I think you should almost never use AI to write
erichgrunewald.substack.com
erichgrunewald.substack.com
This is the best articulation I've seen of why simply reviewing and copy-editing does not provide remotely the same value as writing from scratch. I spent a considerable amount of time over the past two weeks reviewing and improving a work document that was the output of an LLM. Given the number of people involved and the final level of effort, I'm firmly convinced that writing it manually would have been faster and resulted in a higher quality product. Getting the wording right matters.
Not that it's necessarily better but it's the kind of thing an editor might pick up and so why would you reject the advice if it's a machine giving it rather than a human?
See also this existing comment: https://news.ycombinator.com/item?id=49768564, which I completely agree with and which is complementary to the above root comment.
What do you see as the difference?
If you want to discuss something else just let me know.
But in your own mind, your thoughts are still incomplete. The act of writing them down (or, I imagine, in an oral tradition somehow committing to a specific narration) is very important. IMO, when you delegate that to an LLM, you are not really thinking. It's the same as if you explained your ideas to someone (eg in an interview), and then they ghost wrote the work for you.
And then if you want to tell a message with a different tone, again, you must consider differently.
The considering can be strengthened with exercise.
I can't imagine worrying about signaling as I write: what a huge distraction.
The problem is: this type of thinking and writing takes a lot more time and attention.
How much consideration should you put into a message? It's a personal answer, but also one that can be constrained by time.
I was talking about the step before that: what are your ideas?
How do you circle an idea in your own mind: your train of thought, in order to express an idea?
Vocabulary to think, and then if we need to share, we use careful to pick words with the reader in mind.
But just in your own head they lack clarity. I think this can be deceptive to people, it’s a blind spot. How often have you sat down to talk over some disagreement with someone only to find their own thoughts are a jumble of disjointed ideas that don’t make sense?
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, structure, word choice, understanding, or narrative flow as it would if you were to write it yourself.
And when it's an LLM, you should not trust it's "opinions" and "knowledge", because it does not have either of those things. The appearance of those things is just that, an appearance.
But in professional settings, a lot more of the informativeness is about the author, and a lot more of the persuasiveness is I'm worth your time and money. So, if the author is an LLM, and obviously so, what exactly are you informing your audience of (about yourself), and what are you persuading them to do (with your article).
I think we now know.
The whole catch is that they can often be regenerated. But LLMs (on their own, in their parametric memory - which is the result of training) don't have any conception of whether what they've generated is a real reproduction of some training material or whether they've invented something false that seemed probable based on their encoded stats. When the probability produces something contrary to what was in the training material, you get hallucinations.
They're very, very good predictive text models and can be very, very powerful when hooked up to other tools or outside databases. But its fundamentally lossy technology and all the books having been fed in doesn't guarantee all of the knowledge from those books can be spat back out.
Every performance- singing a song, playing an instrument, performing a stand-up routine, giving a speech, performing a theatrical role are all things that are best done from memory but require practice.
There are pneumonic tricks you can use- I've seen some people do it for tricks like memorizing the order of a deck of cards- but it's less useful for long term recital because it helps with order but not comprehension or fast indexing.
I don't really understand this side of the debate other than as a gotcha tbh.
If LLM use atrophies your brain and skills that's bad. If it has a repulsive writing style that's bad.
I'm not sure what the debate about whether an AI is a statistical parrot unlike humans accomplishes. Is relying 100% on a bad human speechwriter somehow better?
The primary complaint seems to be that LLMs are held to a higher standard than humans, though I don't particularly buy that line of reasoning.
LLMs are great to make drafts if you give them the source materials. They're great at validation if you give them the tools. They are great at layouting if you give them linters.
Encode architecture and decisions in your workflow, then proofread what your agents have been working on. Not the other way around.
Better validation and testing means more work will transform from exhaustive decision work to automateable gruntwork.
And there are countless other jobs out there where people are the “voice” of another party. From those who manage social media presence, to PR firms, to copyrighters, all who put statements out on others behalf.
There was already an industry of professionals whose job it was to have others write things on our behalf, and that existed long before LLMs were a thing. What LLMs did was make that that service available much more cheaply.
And no LLMs are not producing ‘the same thing’ as a speechwriter. If you are unable to tell the difference I’d suggest doing some a lot more reading of human books. LLMs produce stultifying pablum without coherence or style.
Nope. The exact opposite in most cases. Eg Jon Favreau (Obamas speechwriter) has said:
“As a speechwriter, your ego has to take a backseat. The goal is to make the speaker sound like the best version of themselves, rather than to showcase your own cleverness."
> And no LLMs are not producing ‘the same thing’ as a speechwriter.
