This is something I find fascinating about TikTok: on that platform you literally get a few seconds to catch the attention of your audience before they skip to the next video.
You can't just find one hook that works and reuse it forever because people will get bored of it - including if that hook is heavily used by other accounts.
This makes TikTok a fascinating brute-force attack on human psychology, with literally millions of people all trying to find the right hooks to catch attention and constantly evolving and iterating on them as the previous hooks stop being effective.
I wrote my story and titled it, "My experience at work with an automated HR system". I sent it to a few friends, only a couple of them read it.
A week later, I renamed it to "The Machine Fired Me". That seemed to capture it better. The goal wasn't to make it click bait, but it was to put the spoiler, and punch line right up front. It blew up!
I had just read Life of Pi, and one thing I like about that book is that you know the punch line before you even pick up a copy. A boy is stuck with a bengal tiger in a boat. Now that the punch line is out of the way, the story has time to unfold and be interesting in its own merit. That's what I was trying to recreate with my own story.
There is content you write for acquisition and content you write for retention and my #1 tip for writers who want to engineer growing an audience is be clear before you sit down to write a piece which it’s going to be.
Content for acquisition, the reader’s relationship is to the topic, they have to be convinced the topic is relevant to their life goals but it’s valuable despite who the topic.
Content for retention, the readers relationship is to you as the writer and the topic is merely there as a MacGuffin to help illuminate some aspect of you that is unique.
Business Insider had this down to a science over a decade ago. They started a series called “So Expensive”, detailing why various things were expensive, the first 4 videos in the series were: Caviar, Saffron, Rolexes and Horseshoe Crab Blood. Statistically, some tens of millions of people have organically had the thought of why the first 3 were expensive but zero people have even wondered why horseshoe crab blood was expensive. The 4th video was a way to test, of all the people who were willing to click on the first three, how many were willing to follow along to the 4th because of a trust in BI? The next 4 in the series was Vanilla, Silk, Louboutins & Scorpion Venom.
Creating all content for acquisition is both too exhausting and also sub optimal for the reader because they want deeper stuff to follow as well. I suggest up to 1 in 3 acquisition articles if you can manage when starting out but then ramping down to no more than 1 in 10 fairly quickly or you burn out.
Okay, because no one seems to be answering the Venice question:
- They had a strong navy (and shipbuilding capacity), making a blockade difficult
- They traded with many nations, so no one group could cut off their food supply
- Fish
- They had a near monopoly on the trade of salt and spices, the former of which was important to everyone and the latter of which was important to aristocrats
(note: I read a few sources but this is not thorough research)
You know I’ve never read an article by Gwern that made me feel like he was sensitive to this idea, one that in my head essentially breaks down to the use of narrative and the leverage of “stakes” that inform the reader of kinds of conflict that make a narrative special.
I’m reminded of a remark made by David Foster Wallace (on KCRW? Or oft-repeated elsewhere) about how he had to come to terms with the purpose of writing not being to show off how smart you are to the reader. Instead your writing has to evince some kind of innate investment to the reader that piques their genuine interests and intrigue.
A lot of writers are tainted by the expectations set in grade school. Write for a grade and good writing is what yields a good grade according to the standards set by the subject which often is not ‘Composition’ but more like ‘Prove to me that you remember everything we mentioned in class about the French Revolution’.
I’ve never felt drawn into an article by Gwern at least not in the way that I have been by some writing by Maciej Cegłowski, for example. Reading Gwern I am both overwhelmed by the adornments to the text (hyperlinks, pop-ups, margin notes; other hypertext doodads and portals) and underwhelmed by the substance of the text itself. I don’t consider Paul Graham a literary griot either. But I find that his own prose is bolstered by a kind of clarity and asceticism that is informative and not entirely void of good style and form.
Lawrence McEnery of the University of the Chicago contributed a lot of good thinking to this kind of stuff though.
