- 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 "load-bearing" less than Claude's outputs, but I'm not sure it's much less AI.
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 text recur at a macro level; too-obvious patterns. Readers just want to know what the prompt was.
Writing the prose: obviously not.
Proofreading: I think this works pretty great, as long as you have a structure to how you're doing it. "How do I make this better or sharper": bad, you're having the model structure your arguments for you. But for a predefined rubric of stylistic things --- writing you know how to do, that's just tedious to check manually? Hard to see a downside.
Most people apparently still don't get that using AI for mere proofreading is the same as using it as a mere knowledge base summary tool. Instead of proofreading, have it be adversarial. Let it try to take apart your argument. The latest generation is extremely good at calling out half-truths and circumstantial evidence across many fields. These days I would not dare to put anything in front of a human peer that hasn't passed at least an AI sanity check review.
I don't want an AI's opinion on what I'm writing. I want its assessments against an objective rubric. The whole point is not to have the model weights do your thinking for you.
- 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 rewriting from scratch. Unsloppification is hard.
- AI for proofreading is wonderful. But I always prompt with "only fix grammar mistakes, do not change the style" and actually look at all changes. If not kept in check, it too eagerly homogenizes style.
Structuring your argument is probably something you should do, not an AI. This is the core of what you're trying to convey, and is exactly the type of thing you should be thinking through.
Of course, this is different than having the argument ready and using an AI to make it read a little better, which at that point, is just details.
The likes of Sir Walter Scott, Peter Ackroyd, Stephen King, Tim Ferris, Robert Greene, Tucker Max, Catherine Asaro, Neil Gaiman, Charlaine Harris, Nancy Holder etc... all rely heavily on research assistants in various capacities including proofreading.
Honest question - where does an Author like the above place in your authenticity framework?
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 reminds me a lot of smokers who all have a great strategy beat the nicotine addiction. Only smoke on the weekends when out partying with friends, only smoke to use the extra sense of energy, only smoke to suppress my appetite, go to the gym 5 hours a week to offset the negative health impacts of smoking etc. etc.
In reality most smokers will converge to about a pack a day.
there's a growing contingent of anti-ai extremists shouting down anything that looks or feels like the thing they hate: actual slop. but there's a missing understanding of where slop really comes from - "you used ai instead of thinking" is just another guess, and mostly wrong since you literally cannot use ai without thinking - the prompts don't write themselves.
slop has been around a lot longer than ai - the contingent needs to realize what it is they hate. instead of "don't use ai to write" it's "don't use ai to write slop" or better: "don't write slop"
To be clear, I (the author of the article) do not hate AI. A lot of my current work is focussed on making AI work well.
However, I don't think it's a good idea to replace our own thinking with an AI, and that's what I'm trying to avoid with the way of working proposed in the article.
No. I hate AI and slop. I hated slop before AI, but with AI slop has proliferated to an unprecedented degree. If AI went away tomorrow, slop would go down with it to 1/1000ths of what it is today.
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 better impression, you'd type it on a typewriter, or better yet, hand-calligraph your writing with a fountain pen.
So LLMs allow you to work at the speed of reading, if not faster. So you can try a paragraph here, completely rewrite the tone. Start a chapter in 1PP and just have it rewritten to 3PP. And sofort and so on. The number of new transformations you can do is endless.
What do a lot of people do? Write a one line prompt and show everyone the oneshot result.
Ok fine, sometimes you don't need to write amazing prose. But you can also be like the early writers who adopted the word processor. If you put the same amount of elbow-grease into using an LLM as you did writing before, you can get some pretty cool results.
>Once upon a time the word processor was invented, and it made it so much easier to arrange your words on paper.
It’s easier to rearrange and rethink and tinker. Hard to imagine now but it was more efficient in the days when you would have to pencil out a memo and have someone type it up, you would make corrections and the final copy would be done in a couple passes. This has been studied, efficiency has actually gone down, in favor of immediacy and direct control.
What it frees you from is organizing your thoughts ahead of time, which takes practice and discipline. Using an LLM is just taking that further, you can avoid organizing your thoughts and ideas a bit more.
Well said, the issue being, the form factor of many AI products disincentivises this behaviour. To give an example, using a CLI tool with a frontier model to write a blog (actually fairly common), you have to re-open it in another editor, potentially reformat it or convert MD>DOCX, etc etc. There's a lot of noise and mess, and people naturally tend towards "not bother" and 1-shot some "amazing" prompt (because the value of their elbow grease is disproportionately low, unless they actually have something to say- many do not).
> Start a chapter in 1PP and just have it rewritten to 3PP
I write differently with pen and paper than I do in a word processor. Sometimes one is better than the other. (Word processors are generally better, for me, at editing. One of my favorite uses for LLMs in this way is taking a picture of some text and sketches I’ve made and OCR’ing and rendering them into editable documents.)