I didn’t say it was the same. I said it was an existing industry that AI has stepped into and AI is taking the low end of market.
It’s just like how AI has captured the low-end of many tech roles too.
> If you are unable to tell the difference I’d suggest doing some a lot more reading of human books.
“Some a lot more”? Very ironically timed editing error there. You can bet an LLM wouldn’t have made that mistake ;)
AI has replaced neither full tech roles nor speechwriters.
I think that’s a good example of the kind of trivial error humans make all the time when not rereading/editing. When humans make mistakes, it’s usually that kind of mistake, which IMO doesn’t really matter in an internet comment.
Unfortunately LLMs make gross errors of style and content and often just don’t make any sense in long form text. That’s a very different category of error.
That would explain why he might say that if directly asked. But not why he volunteers that information nor talks about his research into Obamas speaking style before writing his first speech on Obamas behalf. Nor any of the other detail the he, and other speechwriters discuss when talking about their job.
You accused me of not reading enough before, and yet here you are making bold claims about a profession you’ve clearly not read enough about.
> AI has replaced neither full tech roles nor speechwriters
As I said in my previous post: i claimed AI had replaced all people. I said it’s captured the low end of the market.
> I think that’s a good example of the kind of trivial error humans make all the time when not rereading/editing. When humans make mistakes, it’s usually that kind of mistake, which IMO doesn’t really matter in an internet comment.
If you’re referring to your error, then it was still mistake regardless of the nature of it. And yes it does matter. It matters because if you want to talk about professional-quality writing, it’s the kind of mistake LLMs wouldn’t make.
> Unfortunately LLMs make gross errors of style and content and often just don’t make any sense in long form text. That’s a very different category of error.
Your comment technically didn’t make any sense. I was able to extract the intended meaning, but that wasn’t because your comment was well written.
And if we are going back to the speechwriting example, such prose isn’t read to an audience without the speaker reviewing and iterating with the speechwriter. And neither should LLM output be submitted without review.
Your argument here is effectively saying “AI isn’t as good as any humans because I expect it to perform better than the top tier of professionals.”
Clearly that’s a biased benchmark.
Whereas what I’m saying is “AI isn’t as good as top level professionals but it already offers a good-enough alternative for the entry level requirements”. Which is a completely different yardstick.
And all of this arose from your original comment that AI mimics voices so isn’t any use while ignoring the evidence provided that professional writers do exactly the same in many fields. Again, you’re deliberately skewing the facts to suit a position you’ve already pre-determined.
Now I’m not going to argue that AI is a net benefit for society nor that AI is an ethical technology. Frankly, I prefer life pre-LLMs too. But that’s a personal opinion separate from the facts of the conversation.
Obama, former president of Harvard Law Review, is far from inarticulate and unpolished. The truth you are ignoring is that speechwriters free the speaker from long hours of speechwriting, analogous to a chauffeur enabling them to work while commuting (chauffeurs are not principally used by incompetent drivers).
Every politician - even Trump - reviews their speeches and edits to better fit their desired goals. Speechwriters are principally labor-savers.
How do you recognize someone as having true opinions and knowledge from someone having just appearance of them?
Going from foreign to native is easier than native to foreign. The latter requires completely different brain paths and a lot more understanding of the language to actually get to something correct.
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 thought.
You can usually express an idea with a few thoughts, or many. The audience, and amount of details chosen should always be kept in mind. Writing helps you to remember vocabulary and word choice when expressing ideas.
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 do believe there's generally a bias to accept something that's already written. The much bigger reason though is why you let somebody else write it to begin with.
It might just be that not thinking carefully about every sentence/wording was the exact thing that made you use AI to begin with.
> I spent a considerable amount of time over the past two weeks reviewing and improving a work document that was the output of an LLM.
I had the same experience with texts where I have a very detailed expectation of the desired result. This is just a general limitation. For code, there's the saying "the precise description of the solution is already the code". Describing X is a simplification of X, oftentimes it's fine guess the gaps. But when it's not, it didn't help to describe X, you have to manifest X itself.
Don’t use AI to write things that you are producing for someone else to consume.
Elsewhere it’s painful and irksome when you didn’t ask for it
Perhaps English is not your first language? To people who are well-read, LLM-ism grate on their nerves like the high-pressure sales pitch you hear on 3am shopping channels.
The other difference is that, if I get AI to write something for me, I expect something AI-written. If I am writing something for you, you expect something that I wrote, not something that AI wrote.
So the intellectual responsibility is diluted to a substantial degree: The text is not the opinion of anyone who can be expected to honestly defend it. It's bullshit in the technical sense.