This wasn’t meant to be a criticism of the author of this post’s own work. But here that’s how it’s left. I haven’t come across any writing of his that’s as intriguing as "Empires Without Farms: The Case of Venice” seems. If anyone has any recommendations, do share.
> When writing, first, make the reader care, one way or another. Because if I am not hooked by the first screen, I will probably not keep reading—no matter how good the rest of it is!
Keeping the reader glued to the screen is not the primary goal of writing. This artificial goal pollutes the connection between writer and reader. It makes them buyer and seller and rewards sales tactics. You don't write for the reader. You write for yourself first. Readers sometimes, just happen to appreciate it about as much you do.
I think "Just… start with the interesting part first" is quite different, and actually much better advice than "make me care". I'm more than done with stupid hooks and attention grabbing techniques, just plainly and honestly state at the outset what the point is of what will follow.
This reminds me of what my PhD supervisor told me as he was trashing my first draft of my first paper: “up until this point in your life you’ve been trained to convince someone who knows more than you that you know something by writing impressive equations and complex concepts. Now you are the expert and if you do that nobody will read your papers. And if someone stops after a few sentences you’ll lose citations too.”
There's a spectrum of writing, corresponding to supply/demand or push/pull. The article is giving advice for oversupplied writing, where the audience doesn't really want to hear you, and you're trying to badger it into reading it anyway - typically, for some sort of personal gain (getting an interview, making a sale, promoting a political cause). Yes, attention hacking is important in this case.
There is also a writing where people are looking for the information, and they are showing up at your door because they already care. Presumably you wrote, because you saw the open question, and want to try answering it. History books, encyclopedias, classic literature by dead people, falls under this. Ironically, so does the example of Venice - you would read about Venice if you were already curious; there is little profit in "making someone care" about Venice otherwise. An attention grabby style would be forced and counterproductive in this.
I have often thought that all good fiction is mystery. This is obviously an overstatement, but I think it’s not too far off. Humans are mystery solvers. If I don’t have a compelling mystery to solve—something like the “what’s going on beneath the surface in this town?” that David Lynch does so well—when I’m reading your book or watching your tv show or playing your game, I’m usually out unless I have a strong prior interest (which simply means that I brought my own mystery).
"First, make me care" is exactly right. But I also know that anytime you have narrative non-fiction on here, someone without fail argues that the author didn't get straight into the details.
Care is the most important trait of people who make great things; it's not money or time. Is not even skill.
I was interviewing a candidate yesterday and I noticed that a project inside their personal website was not working. I told him my opinion on care and he said that he hasn't had the time to deploy it, since he's been working on it for 2 weeks already and it was working on his local machine.
A few hours after the interview, the project was online.
The bitter pill of realizing the importance of care is that this applies not just to literary works, like Gwern's case, but it also applies to any creative endeavor: writing, music, drawing, and yes, software engineering.
That CLI tool without a tutorial. That product with a confusing sign-up flow. The purchase without a confirmation dialog such that I don't feel I was just scammed.
It's all the same. Lack of care.
I've also noticed that when caring is there, skills follow.
Know your audience is right, and Gwern misses the most interesting thing about Venice- that it was a merchantile Republic with reasonable independance from the Catholic Church. Lots of the political ideas which influenced British and American democracy came from the Italian city states. Ruskin's Stones of Venice and Bowsma's "Venice and the defense of republican liberty" capture this well, as do parts of Quentin Skinner's "The Foundations of Modern Political Thought."
“Venice built a maritime empire from a city that couldn’t feed itself; so who fed it—and why didn’t its enemies simply starve it out?”
I love ancient history and would read a good book about the Venetian empire, but the sentence answers the final question. Venice was a maritime empire (it's capital on an island), that's why its enemies could not starve it out. All in on finding out who fed it.
> We could easily have a 2026 LLM deliver high-quality editing advice to fix this up extensively, but it would still be mediocre.
This seems to suggest that a human needs to be in the creative loop, but that could be short sighted. LLM training has humans in the loop which optimize for not being bored. That's a reason for LLM texts typically being recognizeable because at the current stage they are a little too simplistically flashy. But give it a few iterations and the machine will excel humans at catching their attention, i.e., making them care.