Switching from first to third person is in part a grammatical shift. But it’s also a real change in perspective. If you’re only making the motions but not trying to see where the new perspective opens up (or closes) views into what you’re saying, you’re doing copy editing more than writing.
At least right now, this hollowness is part of what makes AI-generated text stand out.
If someone uses LLM to help wording their idea in a way that's easier to understand, I am fine with it as long as they don't copy-paste it verbatim into their blogpost.
I am getting a little ranty here, but the usual LLM-isms like "not this, but that" or "that's the real footnote here" simply add zero value to the text, and I can't imagine anyone who has proofread the LLM-generated text even once will opt to keep those sloppy paragraphs. If I see them, I can't help but to think that the author had oneshotted the text. And it strongly suggests to me that the author has put zero effort into their work.
I disagree, the problem is the tool. Generating text instead of thinking about what text to use explicitly is not writing, and it's not worthy of someone else reading it.
Fundamental to exchanging ideas is discovering which ideas should be exchanged and how to express them. The process of writing itself is that exercise. If you use a tool to outsource the thinking then you're not thinking.
I definitely think that "not thinking about what you're writing" is a big danger with AI-generated text, but it's not a given. If you generate a block of text and then go through and edit it to better reflect your ideas and voice - that's the thinking process in action. It's also entirely possible to generate slop by manually typing every word - see any corporate communication ever.
You will NEVER be a good to great writer if you use AI to seed your writing and then edit it.
_Everything_ AI generates, is generalized slop. All of your ideas will be boring and derivative. Mediocre writers will be improved, but the best would only have their ability to think dulled.
Just think through it from the perspective that AI is generating art for you instead of writing. What would all art look like if good artists used it for ideas and "edited" it? It would look like generic, derivative shit. A perpetual crutch that the person can never abandon.
Genuine experiment: Do characters "have a color" for you, maybe? [1][2]
1) "The quick brown fox jumps over the lazy dog." <- Typed by hand.
2) "The quick brown fox jumps over the lazy dog." <- Generated by ai, cut-and-pasted in. [3]
For you, would you classify the former as genuine and the latter as slop?
You'd think I'm being silly, but both yes and no are not unthinkable, and honestly I've actually seen odder arguments in court filings and language standards.
I just fundamentally do not agree with you. If you are editing AI generated text to "better reflect your ideas and voice" you are still not using your ideas or your voice. There's no 'thinking process in action' if you've delegated the hardest part of the task to a text predictor!
It really bothers me that people think it's ok to use AI as the starting point for writing. In my mind, it's the worst thing you can do. It poisons the entire effort, and frankly it's disrespectful to give it to anyone to read. Stop doing it!
There are just times that we stumble upon writer's block.
Personally I have had multiple instances of me knowing the idea I want to express, but utterly failing to articulate it in a way that's understood by most people.
And I unfortunately do not have a human assistant, so for a second opinion on how to articulate the idea better, I turn to an LLM. (Yes, I am using my own idea!) I write something, ask the LLM whether it understood what I said, and make adjustments based on that. In this case, the text is 100% written by me, and the LLM is no more different than a random passer-by giving me advices.
But there are also times when the idea is so complicated that I don't really have a good idea on how to untangle all of that. Ideas that live in your brain as a bunch of abstract concepts, if you can get it. If I were to untangle all of that myself - it's fairly good chance that I'll just leave it for "another day" (read: abandoned). I won't be investing any more thinking into it, LLM or not.
So LLM it is. I dump my ideas (as a bunch of incomprehensible human gibberish ;P) into it, make it make sense of the ideas, generate an ideas-untangled paragraph, which I use as an inspiration (or basis, or template, or whichever word you think suits it) for writing my own text with my own voice and my own style. I was able to continue thinking because the LLM keeps my attention from drifting away!
(I always suspected I have some kinda ADHD, but shrug)
I understand where people are coming from when they reject all forms of LLM assistance in writing, it's just that I hope people don't turn it into a dogma of some sort.
I have never ever seen anyone to type it on a typewriter in order to make a better impression. That was not a thing.
Likewise, the only handwriting for better impression I have seen was in stuff like love letter. It could possibly done also in super high formal situation (diplomacy, school diploma) or some such, but otherwise it was not a thing.
For most of the 20th century, ordinary people generally had the option of either handwriting or typing. Starting roughly in the 1980's consumers could get their hands on dot matrix printers, and in the 1990s inkjets and lasers started coming in. (For corporate the timeline shifts a bit earlier) .
Relevant for this conversation, dot matrix printouts were considered "cheap", and people were warned not to submit dot-matrix printouts for things like job applications, formal correspondence, etc.
If you were a kid and you tried to hand in a word processed/printed essay in school, even into the early '90s sometimes, you'd get pushback that's very similar to the AI pushback of today. "How do I know YOU wrote it?", "The computer is doing the work for you". They'd force you to rewrite it, properly, in pen.