They can be broken and jumbled, or terse AF, or combine to be Pulitzer-quality prose. Either way: I want to consume their own unique human expression of a concept.
There's value in raw human expression, including in the missteps.
If I instead want the regurgitated waggyings of a bot, then: I know how to get that on my own.
We've all got web browsers and pocket supercomputers. We've all (well, most of us) been alive and aware of LLMs since their recent rise from infancy.
It's a no-brainer for us to paste some paragraphs into our favorite chatbot and get a summary or an artificial expansion or whatever else we wish to have. If that's what our goal is, then we can do that on a whim -- and we can still retain the original expression.
But doing it on our behalf is deleterious, unsettling, and unhelpful. It has negative value to the beholder.
There's a (probably subconscious) reason for you wanting that (and I just posted this elsewhere on a different story, but it bears repeating): https://news.ycombinator.com/item?id=49767867
The same is true if the thing they both want is a stale cupcake bought from a gas station. But this still doesn't mean that person B wants to be presented with person A's stale thing, because the overhead of the social transaction isn't worth the object. Which is to say, it makes a terrible gift.
The exercise of writing is an important step of actually understanding your own thoughts properly. Readers are reading to get a sense of your own experience and ideas on the topic, if it turns out you didn't actually write the final thing yourself they'll always feel a sense – in whatever way – of being conned by the author. Intentional or not.
Technical documentation, where what the reader wants to come away with is the clear facts on what something does and how to use it/whatever, is a different matter. But even then editing is key – overlong text, unnecessary information, these will again fight against what the reader is there to get.
With that said, if you do want to do it, a disclaimer is at least upfront and may well show there's no intentional attempt to deceive.
I don’t agree with this one. Writing technical documentation can be a slog, but quite often you will discover very annoying design decisions in the process. Like you’ll write a tutorial for a CLI and realize that actually some flag makes no sense, or some default should be changed.
This also goes for dogfooding. If you’re using AI to write and run the tools, try to use them yourself and see how nice they are. AI will happily run whatever jank it’s working with.
Docstrings perhaps, but even then I find myself editing heavily to avoid cruft sneaking in.
TLDR: Communication from a human is not only the material being communicated, it's also a signal that the human understands the message they are sending you. When you pass that message via an intermediary (regardless of whether the intermediary is a human or an LLM), the receiver is never sure that what they received is what you intended.
It's doubly worse when you review it, notice a few additions and think "Great - that makes me sound smart and knowledgeable about this topic", because the receiver now has an inaccurate view about your skill.
At any rate, using an LLM to construct a message for someone else to read is basically sending the wrong signals about your skill levels.
(This probabyl explains why so many people are so ready to use an LLM to write the material - in their minds, it makes them sound smarter than they actually are, but as we've seen from people on this forum vigorously defending the practice with "pangram is wrong" and similar nonsense, they themselves are not skilled enough to recognise slop)
100% right. The fundamental point here is:
- it's OK to be wrong
- it's OK to be wrong and not know
- it's a little less OK to be wrong when you should have known
- it's much less OK to be wrong when you haven't put any effort into being right
- it's far less OK to be wrong when you are pretending to know either way
Relying on AI writing risks pushing you further down the path to worse kinds of wrongness.
I'm not saying you should never do this: our time is valuable, and we shouldn't spend it on things that are not genuinely worthwhile to us if we can help it. But we're still losing something by having an LLM write for us, even if the intended audience is just ourselves.
I do not personally even think AI is useful for editorial purposes; I would rather read and re-read my own text and edit it down (which, despite my overlong comments here, I actually do), than have someone else do it. The reason is simple: editing is my search for the clearest way to express my thoughts.
I tend to turn grammar checkers off, but leave spelling checkers on, because they catch typos and a handful of my spelling gremlins like "liaise".
One of the most egregious negation issues I run into a lot is when I (or someone) makes a statement of the form: "not X" or "X is thus not true", and the AI then proceeds to interpret or summarize this as 'whatever is the opposite of X is the case'". This will cause it to go down a useless path investigating or disputing the opposite of X, which generally has no relevance or bearing on anything.
It also often very harmfully will replace your carefully chosen words with weirdly specific academic operationalizations or formalisms, then again waste huge amounts of text refuting / showing "problems" that result from that formalism, all of which again have no bearing or relevance on the original statement. An example would be you saying something like "intelligence, generally, must surely explain some of the differences in X", and then it will go "actually IQ does not correlate with X", unless you specifically tell it not to conflate psychometric IQ with intelligence generally.