As others wrote: Tiktok is already excellent at that. While the content in there still has humans in the loop, the choice what to show to whom and when is already entirely mechanical.
People our so tired of sensational intros and baiting questions which bury the actual lede up to the point where you discover it requires an annual subscription to find out the actual answer, that now it's actually counterproductive to start with an interesting "question".
It's facts first or gtfo. Prove to me that I'm not going to waste my time until you deliver what you promised, by delivering enough of that relevant background up front, otherwise I don't have time for your shenanigans.
I found it really ironic that the author started with something attention-grabbing about a topic completely unrelated to his point, but then made a series of such mundane statements to begin his own writing that I didn't care enough to go past his first screen. He didn't make me care about why I should make people care.
What I find extremely off-putting and overused is the pattern of making you care about an article by saying something about the person being interviewed, usually related to the interview itself. Think "he was a balding man[...] drinking his matcha latte[...]" It's always something which has zero bearing on the situation in question.
Whenever I see this, I immediately turn to cmd-a + cmd-c + `pbpaste | llm 'summarize this'`
I felt like the movie Marty Supreme completely failed to make me care about the main character until the final act where the filmmakers had to pull out all the big easy stops to force me to care about him. A third of the way in to the movie I was wondering if it was going to be explained at any point why I should be interested in this guy or care about his difficult and fairly unremarkable personality. A lot of the time it seems as if creatives assume that if you’re watching/reading/engaging with their movie/book/artwork then you already care enough to care.
Suppose you fed this article into an LLM, along with whatever other documents you had, and asked it to come up with some good candidates for opening sentences? And picked one, and let it take it from there?
I assume you'd get a mess, but it might be an interesting mess.
So LLMs can't do that? Every LLM-written historical or pseoudo-historical (i.e. made up) thing that comes up in my Facebook feed does start with a "hook" like that. Doesn't make them great articles but obviously you can prompt them to do it.
This insight is what caused the rise of the clickbait headline and its predecessors in eras past. You need a hook or there's no point reading the tale.
Adrian Wooldridge (the Economist) in "Aristocracy of Talent: How Meritocracy Made the Modern World" argues, rather successfully IMO, that what made Venice the maritime super-power was meritocracy. Indeed, he argues, that the fall of the Venitian empire came swiftly when the Doge was forced to place only Venitians (birthright) to top positions, instead of the most "capable". Hence the available talent pool shrunk.
ps. this book came out as a response to Michael Sandell's "The Tyranny of Merit: What's Become of the Common Good?" which was a best seller at the time.
For the last 30 years I've decided that the best stuff (most engaging books, stories, experiences in life) require investment. It gets worse (you go through some pain while exercising) before it gets better.
That's essentially a definition of a good life to me - finding the things worth sacrificing resources and getting the payoff.
So 'first make me care' to me is a manifest of Gen Z - tiktok - brainrot approach.
From my perspective you miss most of the really good stuff by cultivating this approach.
I.e. my favorite books - Tai Pan, Noble House; tv series - Better Call Saul! - require you to go through so much of initial boredom.
It's also the same discussion as 'learn to code vs only do AI Slop' or 'learn math and algos vs only import functions from libs and never check what's inside'.
*Exceptions apply, ofc. There are things that hook you and progressively ad depth, but it's really rare. I.e. Arcane tv show is both easy to access and quite deep.
Edit:
...so I can imagine math teacher that first tell you what are some amazing uses of derivatives and integrals - PIDs, SGD, better estimation, wave functions, generalized description of problems, accessing interesting physics etc. And after that they make you grind. I think it would be quite great. But it is so rare, that you have to make a leap of faith and assume most of the good stuff is boring initially.
The hook was great, but article was mediocre. I glazed over at the mention of LLMs in the second paragraph, skimming the article through to the end didn't improve things.
If your readers now care, don't disappoint them...