Okay, this is just time/place/situation dependent. I remember typing job applications, that was something that communicated seriousness.
Handwritten pages always signify special care, indeed as you say that’s why you see that in diplomas and such. There was a time when I kept hand written letters going for my close college friends, on good stationary. I think we are just in different situations/places/demographics.
[edit] also I kept my job application rejections (paper mail). These were very instructive as they showed a pattern of rote rejection that made me less discouraged.
> 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 coding (and thinking) than before as I do not get hamstrung on, how do I convert this Word doc to PDF in code again? I spend more time in higher order thinking. Same with when I read text written by an AI. I don't believe that writing is any more sacrosanct that coding. And reading and adjusting AI output requires substantial thinking.
Anyone who has learned a language knows this anecdotally. The linguistic structures need to be marshalled by active cognition (editing and writing). Merely reading and watching is not enough.
Perhaps not 1:1, but, personally, I can read for much longer than I can write. And so I'd argue that this duration increase matches any deeper thought that comes with shorter writing bursts.
That should tell you something about how much effort (thinking) is required for the activity. The fact you writing is t something you can do as much as reading means you use your brain a bit more. I would claim that consuming information always uses less thought then generating information. For a human brain at least.
I've been watching chess videos for many years but my own chess playing is just recreational.
Writing/programming as a specialization of just going through the damn thing is actually sacrosanct IMO. Spectatorship as a generalization of reading is second-class. You do become a worse programmer as you move to "higher order thinking", it just matters less in a world where LLMs exist.
> And reading and adjusting AI output requires substantial thinking.
It does, but less of that thinking translates to skills transferred. What amount of the creativity and magic of Tolkien transferred to his editor? I'm sure it took substantial thinking to edit Tolkien's work.
AI can’t write for shit because it doesn’t have your experience for context and it doesn’t know what you actually want to say.
Hell, most people barely have a rough idea of what they are trying to say when they sit down to write, and they have all of the context.
Writing and coding do have many similarities, at a high level. But the things that make AI good at coding don’t really apply to writing. The analogy doesn’t work well here.
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 able to access the artifact I built here: https://claude.ai/artifact/RK7XfnaoxVV1Uji84ETaLr - it uses "Claudeception" - the feature of Claude artifacts that lets them use Claude prompts - so it will be billed against your Claude account (at a cost of probably a fraction of a cent per run.)
"The marks are suggestions to weigh, not corrections."
"that could go without changing what you mean. Your text stays exactly as you wrote it"
"Marking works when this page is open in Claude and you are signed in. It isn't available in this view"
The last one may be why I even gained the ire to interrogate the copy of the page. I was signed in g'damnit!
edit: I've seen you get quite unfairly treated in the comments section of this place, you always seemed so kind in response. Thanks for all the writing over the years!
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 careful approach as well, which also starts with thinking through the problem you're solving!
The details of workflows for coding and writing are different (in my opinion), but at a high level the same concept applies; don't let an AI do the important thinking for you!
I'll probably follow up with another blog post with what I've seen work well for coding.
I use it to write in German, because my German sucks.
I first write the thing myself, then feed it through Deepl Write to see how the AI would write it. It's been very helpful.
I write everything else myself. However I don't know how I feel about spending my time on Earth typing personal responses to AI-generated messages. Do those people deserve my time and energy?
While I do agree with the premise of the article, if find this way of thinking applicable for writing code as well. Your code is meant to not only be read by machines, but also other fellow humans. Writing code requires substantial amount of thinking as well. Shouldn't we write code without AI as well?
> Writing code requires substantial amount of thinking as well. Shouldn't we write code without AI as well?
You absolutely should. Using an LLM to write your code for you is not a good idea. It produces bad code, and it also robs you of understanding which will be valuable down the line.
Okay but I’ve spent 8 years implementing bad code so far. I’ve got a lot of understanding but only half the code is done. Down the line I’ll be dead. Why not implement it now with AI?
Moreover is writing code the only way to understand it?
Thinking and writing are complementary processes. If you write things where you don’t have to think, let the AI do the typing. No harm done. But any writing, where you actively think while you type not about what the next word should be but about the entire topic, so it yourself. Otherwise you’re not thinking and that’s pretty much the only value-add that’s left.
Don’t use AI to write and don’t use it to edit either. LLMs can’t understand your writing, so the edits and questions about your work will be less helpful than asking another person with knowledge to review.
I don't know what this is supposed to mean. An LLM can certainly understand the structure of your writing. It can accurately diagram sentences, identify the grammatical agent, note transitions between sentences and paragraphs, determine the position of modifiers, evaluate topic continuity, spot hedges and qualifiers, count word and construction use and identify overuse, and evaluate openings and stress positions of sentences.