Sometimes this is helpful, but the more specific / technical the domain, the more often you specifically have to prevent it from going down stupid paths that should be obvious given the expert context and wording, because it can seem almost hungry to try to catch you in some kind of insipid 'gotcha'. Much of these issues often clearly arise immediately from the first-pass "reword what the user said" part, given the reasoning traces.
If we don't want to be writers, then we have to be editors. And editing is an entirely different job and it's not an easy one. In many ways it's harder.
Especially when LLMs love writing novels when all we need is a short story or less.
The LLM will always give you a full rewrite — don't use it. It always does too much, and persuading it to tone it down is a constant battle.
Writing code is thinking, AI code is often vague and wrong in hard-to-notice ways, and the (human) reader of code is the one that pays the cost for this later.
The cost/benefit analysis may still work out differently for code though...
My reason: code can be checked objectively. I can run it and confirm it works. I don't get attached to it. I don't feel pride in it (even when I write it by hand). Code just is. It's lifeless, inert, and entirely replaceable.
How do I do the equivalent for prose? How can I tell if my words "work"? Do they clearly convey my ideas to the intended audience? There's an element of subjectivity here forces me to identify personally with the prose.
Code has no such personality. I don't tie my identity or ego to code the same way I would an essay.
Running the code only confirms that it works with the precise input, in the precise environment, under the precise circumstances you run it under. It doesn’t ensure that the code is correct. Thinking through the code, on the other hand, lets you consider all possible cases. It’s the difference between experiment and (mathematical) proof.
For an objective correctness proof, using a formal language is indispensable.
In practice, for code, testing code is the way to go. Formal proofs work, too, but the barrier to entry is high and it’s overkill for most business applications. LLMs can be very good at writing tests, if you read the test thoroughly and analyze them.
Hard for me to imagine. You feel no pride in using a tool to accomplish a goal?
> Code has no such personality. I don't tie my identity or ego to code the same way I would an essay
Code certainly does have a personality. When working with teams for a while you can absolutely get a sense for which person wrote what code in a codebase, just by subtle little tells.
You may not tie your identity or ego to it, bully for you, but for me I take a lot of pride in writing clear and maintainable code that contributes to big projects in meaningful ways.
Maybe the problem with software is there's too many people who treat writing code as a mere means to an end, instead of a very important part of the process.
Code is just sitting in a git repo somewhere, not necessarily running. That's a big distinction for me. Consider that code volume has increased 14x on github this year, but we see nowhere near that increase in the actual usable software. Code is cheap and getting cheaper. Running software ain't.
Another way to put it: I'm only interested in code so far as the value it provides. That value, not the code itself, is the source of pride. If I can provide similar value without any code at all, I'd gladly do so.
The danger for me in LLM code is the same as in writing, it's just that I'm not typically writing at the same scale as when I'm building something. The final piece when I'm writing is usually a message or a 1,000 word article at most. So I'm naturally going to analyze it quite intensely, because I can afford to. And I don't really use LLMs for this at all. I use them for things around writing (research, interrogating ideas, situationally specific stuff, mapping, visuals, publishing, etc.) And the code equivalent to an essay or message would probably be something like a single script, or the sort of thing I'd write as example code when I'm teaching. In those settings, again, LLMs can be helpful, but I'm still going to be really opinionated at a highly detailed resolution.
But a codebase is more comparable to a novel than an essay. Or more directly, the writing in a codebase is usually the documentation, which grows commensurately with the codebase. And the real LLM risk here is the drift that can happen over the course of many epics or "chapters" as the LLM writes "code that works but is imprecise and probably shouldn't work this way" or introduces weird new terminology that neither of us can precisely define. Worse, this usually becomes obvious down the line, and I have to parse through the verbose constructed world the agent has created to trace the issue back. That's a big cognitive tax, because I'm holding these weird parallel worlds of "How did the LLM's alien brain get here within the bounds of the contracts" and "What do I really want this to look like".
So I think it's fundamentally the same phenomenon, and we're all developing our skills around working with it in real time.
All of the arguments here apply to writing code; Why would the quality of writing differe when writing prose vs code?
I cannot read LLM slop prose without getting mentally fatigued trying to figure out what it is saying, and I've discovered that the quality of the code it produces has the same effect on me.
LLM writing takes your prompt and adds stuff you didn't write, burying your meaning and making it harder for the reader to get your message.
Way more effective to just publish the prompt.
Of course, it's great if you're just trying to fill space or satisfy demands of a bullshit job.
I compared at final edited output less boilerplate vs sum(prompts).
If I had to guess I'd say far more people are better at editing than they are at writing de novo. My approach probably limits the creativity of the output but I'm not writing a novel.
