I'm far more cynical. I had a blog many years ago, in the first wave of blogging. In retrospect, I consider it mostly to have been a mistake, for various reasons. The biggest one nowadays is what I call the "negative lottery". Except in a few circumstances, an ordinary person doesn't have much to gain, but a huge amount to lose if, e.g. a social-media mob comes after you. You're also doing what I call "fighting the power law", in that a tiny tiny number of writers will be very successful, and everyone else is going to be saying how happy they are to have a few readers (or not, like me). And, speaking for myself, I'm extremely unhappy with the system where the powerful up "above" are free to viciously lie and even outright libel anyone "below" them, and their targets can't even effectively fight back.
I view it all like the genre of popular business books - "Have a clear vision, execute your plan, work hard - and have the backing of rich family members".
I could go on, but I'm going to take my own advice and stop here.
I've been writing a blog since 2007 and a substack for about two months. I've also started writing for the local newspaper for the last six months. That is the most rewarding thing because people come up to me on the street or in the grocery store to tell me how they've enjoyed my writing. I've got about twenty subscribers so far on substack, and I don't anticipate getting a huge number in the future. Maybe it's my age - I ain't a young'n anymore. So far I've refused to use AI, except for googling things, if that can be called AI. I've developed my writing style on my own, over many years, and I distrust artificial help. It seems to me that it would lead to a dependency, and besides, I trust my instincts - I am a voluminous reader. If, in ten years, I'm still writing but everyone else is using ChatGPT, at the very least, I will stand out for not using it. Mind you, I wouldn't mind having more subscribers. My substack is crjustice5.substack.com in case anybody is wondering.
By the way, point of information, when you discuss "Scott Alexander ... on the importance of courage.", are you familiar with the extensive harassment campaign against him? As you note, it's not the same as literal war - but I'd rebut that's a low bar. People have a right to be worried about attempts to seriously harm their lives, even if it's short of death. That's not the same as random ranters who can be ignored. Part of what I'm referring to is related in this article:
It's very long, so readers might want to just go to the section "Scott Alexander".
"In February 2021, after Scott rearranged his life and quit his job in order to minimize the disruption from his name being revealed, ..."
That's pretty heavy consequences for just writing a blog. And the people going after him aren't even the worst examples of the type (I feel like I'm in a fantasy story discussing levels of monsters - "That's just a gibbergoth, it's certainly dangerous, but nowhere near as problematic as a wokewarrior").
Before someone says that Scott triumphed, so all's well that ends well, let me preemptively reply that's a survivorship bias outcome. If the attacks on him had succeeded, it'd be a different story. Plus Scott has some very powerful supporters.
Really, it strikes me as a very poor case to assert, paraphrased, "Significant attempts to hurt you are statistically rare, most likely nobody will care about what you say at all" (how is that supposed to be a recommendation?).
LLMs can't "fact check" a claim. And Deep Research hallucinates and is very likely to leave out key sources (or include garbage ones) due to the vagaries of what it happens to have access to. When people who really know a subject area assess its output, it often looks pretty shabby.
it's a bit ironic that LLMs or modern AI is best at creative and artistic expression (things many people thought would be very difficult to impossible to mimic this well; things some people considered uniquely human domains) rather than cold logic, math and facts which we expected it to handle with ease. scifi and i guess our general experience with computers gave us the wrong idea of what to expect.
Second the point about AI. I've been writing a lot of book reviews, and when I read a (work!) book I tend to end up with about 20 pages of messy notes--summaries, criticisms, idle thoughts, etc. In the past I basically just forgot about them because it was too much work to read through the whole thing. I've been asking Claude to try and help me make sense of all these notes--identify the main objections I seem to have, suggest ways of summarising the main claims of the book, etc. It makes everything go about 10x as fast. Also useful for more fine-grained writing help: fix this sentence which I know isn't quite grammatical, suggest an example to use here, and so on.
One thing I've found it is weirdly bad at: searching a text to find good quotes to illustrate a point you want to make. It doesn't even seem to be good at looking through my notes and finding good quotes I've purposefully written down.
