meets the spec I go over everything with a fine-toothed comb - tweaking things to a style I like better, rewriting pieces or interfaces that…
-
This resonated with me - I’ve been avoiding things that affect my word choice since mobile messaging apps started introducing “next logical word choice” features for the exact reasons outlined here. If writing approximates thought, I don’t want some third party process constantly butting in and derailing that train, even if I’m just texting…
-
I disagree, but think that there's a specific way to use AI to write. I use AI to get critiques on my writing, but never want it to actually produce text for me. I experimented with this by making https://www.writelucid.cc, and I really like it, I use it for all my writing nowadays.Getting comments on specific sentences/passages and being able to…
-
I really did like the author's framing there, but I think there is a different, simpler way to put it:Why would you think that asking someone else (that is, another human) to write something (and then reviewing it) is the same thing as writing it yourself?You may trust the other writer's opinions and knowledge, but it will not have the same tone…
-
A workflow I've recently discovered for myself that provides a kind of middle ground: I'll ask the LLM to write a first draft in a language that isn't the one the piece should ultimately be in. Then I'll use the LLM's first draft as a blueprint, to write a first draft myself in the intended target language. This fights my brain's temptation to…
-
From TFA: "Eric Schwitzgebel writes that . . . There’s a huge cognitive difference between nodding along while reading something and actually productively generating a text. Two reasons: First, once the text is on the page, it’s easy to passively let the approximate word suffice, rather than thinking about word choice in the same effortful, active…
- Sep 15
-
Don't use AI to write your posts.