Also, don't use AI to translate into a language you can't read back. It glosses over assumptions you didn't know you made, and the output might not mean what you intended.
We used to have machine translation that worked well enough for readers to know the overall meaning most of the time, but when it failed, it failed spectacularly and the reader would know not to trust the output.
Now with LLM translations we can’t trust anything. Even simple translations which used to work perfectly now shoves extra LLM-isms into the translation making the prose ugly.
It feels like you hit a wall. The moment for me was Astra first day of release. It felt like everything I've done up to that point was meaningless, or perhaps that's too harsh; let's call it art projects.
I've been carrying more notebooks, being fine with terminals being idle and focused on 'what is actually being created', and what will remain?
I'm on version 6 of a writing an agent voice, style, tone, and orientation guide I have been developing over the course of the past year. Every time the agent writes something using the guide, I edit it and then update the guide where needed. Slowly, the agent is beginning to sound eerily close to my natural writing voice. No, I didn't use it for this comment :)
Why is everyone suddenly asking everyone else to not use AI to write literature, essays, comments, emails, letters… but apparently writing code with AI is fine. Oh that sounds natural to you? Not to me, it doesn't. It doesn't feel natural to me writing anything with AI as a human being. But here we are.
AI for proofreading/spelling correction and reordering works quite well. My words, but perhaps structured in a way which makes more sense, with sentiment replaced by synonyms.
Basically an automatic "second-pass" which I would've caught had I re-read it, but automatically.
I said this once in my recent "Trustworthy AI" class that... "I think writing with AI is one of the most detrimental things you can do".
The reaction was pretty much silence and most people probably thought I was some angry, arrogant, anti-ai purist... which isn't the case. It can be acceptable to use a language model in code generation but requires new forms of discipline and tooling.
But for for just regular old writing... Should never, ever be generated. I'll likely never change that stance.
It's a word processor that lets you selectively accept AI suggestions, or just talk to the agent in margin comments. I've been working on this full time for over a year now.
That's not the argument in the article. Not only that but coding is deterministic up to a certain point. Once the project becomes complex, it becomes harder and harder to account for all edge cases. Sometimes it requires heuristics that are not deterministic, etc. There are infinite ways to implement the same requirements.
That argument is like saying that weather is deterministic. It is if you know every single variable, but that's impossible.
What is a soul, how do you measure and quantify it? If you can’t, how would you know whether the grammar mistakes change the soul? I’ve heard people say they keep the grammar mistakes in to prove the existence of a soul, so presumably correcting them would alter the soul in some measurable way.
I'm not sure where people are going with using LLMs for research on a topic.
LLMs construct documents like a middle manager trying to get you fired, so it's a bad idea to get a system to try to research something for you. Anytime I've used the ol' panopticon-homunculus for something applied and niche I always follow the source linked to verify it. And the source is vague and generally unverifiable. Humans can actually link other documents that are actually readable to a greater degree.
I think a large part of the blame is the dataset for large scale LLMs being thoroughly and disproportionality poisoned with popular media sites like Reddit.
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 "load-bearing" less than Claude's outputs, but I'm not sure it's much less AI.
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 text recur at a macro level; too-obvious patterns. Readers just want to know what the prompt was.
Writing the prose: obviously not.
Proofreading: I think this works pretty great, as long as you have a structure to how you're doing it. "How do I make this better or sharper": bad, you're having the model structure your arguments for you. But for a predefined rubric of stylistic things --- writing you know how to do, that's just tedious to check manually? Hard to see a downside.
Most people apparently still don't get that using AI for mere proofreading is the same as using it as a mere knowledge base summary tool. Instead of proofreading, have it be adversarial. Let it try to take apart your argument. The latest generation is extremely good at calling out half-truths and circumstantial evidence across many fields. These days I would not dare to put anything in front of a human peer that hasn't passed at least an AI sanity check review.
I don't want an AI's opinion on what I'm writing. I want its assessments against an objective rubric. The whole point is not to have the model weights do your thinking for you.
They are much worse at citing sources compared to search engines
If you can't find the source and verify it, don't use it. Seems simple.
I didn't see in the article where he recommended AI to structure an argument? He did suggest it can be helpful for structuring data.
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 rewriting from scratch. Unsloppification is hard.
- AI for proofreading is wonderful. But I always prompt with "only fix grammar mistakes, do not change the style" and actually look at all changes. If not kept in check, it too eagerly homogenizes style.
Structuring your argument is probably something you should do, not an AI. This is the core of what you're trying to convey, and is exactly the type of thing you should be thinking through.
Of course, this is different than having the argument ready and using an AI to make it read a little better, which at that point, is just details.
The likes of Sir Walter Scott, Peter Ackroyd, Stephen King, Tim Ferris, Robert Greene, Tucker Max, Catherine Asaro, Neil Gaiman, Charlaine Harris, Nancy Holder etc... all rely heavily on research assistants in various capacities including proofreading.