That’s basically what tv shows do every time. The pilot is great/awesome then the 10 episodes are boring and the very last one get exciting enough for you to wait the second season.
I'm beginning to think that origin stories are an underrated way to find these angles. Like why exactly did you start thinking about this topic. I guess the recipe bloggers were on to this with their long rambles about where they first tried this dish (albeit it may have been for SEO too...)
A good writer should be able to write a catchy hook.
A wise person should be able to read text that is flavored like cardboard. In general, one thing I dislike about this moment, is the incredible emphasis we place on the first five seconds of everything because everyone is thinking about all the other things we could be doing.
But many things are great because of how they feel 20 years later. The first five seconds of playing a musical instrument is horrible, but 20 years later and it is sublime. Emacs, Vim, are both notoriously forbidding, and yet, they are wonderful tools. Some things can be massaged to meet both criteria (I can imagine some emacs configuration that made it less painful and surfaced its true power in the first five seconds maybe), but other things are hard by nature and derive their value from how we have to adapt to them instead of how well they are adapted to us. I feel like the AI era is going to just accelerate this trend where everything we interact with is a slick surface and many people will never experience depth.
Perhaps I am too much of a curmudgeon, but the example first sentence made me not care at all - not about Venice, but about the writer's approach, which seems to want to conjure breathless mystery about something I could easily look up on Wikipedia (or read in tl;dr comments in this thread).
It ISN'T Venice you need to make me care about, it's YOU! Why should I spend any of my time on you?
A good first sentence should make me care about your perspective, at least for non-fiction about subjects well-studied.
Fiction, obvs, differs. Scalzi's Old Man's War had such a great first sentence I devoured the series.
Expanding Gwern's well made point, just a little, is that writing of any length needs a thesis statement in the first paragraph or so.
Thesis statements are not a new techniques, and these days they are needed much more because there is so much to read. Many articles don't state their thesis at all or not for a long time.
I don't have time to read that far to find out if it's worthwhile to me. Unless you are Satoshi Nakamoto, I'm not going to read far to find out.
I hate this. Basically it’s saying use constant clickbait to keep your reader reading.
Increasingly, readers don’t have time for this shit. Be direct, and if the reader doesn’t care, they were never meant to be your reader anyway. Someone will care, write for them.
First, make me care
(gwern.net)788 points by andsoitis 25 January 2026 | 238 comments
Comments
You can't just find one hook that works and reuse it forever because people will get bored of it - including if that hook is heavily used by other accounts.
This makes TikTok a fascinating brute-force attack on human psychology, with literally millions of people all trying to find the right hooks to catch attention and constantly evolving and iterating on them as the previous hooks stop being effective.
A week later, I renamed it to "The Machine Fired Me". That seemed to capture it better. The goal wasn't to make it click bait, but it was to put the spoiler, and punch line right up front. It blew up!
I had just read Life of Pi, and one thing I like about that book is that you know the punch line before you even pick up a copy. A boy is stuck with a bengal tiger in a boat. Now that the punch line is out of the way, the story has time to unfold and be interesting in its own merit. That's what I was trying to recreate with my own story.
Content for acquisition, the reader’s relationship is to the topic, they have to be convinced the topic is relevant to their life goals but it’s valuable despite who the topic.
Content for retention, the readers relationship is to you as the writer and the topic is merely there as a MacGuffin to help illuminate some aspect of you that is unique.
Business Insider had this down to a science over a decade ago. They started a series called “So Expensive”, detailing why various things were expensive, the first 4 videos in the series were: Caviar, Saffron, Rolexes and Horseshoe Crab Blood. Statistically, some tens of millions of people have organically had the thought of why the first 3 were expensive but zero people have even wondered why horseshoe crab blood was expensive. The 4th video was a way to test, of all the people who were willing to click on the first three, how many were willing to follow along to the 4th because of a trust in BI? The next 4 in the series was Vanilla, Silk, Louboutins & Scorpion Venom.