Only ever use AI to make yourself think harder, and more.
I agree that being lazy about writing and simply using a short prompt or list is detrimental if you are replacing your own output but you don't have to do that and you don't have to accept any of the output either. The best chats I've had usually start with a large amount of my own writing up front, thinking through the idea, listing several alternative ideas, asking a lot of questions, jotting down related topics, etc. and then reading the resulting output and critiquing it, asking for clarification, doing my own research on it, even just discussing it with the AI and doing this for a number of turns until I feel like I've exhausted the topic in the chat. I then often take what I've shaped in my own mind from that process and write something myself, either that or take the best parts of the output and edit, rearrange, reinterpret, and add to it in order to produce something.
I guess it all comes down to whether you are actively engaging with the material, regardless if that material is the result of a Google search, pulled from a book or generated by AI. If you just read it or copy paste it and don't engage with it and think about it then it doesn't do much good.
LLMs are helpful because their idiom-list is vast and they are indefatigable and infinitely patient. You can keep iterating on a sentence and it will keep giving you fresh takes. Eventually, whether through careful steering or brute-force iteration, it will come up with something that has the right resonance. Like a word on the tip of your tongue, you recognize it when you hear it.
For me, the end result of this writing-process is often a document where much of the language first came from the agent, but the voice is recognizably mine, and I feel a clear sense of authorship. It is still my work because of the microscopic attention I paid to every word. The marble is the agent's, but the chisel and mallet are in my hands.
https://www.pangram.com/history/3d55b442-d151-4490-9c71-28d0...
This expresses quite well an experience that I have sometimes had. Your first words can trap you in a box that you know isn't right, but you can't figure out how to escape.
“I think you should almost never use AI to write -- that is, to do the thing you’re doing when you type words on a page -- whether for a blog post, a research report, a memo,”
I have a much better theory: as it happens, Substack automatically converts "--" to em-dash as you type. I'm guessing the author knew that and wanted to use this feature, but composed the article in another editor and then copied-and-pasted the whole thing. Because of how Substack implements the feature, the substitution doesn't actually kick in on Ctrl-V, so all these "--" remained in place.
“Another promising approach is to conduct on-site inspections regularly—say, every six months— to make sure chips are not just installed to fool inspectors and then removed and smuggled, as has happened before. However, conducting on-site inspections at this scale would likely require additional budget for export enforcement, or for US authorities to implement a third-party export auditor program.”
Same needs to go for LLMs, leverage them for productivity both make sure you are owning their output.
That said, I have a lot of situations at work where I am asked to simplify something I am an expert in or convey something for a different audience, particularly as I prepare for presentations and I do find it helpful in helping me step down my writing or work. YMMV.
I also skip words a lot and can miss that even after 2-3 editorial passes and it's very helpful at that vs. say normal spellcheck.
-- EDIT --
The thing I am mostly scared of is that people who create decisions will start writing them with AI and world will stop making sense in 3... 2... 1...
In the second case, these sentences were never a "voice in your head", and you never invested any effort into shaping them, so it feels more like janitorial work / drive-by nitpicking than craftsmanship.
I think one of the underappreciated things about writing "substantive" work is that the work you see at the end isn't the first attempt. And I don't mean the first draft of the piece. I mean that almost always, writers iterate on the same topic many times, either with complete published pieces or abandoned drafts or even just conversations and sessions of unproductive daydreaming. It's a cliche that your best work typically also comes out fastest, but it's not because of divine inspiration, it's because you've whittled the big idea you actually care about down so much in your mind that you instinctively know exactly how to write it.
My experience has been that for people who don't work this way or don't write a lot, LLMs can give them this incredible feeling of leaping straight from inkling to "substantive" writing. And because they haven't built up those muscles or "taste", they don't immediately recognize that it's imprecise and hard to follow.
That's not to say they're not brilliant in their own right, just that they haven't spent a lot of time on this particular thing. Sort of like a very gifted programmer who doesn't have a ton of experience yet (speaking as someone who is gifted at nothing and frequently has to do things they're inexperienced at).
So I don't think the problem is that LLMs just write bad. It's that LLMs are so wonderfully powerful that they allow you to confidently leap forward to create something that is a little beyond your experience.
And that's why I have different reactions to heavily AI generated writing. When it feels like marketing at scale, it grosses me out. But when I feel like it's just someone who is excited to write an idea and maybe doesn't have a lot of experience doing it, I'm not judgemental. My hope is that it makes them more excited about writing, and that trying to make their next piece better will lead them inevitably to start thinking about where the last piece fell short. And if there's some slop along the way, eh, I'm not compelled to read it.