Honest question - where does an Author like the above place in your authenticity framework?
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 reminds me a lot of smokers who all have a great strategy beat the nicotine addiction. Only smoke on the weekends when out partying with friends, only smoke to use the extra sense of energy, only smoke to suppress my appetite, go to the gym 5 hours a week to offset the negative health impacts of smoking etc. etc.
In reality most smokers will converge to about a pack a day.
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 about ideas that other people or AI may have helped me come to.
there's a growing contingent of anti-ai extremists shouting down anything that looks or feels like the thing they hate: actual slop. but there's a missing understanding of where slop really comes from - "you used ai instead of thinking" is just another guess, and mostly wrong since you literally cannot use ai without thinking - the prompts don't write themselves.
slop has been around a lot longer than ai - the contingent needs to realize what it is they hate. instead of "don't use ai to write" it's "don't use ai to write slop" or better: "don't write slop"
To be clear, I (the author of the article) do not hate AI. A lot of my current work is focussed on making AI work well.
However, I don't think it's a good idea to replace our own thinking with an AI, and that's what I'm trying to avoid with the way of working proposed in the article.
No. I hate AI and slop. I hated slop before AI, but with AI slop has proliferated to an unprecedented degree. If AI went away tomorrow, slop would go down with it to 1/1000ths of what it is today.
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 better impression, you'd type it on a typewriter, or better yet, hand-calligraph your writing with a fountain pen.
So LLMs allow you to work at the speed of reading, if not faster. So you can try a paragraph here, completely rewrite the tone. Start a chapter in 1PP and just have it rewritten to 3PP. And sofort and so on. The number of new transformations you can do is endless.
What do a lot of people do? Write a one line prompt and show everyone the oneshot result.
Ok fine, sometimes you don't need to write amazing prose. But you can also be like the early writers who adopted the word processor. If you put the same amount of elbow-grease into using an LLM as you did writing before, you can get some pretty cool results.
>Once upon a time the word processor was invented, and it made it so much easier to arrange your words on paper.
It’s easier to rearrange and rethink and tinker. Hard to imagine now but it was more efficient in the days when you would have to pencil out a memo and have someone type it up, you would make corrections and the final copy would be done in a couple passes. This has been studied, efficiency has actually gone down, in favor of immediacy and direct control.
What it frees you from is organizing your thoughts ahead of time, which takes practice and discipline. Using an LLM is just taking that further, you can avoid organizing your thoughts and ideas a bit more.
See also: https://news.ycombinator.com/item?id=49795122 "Socrates vs. the Written Word (2011)"
Well said, the issue being, the form factor of many AI products disincentivises this behaviour. To give an example, using a CLI tool with a frontier model to write a blog (actually fairly common), you have to re-open it in another editor, potentially reformat it or convert MD>DOCX, etc etc. There's a lot of noise and mess, and people naturally tend towards "not bother" and 1-shot some "amazing" prompt (because the value of their elbow grease is disproportionately low, unless they actually have something to say- many do not).
Are people really publishing blogs in docx format?
> Start a chapter in 1PP and just have it rewritten to 3PP
I write differently with pen and paper than I do in a word processor. Sometimes one is better than the other. (Word processors are generally better, for me, at editing. One of my favorite uses for LLMs in this way is taking a picture of some text and sketches I’ve made and OCR’ing and rendering them into editable documents.)
Switching from first to third person is in part a grammatical shift. But it’s also a real change in perspective. If you’re only making the motions but not trying to see where the new perspective opens up (or closes) views into what you’re saying, you’re doing copy editing more than writing.
At least right now, this hollowness is part of what makes AI-generated text stand out.
You nailed it on what I've always wanted to say.
If someone uses LLM to help wording their idea in a way that's easier to understand, I am fine with it as long as they don't copy-paste it verbatim into their blogpost.
I am getting a little ranty here, but the usual LLM-isms like "not this, but that" or "that's the real footnote here" simply add zero value to the text, and I can't imagine anyone who has proofread the LLM-generated text even once will opt to keep those sloppy paragraphs. If I see them, I can't help but to think that the author had oneshotted the text. And it strongly suggests to me that the author has put zero effort into their work.
Agree completely. The problem is not the tool, it's all the pointless filler (or even hallucinations) that the tool generates that people leave in.
I disagree, the problem is the tool. Generating text instead of thinking about what text to use explicitly is not writing, and it's not worthy of someone else reading it.
Fundamental to exchanging ideas is discovering which ideas should be exchanged and how to express them. The process of writing itself is that exercise. If you use a tool to outsource the thinking then you're not thinking.
I definitely think that "not thinking about what you're writing" is a big danger with AI-generated text, but it's not a given. If you generate a block of text and then go through and edit it to better reflect your ideas and voice - that's the thinking process in action. It's also entirely possible to generate slop by manually typing every word - see any corporate communication ever.