Creating all content for acquisition is both too exhausting and also sub optimal for the reader because they want deeper stuff to follow as well. I suggest up to 1 in 3 acquisition articles if you can manage when starting out but then ramping down to no more than 1 in 10 fairly quickly or you burn out.
- They had a strong navy (and shipbuilding capacity), making a blockade difficult
- They traded with many nations, so no one group could cut off their food supply
- Fish
- They had a near monopoly on the trade of salt and spices, the former of which was important to everyone and the latter of which was important to aristocrats
(note: I read a few sources but this is not thorough research)
I’m reminded of a remark made by David Foster Wallace (on KCRW? Or oft-repeated elsewhere) about how he had to come to terms with the purpose of writing not being to show off how smart you are to the reader. Instead your writing has to evince some kind of innate investment to the reader that piques their genuine interests and intrigue.
A lot of writers are tainted by the expectations set in grade school. Write for a grade and good writing is what yields a good grade according to the standards set by the subject which often is not ‘Composition’ but more like ‘Prove to me that you remember everything we mentioned in class about the French Revolution’.
I’ve never felt drawn into an article by Gwern at least not in the way that I have been by some writing by Maciej Cegłowski, for example. Reading Gwern I am both overwhelmed by the adornments to the text (hyperlinks, pop-ups, margin notes; other hypertext doodads and portals) and underwhelmed by the substance of the text itself. I don’t consider Paul Graham a literary griot either. But I find that his own prose is bolstered by a kind of clarity and asceticism that is informative and not entirely void of good style and form.
Lawrence McEnery of the University of the Chicago contributed a lot of good thinking to this kind of stuff though.
This wasn’t meant to be a criticism of the author of this post’s own work. But here that’s how it’s left. I haven’t come across any writing of his that’s as intriguing as "Empires Without Farms: The Case of Venice” seems. If anyone has any recommendations, do share.
Keeping the reader glued to the screen is not the primary goal of writing. This artificial goal pollutes the connection between writer and reader. It makes them buyer and seller and rewards sales tactics. You don't write for the reader. You write for yourself first. Readers sometimes, just happen to appreciate it about as much you do.
There is also a writing where people are looking for the information, and they are showing up at your door because they already care. Presumably you wrote, because you saw the open question, and want to try answering it. History books, encyclopedias, classic literature by dead people, falls under this. Ironically, so does the example of Venice - you would read about Venice if you were already curious; there is little profit in "making someone care" about Venice otherwise. An attention grabby style would be forced and counterproductive in this.
"You won't believe the weird trick that the city of Venice did to feed itself"
I was interviewing a candidate yesterday and I noticed that a project inside their personal website was not working. I told him my opinion on care and he said that he hasn't had the time to deploy it, since he's been working on it for 2 weeks already and it was working on his local machine.
A few hours after the interview, the project was online.
The bitter pill of realizing the importance of care is that this applies not just to literary works, like Gwern's case, but it also applies to any creative endeavor: writing, music, drawing, and yes, software engineering.
That CLI tool without a tutorial. That product with a confusing sign-up flow. The purchase without a confirmation dialog such that I don't feel I was just scammed.
It's all the same. Lack of care.
I've also noticed that when caring is there, skills follow.
And what would happen if everyone followed it? Clickbait titles like "the third one will surprise you" and TikTok.
I love ancient history and would read a good book about the Venetian empire, but the sentence answers the final question. Venice was a maritime empire (it's capital on an island), that's why its enemies could not starve it out. All in on finding out who fed it.
I came away not having a resolution to the hook - violating the articles second principle.
This seems to suggest that a human needs to be in the creative loop, but that could be short sighted. LLM training has humans in the loop which optimize for not being bored. That's a reason for LLM texts typically being recognizeable because at the current stage they are a little too simplistically flashy. But give it a few iterations and the machine will excel humans at catching their attention, i.e., making them care.
As others wrote: Tiktok is already excellent at that. While the content in there still has humans in the loop, the choice what to show to whom and when is already entirely mechanical.