Getting comments on specific sentences/passages and being able to debate them has been great for my thinking, and I've learned a lot about what works and what doesn't work while writing with it.
> blog post, a research report, a memo, a thoughtful email, a novel
Normal people don't write blog posts, research reports (because normal people don't do research), or novel. They also don't write "thoughtful emails"; they just write low effort normal emails.
Good or bad please let me know what you think of this format!
1. Agree writing process is essential, but just getting your thoughts down is the start.
2. Human writing is very often vague and wrong in hard to notice ways, but I would say it's more likely to be wrong in easy to notice ways, which is ... better?
3. I think it's rude to not use the best tools to convey the message most clearly and most respectfully of my time. I often use AI to help me write emails and nearly every time it provides more concise well structured versions of what I have to say. 99% of the time I'm just trying to convey a message clearly, that's all. When I write my draft there is often words I can delete or phrases that I can shorten. It's hard to spot but when you have AI point them out it's obvious. This is the role of an editor (check out Stephen King book On Writing which makes this point). Unfortunately I don't have an editor, but I have AI that helps trim the fat, get to the point and save my readers time.
If these tools are available to you and you don't use them, then I think that would be disrespectful. I don't take any solace thinking how much time and effort went into what someone wrote me. If I could save them time and have AI write or edit, and save me time reading it since it would be more concise and better structured, it's a win-win.
I think “cutting down on words” depends on personal writing style. AI tends to write more than I would, so I usually have the opposite problem.
I would have much preferred if the rep had written a shorter document and reached out about things they were uncertain about. Instead thet wasted my and my team's time as we first tried to make sense of the document on good faith before realizing the problem.
I don't care if people use LLMs per se, but if this is functionally the result, then in this aspect of my work things would go much better if people did not use them. They are a great interface for talking to a machine, but due to the laziness they breed, they are a terrible interface for talking to other humans unless approached with a great deal of discipline.
I think you should write down your thoughts, tell it to make the points you spelled out and be concise and respectful of the reader's time and attention. When I do this, it works very well. But I agree if you just have it respond and reason it tends to often produces bad responses.
If you're talking about the general plan, structuring paragraphs, organizing your thoughts then I totally agree. Claude, in particular, is very hard to read.
But if English is a second language for you and you still get confused about what is the proper preposition or phrasal verb then you need AI. I suggest Grammarly.
You're probably extending the logic from the idea that a programmer-llm is a tool. That is at least a defensible position. A written program isn't written to be experienced by people (usually), rather to create certain outcomes.
On the other hand, the written word is the entire experience of the person reading your thing, and that can all be created by an AI without your input. It has it's own biases and motivations and "alignments" (so to speak). Thus it is not a tool and actually a writer itself. And when it writes, you miss out on the experience of having written it, thus actually you end up becoming the tool (pun intended) that LLM is using to express itself.
But as a thought experiment, replace “writing” in this article with “coding”, and ask yourself why the same points wouldn’t hold? Maybe it’s because I enjoyed writing essays in college, before I wrote software as a job, but I’ve never seen much difference between writing code and writing natural language. They’re both a means to an end, some lossy way of encoding thought. One is machine executable and one isn’t (arguable maybe, at this point) but I think they both suffer from the same pathologies outlined here when AI takes over, eg, “writing is thought”, “AI writing can be subtly wrong” etc.
Yet the preponderance of hot takes I see these days are along the lines of “code review is dead”, “writing code by hand is dead”. Perhaps people largely don’t see code as communication? It bothers me to imagine a world where the vast majority of textual communication produced becomes machine generated, whether that’s natural language or code.
Claude on the other hand is atrocious at sounding human, it’s like it’s deliberately following LLM -speak stereotypes no matter how you tell it to write. I finally told it to stop using “Claudeisms,” which gave better though not great results.
To take it to the extreme, a “3x worker” is by definition relative, and someone malicious could achieve that by overloading a colleague with generated text that they themselves hadn’t bothered to read in full. You can stay at your baseline but be “3x” if you’re sufficiently able to sabotage your coworkers.
I’ve been on the receiving end of needing to review “documentation” that was generated almost entirely by LLM. It wasn’t anything that was done with malicious intent, but it’s still exhausting, and you need to exercise an even greater degree of caution because of the potential subtle mistakes or inaccuracies that a real human would never make.
Oh, but I would!
https://news.ycombinator.com/item?id=49747070
I think people are stuck in the idea that the way you "use AI to write" is to let it generate alternative words for your thoughts. That's a terrible way to write. I advocate for a rule: "any word an LLM suggests to you is disqualified, even if it's better than the word you've chosen". Nobody is vigilant enough to keep LLM word selection from bleaching out their personal style.