One thing about advice is it works differently coming interpersonally from a single person like a friend, than from a public writer offering advice to an entire audience. And when you receive advice from any source you may hear it not only as a recipient, but as a potential sharer already anticipating wanting to pass on this advice to others. The benefits are not only social; there is inherent satisfaction in offering your own chosen wisdom to somebody. It can even be an affirmation of personal values.
I liked your own advice, aside from the LLM one - but do you really think people like Nate Silver, Matt Yglesias and Rob Henderson are appropriate models for the vast majority of beginning bloggers? (I know nothing about Ed West). These guys are basically celebrities; the first two honed a distinctive brand over the last decade or two and not only from blogging, while Henderson has a very unique and compelling story and first came to attention through a popular book. Even someone like yourself, who started small, was at least an established academic who could use that network to help spread the word. I'm much more inclined to look at people who are just random schmoes but have gotten up to a few thousand. Or if more popular, someone like Erik Hoel or Henrik Karlsson, or Sam Kahn, who devote many posts to reflecting on the art and culture of blogging.
Hmm good points. On the second one, to be honest I didn’t give much thought to who I listed after “e.g”. - I just linked some big names that came to mind when I was writing. I think your observation is probably correct. What I wanted to convey there is just that there are many more successful bloggers than me who have given good advice. If I remember correctly their advice is pretty widely applicable, but it’s true that for many aspiring bloggers others (including those you mention) would give more relevant advice.
About the advice in #12: it could be interesting, as a future post, to walk readers step-by-step through your entire process of using an LLM for some selected piece, to show exactly what this looked like under the hood. Not to mention a refreshing change-of-pace from the usual.
Re. the inherent satisfaction of offering advice: in general I think you don't pay enough attention to pleasure (as in, pleasure for its own sake), as a powerful motivator for behavior along with any evolutionary or social functions.
i don't think it's puzzling. giving "free" advice shows others that you are a useful and valuable member of the community. they are more likely to help you in return, ally with you, follow you, see you as a leader (higher status), and generally want to keep you around. i can see it being adaptive.
i've always wanted to write/start a blog myself even though i think nobody will read it haha. writing about stuff helps you learn and clean up your thinking too, so even if nobody reads it, i'd improve from it. readers would be a bonus.
my problem is the opportunity cost. i just don't feel like i have the time. i also tend to get hyperfocused and i know i won't be able to balance my blog with work and personal life.
i'll probably never get around to it but it's okay.
i'm glad for your blog/substack though; it's one of my favorites.
My partner tells me she would cry for days if she received some of the angry and hostile comments I’ve received throughout my blogging career.
Dan i read 30-40 substackers extensively and you are the most politically correct,moderate and normal looking face out of the bunch,to the point i am surprised i enjoy your stuff so much. What can they even hurl at you. And substack is less toxic that other platforms
Ha - lots of people really hate politically correct, moderate, and normal content. But yes, I'm sure I don't get it anywhere near as badly as most actually controversial writers.
"Another fear some writers have is that people will read their blog and hate it and so hate them, and then maybe try to cancel them. People are typically far too paranoid about this. Nevertheless, if you’re terrified of getting negative reactions from what you write (i.e., you’re a coward), blogging is probably not for you, at least when it comes to non-fiction, opinion blogging."
...
"However, subject to those constraints, the most interesting writers are interesting in part because they take risks. "
...
"In general, quite a bit of self-promotion is necessary to reach and grow an audience."
I'm far more cynical. I had a blog many years ago, in the first wave of blogging. In retrospect, I consider it mostly to have been a mistake, for various reasons. The biggest one nowadays is what I call the "negative lottery". Except in a few circumstances, an ordinary person doesn't have much to gain, but a huge amount to lose if, e.g. a social-media mob comes after you. You're also doing what I call "fighting the power law", in that a tiny tiny number of writers will be very successful, and everyone else is going to be saying how happy they are to have a few readers (or not, like me). And, speaking for myself, I'm extremely unhappy with the system where the powerful up "above" are free to viciously lie and even outright libel anyone "below" them, and their targets can't even effectively fight back.