You will NEVER be a good to great writer if you use AI to seed your writing and then edit it.
_Everything_ AI generates, is generalized slop. All of your ideas will be boring and derivative. Mediocre writers will be improved, but the best would only have their ability to think dulled.
Just think through it from the perspective that AI is generating art for you instead of writing. What would all art look like if good artists used it for ideas and "edited" it? It would look like generic, derivative shit. A perpetual crutch that the person can never abandon.
Genuine experiment: Do characters "have a color" for you, maybe? [1][2]
1) "The quick brown fox jumps over the lazy dog." <- Typed by hand.
2) "The quick brown fox jumps over the lazy dog." <- Generated by ai, cut-and-pasted in. [3]
For you, would you classify the former as genuine and the latter as slop?
You'd think I'm being silly, but both yes and no are not unthinkable, and honestly I've actually seen odder arguments in court filings and language standards.
[1] https://ansuz.sooke.bc.ca/entry/23 "What color are your bits"
[2] https://journal.stuffwithstuff.com/2015/02/01/what-color-is-... "what color is your function"
[3] Sonnet 5, Medium, Sept 2026.
I just fundamentally do not agree with you. If you are editing AI generated text to "better reflect your ideas and voice" you are still not using your ideas or your voice. There's no 'thinking process in action' if you've delegated the hardest part of the task to a text predictor!
It really bothers me that people think it's ok to use AI as the starting point for writing. In my mind, it's the worst thing you can do. It poisons the entire effort, and frankly it's disrespectful to give it to anyone to read. Stop doing it!
There are just times that we stumble upon writer's block.
Personally I have had multiple instances of me knowing the idea I want to express, but utterly failing to articulate it in a way that's understood by most people.
And I unfortunately do not have a human assistant, so for a second opinion on how to articulate the idea better, I turn to an LLM. (Yes, I am using my own idea!) I write something, ask the LLM whether it understood what I said, and make adjustments based on that. In this case, the text is 100% written by me, and the LLM is no more different than a random passer-by giving me advices.
But there are also times when the idea is so complicated that I don't really have a good idea on how to untangle all of that. Ideas that live in your brain as a bunch of abstract concepts, if you can get it. If I were to untangle all of that myself - it's fairly good chance that I'll just leave it for "another day" (read: abandoned). I won't be investing any more thinking into it, LLM or not.
So LLM it is. I dump my ideas (as a bunch of incomprehensible human gibberish ;P) into it, make it make sense of the ideas, generate an ideas-untangled paragraph, which I use as an inspiration (or basis, or template, or whichever word you think suits it) for writing my own text with my own voice and my own style. I was able to continue thinking because the LLM keeps my attention from drifting away!
(I always suspected I have some kinda ADHD, but shrug)
I understand where people are coming from when they reject all forms of LLM assistance in writing, it's just that I hope people don't turn it into a dogma of some sort.
> Of course if you wanted to make a better impression, you'd type it on a typewriter, or better yet, hand-calligraph your writing with a fountain pen.
Why do people feel the need to invent the past that never existed?
I lived in that past myself, what are you thinking about?
I have never ever seen anyone to type it on a typewriter in order to make a better impression. That was not a thing.
Likewise, the only handwriting for better impression I have seen was in stuff like love letter. It could possibly done also in super high formal situation (diplomacy, school diploma) or some such, but otherwise it was not a thing.
For most of the 20th century, ordinary people generally had the option of either handwriting or typing. Starting roughly in the 1980's consumers could get their hands on dot matrix printers, and in the 1990s inkjets and lasers started coming in. (For corporate the timeline shifts a bit earlier) .
Relevant for this conversation, dot matrix printouts were considered "cheap", and people were warned not to submit dot-matrix printouts for things like job applications, formal correspondence, etc.
If you were a kid and you tried to hand in a word processed/printed essay in school, even into the early '90s sometimes, you'd get pushback that's very similar to the AI pushback of today. "How do I know YOU wrote it?", "The computer is doing the work for you". They'd force you to rewrite it, properly, in pen.
Okay, this is just time/place/situation dependent. I remember typing job applications, that was something that communicated seriousness.
Handwritten pages always signify special care, indeed as you say that’s why you see that in diplomas and such. There was a time when I kept hand written letters going for my close college friends, on good stationary. I think we are just in different situations/places/demographics.
[edit] also I kept my job application rejections (paper mail). These were very instructive as they showed a pattern of rote rejection that made me less discouraged.
That past certainly did exist, within living memory (mine, for one).
> 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 coding (and thinking) than before as I do not get hamstrung on, how do I convert this Word doc to PDF in code again? I spend more time in higher order thinking. Same with when I read text written by an AI. I don't believe that writing is any more sacrosanct that coding. And reading and adjusting AI output requires substantial thinking.