But LLMs can do things for language (natch) that deterministic programs can't. Those things are helpful and you should consider taking advantage of them. Here's a short list of ways an LLM can potentially improve your writing:
* It can instantly spot overused words and turns of phrase, or, better still, worthless filler and throat-clearing like "just" and "very" and "it's important to note".
* It can rescue active verbs trapped inside nouns, where sentences are wrapped in zombie verbs like "make" or "reach" or "start" or "have", carbonite-frozen verbs like "decision" or "agreement" or "distortion".
* It can match the subjects of your sentences with actual characters in the action of your story or argument, flagging all the times you accidentally nail the subject of a sentence down to some random part of the scenery. It can check to see whether the new detail each sentence adds (if it adds any at all --- something else it can check) is in the stress position of the sentence.
* It can check your transitions and flag places where the openings of sentences and paragraphs are abrupt. For that matter: it can check the topic sentences of paragraphs and the flow from graf to graf.
* It can check for passive voice, but also note the (many) instances where passive is the right call for what you're trying to say.
The model never gets tired. It generally never forgets the rules. It can instantaneously diagram out a sentence and work out an accurate model of the semantics of your writing. It's doing things you cannot do with a grammar checker.
You're not going to want to act on everything an LLM flags. LLMs have an idiosyncratic sense of style (I keep saying they write every sentence like it's the headline of a magazine article). Some of these quirks of writing will be part of your voice, and you'll need to keep them.
You're not going to want the LLM to give you alternate words and sentences. That way lies Velveeta. And many times, the LLM will flag a usage issue and your response won't be to act on the suggestion, but just to rewrite the sentence or paragraph entirely --- or, better yet: just delete it, which is an awesome feeling.
This is using an LLM as a copyeditor. I've worked with professional copyeditors, and the feeling of working with an LLM copyeditor is comparable, except that the LLM is much more thorough, and I don't feel bad about ignoring it when we disagree. This style of working doesn't allow the model to infect your writing; it's just making you better informed about the words you're choosing. I think more people should try it.
You do not.
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 power they have against it (e.g. downvoting).
My 75 y/o mother wrote a kids' book about the animals on my brother's farm, and she was singing praises about how much AI was helping her. I reminded her that I would rather have her grandchildren reading _her_ words and not some clanker's words. I don't think I am alone in valuing the human soul.
The only cure: be the person people can turn to instead of the machine, whether that's helping them write or helping them become better writers. And, yeah, you'll have to do that cheaper than the AI (netted against the social cost of being the guy/gal who got caught slopping).
I do appreciate old things, antiquity, retro technology, and nostalgia. But I also think people sometimes hold onto things long after their practical purpose has passed. That’s not necessarily bad. Typewriters, handwriting, calligraphy, old forms of art and writing can still have value as crafts and forms of expression. They just don’t necessarily need to remain the primary way we communicate.
The more interesting problem to me is communication itself. We often assume that because two people speak the same language, they’re actually transferring information effectively. In reality, two people can hear the exact same words and walk away with completely different understandings because their knowledge, context, and mental models are different.
I think the future should be less about preserving old methods of expression and more about finding better ways to actually transfer meaning between people, while still allowing each person to understand that information in their own context.
The old ways will probably stick around as art, nostalgia, and tradition, just like old forms of lore and craftsmanship. That’s fine. They’re worth keeping for those reasons. But I don’t think we need to confuse preserving the art with preserving the limitations of the old method.
I use one for the second: photograph a shop receipt, get back shop, date, total and line items as structured fields. The output is checkable against the piece of paper in your hand, it costs about a third of a cent per receipt, and when it is wrong you see it immediately because the total does not match.
That is almost the opposite of writing. There is a ground truth, the human is the verifier rather than the audience, and a wrong answer is cheap to spot. I agree with the essay about prose, and I think the reason has less to do with AI than with whether anything exists to check the output against.
Every generation of abstraction promises to remove the layer below it, and every generation discovers that the layer below leaks eventually. The engineers who thrive aren't the ones who adopt fastest — they're the ones who can drop down a level when the abstraction breaks, and who remember what the old way was actually doing.
The risky position isn't 'old person who won't learn new things.' It's 'person who only knows the new things' when the new things have a bad Tuesday.
Concrete version. I submitted an essay here tonight from a fresh account. It was accepted, it sits in my submissions with a point on it, and logged out the item page is empty. The account is mine, the writing is mine, and neither fact gets it in front of a reader.