I view it all like the genre of popular business books - "Have a clear vision, execute your plan, work hard - and have the backing of rich family members".
I could go on, but I'm going to take my own advice and stop here.
I've been writing a blog since 2007 and a substack for about two months. I've also started writing for the local newspaper for the last six months. That is the most rewarding thing because people come up to me on the street or in the grocery store to tell me how they've enjoyed my writing. I've got about twenty subscribers so far on substack, and I don't anticipate getting a huge number in the future. Maybe it's my age - I ain't a young'n anymore. So far I've refused to use AI, except for googling things, if that can be called AI. I've developed my writing style on my own, over many years, and I distrust artificial help. It seems to me that it would lead to a dependency, and besides, I trust my instincts - I am a voluminous reader. If, in ten years, I'm still writing but everyone else is using ChatGPT, at the very least, I will stand out for not using it. Mind you, I wouldn't mind having more subscribers. My substack is crjustice5.substack.com in case anybody is wondering.
By the way, point of information, when you discuss "Scott Alexander ... on the importance of courage.", are you familiar with the extensive harassment campaign against him? As you note, it's not the same as literal war - but I'd rebut that's a low bar. People have a right to be worried about attempts to seriously harm their lives, even if it's short of death. That's not the same as random ranters who can be ignored. Part of what I'm referring to is related in this article:
https://www.tracingwoodgrains.com/p/reliable-sources-how-wikipedia-admin
It's very long, so readers might want to just go to the section "Scott Alexander".
"In February 2021, after Scott rearranged his life and quit his job in order to minimize the disruption from his name being revealed, ..."
That's pretty heavy consequences for just writing a blog. And the people going after him aren't even the worst examples of the type (I feel like I'm in a fantasy story discussing levels of monsters - "That's just a gibbergoth, it's certainly dangerous, but nowhere near as problematic as a wokewarrior").
Before someone says that Scott triumphed, so all's well that ends well, let me preemptively reply that's a survivorship bias outcome. If the attacks on him had succeeded, it'd be a different story. Plus Scott has some very powerful supporters.
Really, it strikes me as a very poor case to assert, paraphrased, "Significant attempts to hurt you are statistically rare, most likely nobody will care about what you say at all" (how is that supposed to be a recommendation?).
LLMs can't "fact check" a claim. And Deep Research hallucinates and is very likely to leave out key sources (or include garbage ones) due to the vagaries of what it happens to have access to. When people who really know a subject area assess its output, it often looks pretty shabby.
it's a bit ironic that LLMs or modern AI is best at creative and artistic expression (things many people thought would be very difficult to impossible to mimic this well; things some people considered uniquely human domains) rather than cold logic, math and facts which we expected it to handle with ease. scifi and i guess our general experience with computers gave us the wrong idea of what to expect.
“To be clear, I’m not one of them. I recently had to remove an attempted joke from an article because it was so unfunny”
This is actually a funny comment all on its own..
Second the point about AI. I've been writing a lot of book reviews, and when I read a (work!) book I tend to end up with about 20 pages of messy notes--summaries, criticisms, idle thoughts, etc. In the past I basically just forgot about them because it was too much work to read through the whole thing. I've been asking Claude to try and help me make sense of all these notes--identify the main objections I seem to have, suggest ways of summarising the main claims of the book, etc. It makes everything go about 10x as fast. Also useful for more fine-grained writing help: fix this sentence which I know isn't quite grammatical, suggest an example to use here, and so on.
One thing I've found it is weirdly bad at: searching a text to find good quotes to illustrate a point you want to make. It doesn't even seem to be good at looking through my notes and finding good quotes I've purposefully written down.
One thing about advice is it works differently coming interpersonally from a single person like a friend, than from a public writer offering advice to an entire audience. And when you receive advice from any source you may hear it not only as a recipient, but as a potential sharer already anticipating wanting to pass on this advice to others. The benefits are not only social; there is inherent satisfaction in offering your own chosen wisdom to somebody. It can even be an affirmation of personal values.
