Yeah, this is a great book. Curious to see how the rise of adblockers/anti-addiction tech will change things. There's a counterintuitive argument that things will become worse: As sophisticated folks tune out, unsophisticated folks become the most lucrative audience and the race to the bottom accelerates. As wise people leave the conversation, the conversation that remains is even more crazy. And unfortunately, even a small number of well-coordinated crazy people can do a lot of damage.
I actually think leaving comments online is a more scaleable strategy than people realize. I leave a lot of comments on LW, the EA Forum, etc. and I'm now no longer surprised when I meet someone IRL and they recognize my name. It took me a while to internalize how skewed reader/writer ratios are online and how many lurkers there really are. I suspect the lurker numbers for any kind of culture war discussion are even higher. It's like two people having a shouting match on public transportation: Everyone wants to watch, no one wants to participate. But the size of the audience means that if you do choose to participate, then you massively amplify your influence.
I once did an experiment where I registered a throway twitter account, searched for the trending hashtag controversy du jour, replied to peoples' tweets and tried to talk them down from their extreme positions. I was surprised by how few tweets there were, how little time it took for me to respond to all of them, how many people engaged with me, and how successful I was at getting people to moderate their positions a bit. I got the impression that if I had the money to hire 100 people to use Twitter full time and mediate every ugly discussion they saw, I'd have a nontrivial chance of moving the needle for a nation of 330 million.
I actually think leaving comments online is a more scaleable strategy than people realize. I leave a lot of comments on LW, the EA Forum, etc. and I'm now no longer surprised when I meet someone IRL and they recognize my name. It took me a while to internalize how skewed reader/writer ratios are online and how many lurkers there really are.
My experience has been similar. I also believe this is true of Wikipedia articles, and that's one reason I still engage. I'm less confident that I am making a difference by hosting fulltexts or scanning books, but I figure at some point I can do a time-series analysis of citations as a proxy.
As far as genetics goes, direct interactions haven't gone too well, especially on Twitter; if you're dealing with someone who flatly denies that GWAS hits replicate or that sibling comparisons prove causal effects or who claims that all hits are population stratification, at this point, there's no reasoning with them. So I try to simply publicize a little all the research going on for the hidden masses, hoping that it'll be Grothendieck's 'rising sea'.
Something about the model proposed here feels slightly off to me, but overall I think this changed my perspective on public discourse a bunch.
I am putting out a $50 bounty to anyone who creates a Sarah Constantin style fact-post about how many active online-commenters there are on major platforms like Twitter, Facebook and Reddit, especially with basic writing and persuasion skills.
Yeah, I'm not at all confident in this model, but I do suspect it's underrated. I'm reminded of something Andrew Ng mentioned in his machine learning class about how he would run into machine learning projects where they didn't have much data, and he would ask them to do a back of the napkin calculation to figure out how much time it would take them to hand-label more data. He said that oftentimes with just a week's time spent hand-labeling data, they'd dramatically increase the amount of data available and improve their algorithm's performance. It's not clever or "scalable", but sometimes the solution that doesn't scale is the best one.
This has definitely been true in my experience with ML/DL so far. If you grit your teeth and put a bit of effort into a reasonably low-latency script for hand-labeling, you can often do a few hundred or thousand datapoints in a quite feasible amount of time and it will be enough to work with in finetuning (eg. PALM) or training on an embedding (eg. StyleGAN latents).
And this is something that a lot of people have been learning with image generation models the past 3 years: it is often way faster to just curate a small set of images to, eg. train a LoRA, than to try to train some sort of uber-model which fixes the problem out of the box or use some complicated fancy algorithm on top of the existing model or brute force sample generation & selection. It's not necessarily that the model doesn't "know" the thing you want, it's that you can't tell it accurately. 'A picture is worth a thousand words', which is a lot of words to have to figure out! (Or right now - we've been working with a guy on InvertOrNot.com as an API service for better dark modes, to automatically classify images for website dark modes, and while GPT-4-V can do it if phrased as a pairwise comparison, one could never afford to offer that in bulk as a free public service as we would like to - so... he finetuned a tiny cheap EfficientNet on ~1000 hand-labeled images and it works great now and runs for free. Because you can't easily "prompt" for it AFAICT, but it's easy to collect a few thousand examples to hand-label for a pretrained CNN.)
Of course, the models or services improve over time and now you can often just zero-shot what you want, but one's ambitions always grows to match... Whether it's art or porn or business, once we can do the old thing we dreamed of, soon we sour on it and demand even more, something even more specific and precise and niche, and the only way to hit that ultra-niche will often be some sort of hand-labeled dataset.
The argument around ad-blockers and anti-addiction tech is an interesting one. While I agree that they make it more difficult for sophisticated people to be affected by the provocation and click-bait, I'm not sure that this automatically makes the unsophisticated audience that remains more lucrative. One of the hopeful things that Ryan mentions in the 2017 edition of this book is that the rise of paywalls is a sign that "serious" media outlets are beginning to realize that the pageview game is a game for suckers. They can't hope to compete with blogs who specialize in outrage, so instead of doing that they're differentiating themselves as "upmarket" publications who charge an up-front fee, in exchange for not showering you with clickbait.