People our so tired of sensational intros and baiting questions which bury the actual lede up to the point where you discover it requires an annual subscription to find out the actual answer, that now it's actually counterproductive to start with an interesting "question".
It's facts first or gtfo. Prove to me that I'm not going to waste my time until you deliver what you promised, by delivering enough of that relevant background up front, otherwise I don't have time for your shenanigans.
Whenever I see this, I immediately turn to cmd-a + cmd-c + `pbpaste | llm 'summarize this'`
I assume you'd get a mess, but it might be an interesting mess.
The book makes for a fine read IMO: https://www.amazon.com/Aristocracy-Talent-Meritocracy-Modern...
ps. this book came out as a response to Michael Sandell's "The Tyranny of Merit: What's Become of the Common Good?" which was a best seller at the time.
So 'first make me care' to me is a manifest of Gen Z - tiktok - brainrot approach. From my perspective you miss most of the really good stuff by cultivating this approach. I.e. my favorite books - Tai Pan, Noble House; tv series - Better Call Saul! - require you to go through so much of initial boredom. It's also the same discussion as 'learn to code vs only do AI Slop' or 'learn math and algos vs only import functions from libs and never check what's inside'.
*Exceptions apply, ofc. There are things that hook you and progressively ad depth, but it's really rare. I.e. Arcane tv show is both easy to access and quite deep.
Edit: ...so I can imagine math teacher that first tell you what are some amazing uses of derivatives and integrals - PIDs, SGD, better estimation, wave functions, generalized description of problems, accessing interesting physics etc. And after that they make you grind. I think it would be quite great. But it is so rare, that you have to make a leap of faith and assume most of the good stuff is boring initially.
If your readers now care, don't disappoint them...
Best example: the walking dead season tv shows.
1. First, make me care.
2. Then provide an indication (e.g. in the video description) giving the time in the video where the question starts to be answered.
If you make me somewhat care, but I have to binary search through your video to skip the rambling, I'm likely to back button out.
It really depends on who the audience is...
If I were reading a book and each chapter started with such a "hook," it'd start to feel like a LinkedIn post.
Chapter 1: I didn't know what it felt like to be alive until I was dead...
Chapter 2: Death was nothing compared to what came next: judgment.
Chapter 3: I thought I knew what judgment was until...
A wise person should be able to read text that is flavored like cardboard. In general, one thing I dislike about this moment, is the incredible emphasis we place on the first five seconds of everything because everyone is thinking about all the other things we could be doing.
But many things are great because of how they feel 20 years later. The first five seconds of playing a musical instrument is horrible, but 20 years later and it is sublime. Emacs, Vim, are both notoriously forbidding, and yet, they are wonderful tools. Some things can be massaged to meet both criteria (I can imagine some emacs configuration that made it less painful and surfaced its true power in the first five seconds maybe), but other things are hard by nature and derive their value from how we have to adapt to them instead of how well they are adapted to us. I feel like the AI era is going to just accelerate this trend where everything we interact with is a slick surface and many people will never experience depth.
Read boring shit.
1)
Using brave on iPhone.
Firefox and Safari works…
It ISN'T Venice you need to make me care about, it's YOU! Why should I spend any of my time on you?
A good first sentence should make me care about your perspective, at least for non-fiction about subjects well-studied.
Fiction, obvs, differs. Scalzi's Old Man's War had such a great first sentence I devoured the series.
Thesis statements are not a new techniques, and these days they are needed much more because there is so much to read. Many articles don't state their thesis at all or not for a long time.
I don't have time to read that far to find out if it's worthwhile to me. Unless you are Satoshi Nakamoto, I'm not going to read far to find out.
Always found them interesting.
prepending a one-liner-about-some-feat that might interest that particular company, before the usual cv afterthat?
hmm. made me think..
Increasingly, readers don’t have time for this shit. Be direct, and if the reader doesn’t care, they were never meant to be your reader anyway. Someone will care, write for them.