We have copious evidence for behaviors like note taking and writing creates deeper thinking and retention that passively watching / reading
https://www.scientificamerican.com/article/why-writing-by-ha...
Anyone who has learned a language knows this anecdotally. The linguistic structures need to be marshalled by active cognition (editing and writing). Merely reading and watching is not enough.
> Okay, but reading also requires substantial thinking.
I don't think anyone would argue reading a book takes as much thinking as writing one…
Perhaps not 1:1, but, personally, I can read for much longer than I can write. And so I'd argue that this duration increase matches any deeper thought that comes with shorter writing bursts.
That should tell you something about how much effort (thinking) is required for the activity. The fact you writing is t something you can do as much as reading means you use your brain a bit more. I would claim that consuming information always uses less thought then generating information. For a human brain at least.
I've been watching chess videos for many years but my own chess playing is just recreational.
Writing/programming as a specialization of just going through the damn thing is actually sacrosanct IMO. Spectatorship as a generalization of reading is second-class. You do become a worse programmer as you move to "higher order thinking", it just matters less in a world where LLMs exist.
> And reading and adjusting AI output requires substantial thinking.
It does, but less of that thinking translates to skills transferred. What amount of the creativity and magic of Tolkien transferred to his editor? I'm sure it took substantial thinking to edit Tolkien's work.
AI can’t write for shit because it doesn’t have your experience for context and it doesn’t know what you actually want to say.
Hell, most people barely have a rough idea of what they are trying to say when they sit down to write, and they have all of the context.
Writing and coding do have many similarities, at a high level. But the things that make AI good at coding don’t really apply to writing. The analogy doesn’t work well here.
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 able to access the artifact I built here: https://claude.ai/artifact/RK7XfnaoxVV1Uji84ETaLr - it uses "Claudeception" - the feature of Claude artifacts that lets them use Claude prompts - so it will be billed against your Claude account (at a cost of probably a fraction of a cent per run.)
"The marks are suggestions to weigh, not corrections."
"that could go without changing what you mean. Your text stays exactly as you wrote it"
"Marking works when this page is open in Claude and you are signed in. It isn't available in this view"
The last one may be why I even gained the ire to interrogate the copy of the page. I was signed in g'damnit!
edit: I've seen you get quite unfairly treated in the comments section of this place, you always seemed so kind in response. Thanks for all the writing over the years!
Hah, yeah the copy on that app is pretty bad Claude writing, I should run it against itself!
This is a neat workflow. Incentivises you to proactively think about your writing, even as you cut it.
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 careful approach as well, which also starts with thinking through the problem you're solving! The details of workflows for coding and writing are different (in my opinion), but at a high level the same concept applies; don't let an AI do the important thinking for you!
I'll probably follow up with another blog post with what I've seen work well for coding.
I use it to write in German, because my German sucks.
I first write the thing myself, then feed it through Deepl Write to see how the AI would write it. It's been very helpful.
I write everything else myself. However I don't know how I feel about spending my time on Earth typing personal responses to AI-generated messages. Do those people deserve my time and energy?
While I do agree with the premise of the article, if find this way of thinking applicable for writing code as well. Your code is meant to not only be read by machines, but also other fellow humans. Writing code requires substantial amount of thinking as well. Shouldn't we write code without AI as well?
If the point of the code is to be read, maybe.
I don't write code for my colleague's pleasure, I do it to solve a problem as fast and efficiently as possible.
I write code to be read by just one guy.
He’s five years older, a bit more tired than me, more going on in his life. Easily annoyed by bad code.
Weirdly enough he has exactly the same name and birthday as me.
> Writing code requires substantial amount of thinking as well. Shouldn't we write code without AI as well?
You absolutely should. Using an LLM to write your code for you is not a good idea. It produces bad code, and it also robs you of understanding which will be valuable down the line.
Okay but I’ve spent 8 years implementing bad code so far. I’ve got a lot of understanding but only half the code is done. Down the line I’ll be dead. Why not implement it now with AI?
Moreover is writing code the only way to understand it?
Thinking and writing are complementary processes. If you write things where you don’t have to think, let the AI do the typing. No harm done. But any writing, where you actively think while you type not about what the next word should be but about the entire topic, so it yourself. Otherwise you’re not thinking and that’s pretty much the only value-add that’s left.
Don’t use AI to write and don’t use it to edit either. LLMs can’t understand your writing, so the edits and questions about your work will be less helpful than asking another person with knowledge to review.
I don't know what this is supposed to mean. An LLM can certainly understand the structure of your writing. It can accurately diagram sentences, identify the grammatical agent, note transitions between sentences and paragraphs, determine the position of modifiers, evaluate topic continuity, spot hedges and qualifiers, count word and construction use and identify overuse, and evaluate openings and stress positions of sentences.
Also, don't use AI to translate into a language you can't read back. It glosses over assumptions you didn't know you made, and the output might not mean what you intended.