So the question is not really about prose quality. It is who answers for it when it is wrong. A named person can. A model cannot, and the platforms are pricing that in quietly rather than by rule.
The tell isn't vocabulary. It's that no sentence is addressed to anyone. Prose written "to an audience" has no concession to a specific objection in it, no place where the writer guessed wrong about you and fixed it. Name one skeptical reader before you draft and most of that tell goes away.
Which is also why the accusation two levels up is unfair. "Orbiting around a point rather than landing on it" is a sentence addressed to someone. Style detectors miss that.
But does it?
Writing certainly can be thinking. Anyone who has wrestled with a paragraph at three in the morning knows the peculiar phenomenon: you begin with an idea, and somewhere around the third comma you discover that you didn't actually have an idea. You had a mood wearing an idea costume.
Excellent. Writing caught the bug.
But why assume the biological hand that pushes words through a keyboard possesses some sacred monopoly on this process?
For thousands of years, human beings have externalized cognition. We invented counting stones, writing tablets, books, maps, diagrams, calculators, libraries, laboratories, and—eventually—computers. Every one of these inventions moved some cognitive operation outside the skull.
Nobody argues that using a calculator prevents mathematics because “the multiplication process is part of thinking.”
Sometimes it does.
Sometimes the calculator merely prevents you from wasting twelve minutes multiplying 487 by 36.
The interesting question is therefore not Who typed the sentence? but Who did the thinking?
And here the anti-AI argument becomes rather less certain.
Suppose I write an argument and give it to an AI with instructions:
Find every assumption I have smuggled in. Tell me where my argument contradicts itself. Give me the strongest argument against my position. Identify every sentence that sounds impressive while saying nothing.
Have I stopped thinking?
Or have I acquired a rather peculiar intellectual sparring partner who never gets tired, never needs coffee, and is perfectly happy to tell me that my brilliant insight is actually three clichés in a trench coat?
The answer depends entirely on what I do with its response.
If I obediently paste the machine's prose into an essay and call it my thinking, then yes, something important has been lost. But if I use the machine to attack my assumptions, generate alternatives, expose blind spots, and force me to make distinctions I had previously avoided, then the machine is not replacing the thinking process. It is perturbing it.
And perturbation is where interesting things happen.
Every nervous system operates through feedback. We don't think in isolation; we think through interactions with environments, symbols, other people, books, arguments, mistakes, and increasingly, machines.
The human mind has never been a sealed container.
The more interesting danger, therefore, isn't that AI will “write for us.” The danger is that we will stop noticing when we have handed over our judgment along with the prose.
That distinction matters enormously.
A calculator does not tell you what you should calculate. A spellchecker does not decide what you believe. A search engine does not automatically determine whether the evidence supports your conclusion. And an AI should not be granted the authority to do any of those things merely because it can produce a remarkably confident paragraph about them.
But neither should we fetishize the struggle of producing every sentence manually.
There is a strange technological superstition hiding here: the belief that cognitive labor becomes more intellectually authentic as it becomes more tedious.
I don't buy it.
The question isn't whether the machine touched the sentence.
The question is whether you touched the idea.
If an AI writes a paragraph that you blindly accept, you have delegated your thinking.
If an AI writes a paragraph that makes you say, “Wait, that's bullshit,” and forces you to figure out exactly why it is bullshit, the machine has just helped you think.
And perhaps that is the real reality tunnel worth examining.
We have spent centuries confusing the tools of thought with thought itself.
The printing press didn't kill thinking. The calculator didn't kill mathematics. The word processor didn't kill writing.
Perhaps the genuinely dangerous technology is not the one that writes sentences.
Perhaps it is the one that convinces us we no longer need to ask what the sentences mean.
I think people massively over-rely on AI and I really worry about the consequences of this. But there is nothing baffling at all about the basic appeal, IMO.
Anyway, I completely agree. Writing isn't something to be delegated.
The problem is not new with LLMs. Businesses have been delegating writing to idiots who don't really care for the entire history of business. Authors and musicians have traded their souls for inauthentic crowd-pleasing results for about as long.
No, the problem with LLMs is that people are all using the same generic models. This only further shows that the future of LLMs is locally trained and locally run. Personal LLMs on a personal computer. Business LLMs on-prem. We will continue see them flourish in ways that generate absolutely no money whatsoever on their own, but do add marginal value when combined with a heavy dose of human creativity.
All the clueless old farts that are still alive by this point will continue to lecture the rest of us with more condescending gee whiz "whaddyaknow" and "whodathunkit" nonsense. As if they've even had their finger on the pulse since the early 2000s.
I'm genuinely not sure what you're implying here. Can you expand?