I liked your own advice, aside from the LLM one - but do you really think people like Nate Silver, Matt Yglesias and Rob Henderson are appropriate models for the vast majority of beginning bloggers? (I know nothing about Ed West). These guys are basically celebrities; the first two honed a distinctive brand over the last decade or two and not only from blogging, while Henderson has a very unique and compelling story and first came to attention through a popular book. Even someone like yourself, who started small, was at least an established academic who could use that network to help spread the word. I'm much more inclined to look at people who are just random schmoes but have gotten up to a few thousand. Or if more popular, someone like Erik Hoel or Henrik Karlsson, or Sam Kahn, who devote many posts to reflecting on the art and culture of blogging.
Hmm good points. On the second one, to be honest I didn’t give much thought to who I listed after “e.g”. - I just linked some big names that came to mind when I was writing. I think your observation is probably correct. What I wanted to convey there is just that there are many more successful bloggers than me who have given good advice. If I remember correctly their advice is pretty widely applicable, but it’s true that for many aspiring bloggers others (including those you mention) would give more relevant advice.
About the advice in #12: it could be interesting, as a future post, to walk readers step-by-step through your entire process of using an LLM for some selected piece, to show exactly what this looked like under the hood. Not to mention a refreshing change-of-pace from the usual.
Re. the inherent satisfaction of offering advice: in general I think you don't pay enough attention to pleasure (as in, pleasure for its own sake), as a powerful motivator for behavior along with any evolutionary or social functions.
Yes fair point re. pleasure. I agree that's a blind spot. And thanks for the AI piece suggestion - I might do that if it would interest people.
i don't think it's puzzling. giving "free" advice shows others that you are a useful and valuable member of the community. they are more likely to help you in return, ally with you, follow you, see you as a leader (higher status), and generally want to keep you around. i can see it being adaptive.
i've always wanted to write/start a blog myself even though i think nobody will read it haha. writing about stuff helps you learn and clean up your thinking too, so even if nobody reads it, i'd improve from it. readers would be a bonus.
my problem is the opportunity cost. i just don't feel like i have the time. i also tend to get hyperfocused and i know i won't be able to balance my blog with work and personal life.
i'll probably never get around to it but it's okay.
i'm glad for your blog/substack though; it's one of my favorites.
I think you'd write a good one! But yes, the costs, including opportunity costs, are fairly high.
My partner tells me she would cry for days if she received some of the angry and hostile comments I’ve received throughout my blogging career.
Dan i read 30-40 substackers extensively and you are the most politically correct,moderate and normal looking face out of the bunch,to the point i am surprised i enjoy your stuff so much. What can they even hurl at you. And substack is less toxic that other platforms
Ha - lots of people really hate politically correct, moderate, and normal content. But yes, I'm sure I don't get it anywhere near as badly as most actually controversial writers.
May I ask what kind of text-to-speech software you use? I appreciate. Thank you.
NaturalReader. Didn't do much research, though. If you find a better one please let me know!
Good read. Thanks for sharing.
"Another fear some writers have is that people will read their blog and hate it and so hate them, and then maybe try to cancel them. People are typically far too paranoid about this. Nevertheless, if you’re terrified of getting negative reactions from what you write (i.e., you’re a coward), blogging is probably not for you, at least when it comes to non-fiction, opinion blogging."
...
"However, subject to those constraints, the most interesting writers are interesting in part because they take risks. "
...
"In general, quite a bit of self-promotion is necessary to reach and grow an audience."
Noted.
https://philomaticalgorhythms.substack.com/p/blowing-that-whistle-my-kafkaesque
"Blowing that Whistle: My Kafkaesque Journey in Causal Machine Learning
When you witness fraud so flagrant you begin to doubt your sanity"
I noticed you linked David Pinsof’s “bullshit advice” post, and seemed to nod to it in the intro. Was this piece inspired by that one?
Great advice - I enjoyed the essay even though I have never ventured into the terrifying world of blogging.
You are gay for robots.