I want LLM translations to die.
We used to have machine translation that worked well enough for readers to know the overall meaning most of the time, but when it failed, it failed spectacularly and the reader would know not to trust the output.
Now with LLM translations we can’t trust anything. Even simple translations which used to work perfectly now shoves extra LLM-isms into the translation making the prose ugly.
this SPECIALLY for arabic the shit I see everyday
It feels like you hit a wall. The moment for me was Astra first day of release. It felt like everything I've done up to that point was meaningless, or perhaps that's too harsh; let's call it art projects.
I've been carrying more notebooks, being fine with terminals being idle and focused on 'what is actually being created', and what will remain?
Similar recent discussion: https://news.ycombinator.com/item?id=49767937
I'm on version 6 of a writing an agent voice, style, tone, and orientation guide I have been developing over the course of the past year. Every time the agent writes something using the guide, I edit it and then update the guide where needed. Slowly, the agent is beginning to sound eerily close to my natural writing voice. No, I didn't use it for this comment :)
Why is everyone suddenly asking everyone else to not use AI to write literature, essays, comments, emails, letters… but apparently writing code with AI is fine. Oh that sounds natural to you? Not to me, it doesn't. It doesn't feel natural to me writing anything with AI as a human being. But here we are.
AI for proofreading/spelling correction and reordering works quite well. My words, but perhaps structured in a way which makes more sense, with sentiment replaced by synonyms.
Basically an automatic "second-pass" which I would've caught had I re-read it, but automatically.
Proofreading yes, automatic "corrections" no
I agree. Writing manually and thoughtfully and then letting the AI figure out the nitty gritty implementation works very well.
My brain immediately went to "I wonder how many people got this article and used AI to summarize it for them."
I said this once in my recent "Trustworthy AI" class that... "I think writing with AI is one of the most detrimental things you can do".
The reaction was pretty much silence and most people probably thought I was some angry, arrogant, anti-ai purist... which isn't the case. It can be acceptable to use a language model in code generation but requires new forms of discipline and tooling.
But for for just regular old writing... Should never, ever be generated. I'll likely never change that stance.
Indeed, even the best models output needs to be checked thoroughly
Here is a good explanation in a deeper way of why using AI to write is not a good thing
https://markgreville.ie/2024/12/10/generative-ai-is-digital-...
Why not ?
Shameless plug of my app! https://revise.io/editor
It's a word processor that lets you selectively accept AI suggestions, or just talk to the agent in margin comments. I've been working on this full time for over a year now.
The article opens with
> Writing requires thinking, don't let an AI do the thinking for you.
And then the author goes on to admit that they use AI for coding. Does coding not require thinking or what?
I agree that we shouldn't be delegating thinking to AI, but stopping at writing seems disingenuous.
> Does coding not require thinking or what?
Code is deterministic, and language is not. If I write:
total = 0 for x in [1, 2, 3]: total += x
…nobody is going to argue with me that “total” is anything but 7.
But language is different. What you perceive and what I perceive from the written word can be completely different (e.g https://youtube.com/watch?v=naleynXS7yo&ra=m )
That's not the argument in the article. Not only that but coding is deterministic up to a certain point. Once the project becomes complex, it becomes harder and harder to account for all edge cases. Sometimes it requires heuristics that are not deterministic, etc. There are infinite ways to implement the same requirements.
That argument is like saying that weather is deterministic. It is if you know every single variable, but that's impossible.
Yes but on the sliding scale of how deterministic is this task- writing, not very. coding, can execute and write some tests (even if gamed/bad).
wow really?
We can write draft by using our thoughts.. and check for grammer mistakes thats it.
Our usual flow of writing should not be disturbed at all..
Fixing grammer mistakes wont change the soul.
> Fixing grammer mistakes wont change the soul.
Pangram disagrees, if your grammar was fixed by AI it's "100% Written by AI"
What is a soul, how do you measure and quantify it? If you can’t, how would you know whether the grammar mistakes change the soul? I’ve heard people say they keep the grammar mistakes in to prove the existence of a soul, so presumably correcting them would alter the soul in some measurable way.
- why stop at writing? put this advice for coding too
- sure its good to ask AI feedback about your module or code review or identify vulnerabilities or better ways of doing your existing code
- but maybe stop asking AI to generate entire projects from scratch?
I'm not sure where people are going with using LLMs for research on a topic.
LLMs construct documents like a middle manager trying to get you fired, so it's a bad idea to get a system to try to research something for you. Anytime I've used the ol' panopticon-homunculus for something applied and niche I always follow the source linked to verify it. And the source is vague and generally unverifiable. Humans can actually link other documents that are actually readable to a greater degree.
I think a large part of the blame is the dataset for large scale LLMs being thoroughly and disproportionality poisoned with popular media sites like Reddit.