My personal prediction is that we'll end up with a two-tier media market, where relatively affluent and more sophisticated readers continue to read things like the New York Times, or the Wall Street Journal, or The Economist, while less affluent or less sophisticated readers read Huffington Post and... whatever ends up taking Gawker's place. In a sense, it'll be like the old split between broadsheets and tabloids, only online instead of in print.
With regards to Twitter, I wonder how important it is that the handle be a throwaway one, with no prior history. I know that, for example, gwern has tried engaging with vocal skeptics of things like GWAS studies on IQ, and it doesn't really seem to have made a difference in their viewpoint. I wonder how much of that has to do with gwern's prior posting history putting him firmly on the "other side", causing people to dismiss his claims on the basis of who he is, rather than what the claims are.
Separate from my other comment:
He then targeted those blogs with a series of provocative actions, such as buying billboards promoting the movie and then vandalizing them himself, calling feminist groups to protest showings of the movie, arranging for provocative advertisements on buses, and other acts designed to go viral on social media.
It occurs to me that his relationship with these feminist groups may not have been entirely adversarial.
That is, my first inclination is to read a situation like this as: he wants people to see the movie, the feminist groups want people not to see the movie, and he slyly manipulates them into doing things that achieve what he wants and not what they want.
And maybe this should have been obvious, but - of course that's not the only thing feminist groups want. They might also want members; newspaper coverage; a highly-visible thing to rally around and point people at. (And if someone is the sort of person who would see the movie if they heard about it, maybe the feminist groups don't even really care if they do happen to hear about it and see it.) Perhaps "all publicity is good publicity" applies to both sides, here.
And if that's right, then this problem seems harder to solve.
A concrete example of this is work he did for the movie I Hope They Serve Beer In Hell, starring Tucker Max.
I think it's important to note that the movie was a box office failure. That's not the only thing that counts, Max was probably glad to be interviewed by the Post regardless, but it does seem like Holiday's approach failed at its main goal. Or if the Post coverage really was the main goal, then it's less obvious why we should care.
Does Holiday ever address this?
(Unimportant nitpick: Max wrote the book and cowrote the film, but didn't star in it.)
https://en.wikipedia.org/wiki/I_Hope_They_Serve_Beer_in_Hell_%28film%29
He does, but it's a weak reply, in my opinion. He claims that the viral marketing campaign was supposed to be only a part of the overall marketing, but that due to factors outside of his control, the funding for the more traditional marketing campaign fell through, and that the viral marketing campaign ended up being the entire marketing campaign for the movie.
Trust Me I'm Lying, by Ryan Holiday, is probably the most influential book I've read in the past two years. This book is a guidebook to the twenty-first century media ecosystem, and shows how the structures and incentives of online media serve to polarize us by stoking fear and anger. First published in 2012, the book is eerily prophetic in many places, as it talks about the toxic influence of online publishing on politics and society.
Ryan Holiday is a marketer and publicist who specializes in manipulating blogs in service of his clients. So what is a blog? A blog is any online publishing platform which derives its revenue from advertising. Blogs range in size from small, fly-by-night local publications all the way up to multi million dollar properties like Gawker and the Huffington Post. Blogs may be independent, such as Huffington Post or Politico or they may be associated with an existing media franchise. For example the Monkey Cage politics blog is hosted by the Washington Post, and MoneyWatch is a financial blog hosted by CBS.
The key observation that Ryan makes is that all blogs, no matter how large or small they are, no matter if they're associated with an existing media franchise or not, are driven by the same economic incentives. The revenue a blog makes can be expressed as (price per page-view) × (number of page-views). Since blogs have little control over how much they are paid per reader, blogs universally tend to work to maximize the number of readers they get. Moreover the sheer number of blogs, and the extremely low barrier to entry creates a brutally Darwinian marketplace where any content that doesn't maximize page-views is rapidly superseded by content that does.
In order to maximize the number of people clicking on their stories (and thus viewing the ads on those stories), blogs exploit every flaw in human psychology they can find. Chief among these is provocation. Ryan cites a study by Berger and Milkman from 2012 which shows that content with high emotional valence spreads much faster than content which is emotionally neutral. The study compared stories on the New York Times website, and found that articles which induced anger were 34% more likely than the median article to make the top-10 most e-mailed list. Articles that induced awe also did well, being 30% more likely than the median article to make the most e-mailed list. Both anger and awe are high-arousal emotions (in a negative and positive direction, respectively). On the flip side, articles that induce low-arousal emotions, like sadness, suffer a penalty. Sad articles were 16% less likely than the median article to end up on the most e-mailed list. These facts about human psychology act as constraints on the kinds of stories blogs will write. Every story has to make people feel a "high-energy" emotion, like anger, or awe. Stories that are thoughtful, practical, useful or beautiful but melancholy fall by the wayside.
Another constraint on blogs is their very structure. Marshall McLuhan's adage, "the medium is the message" applies just as much to blogs as it does to television. For blogs, the medium is the stack. Most blogs are arranged in a reverse-chronological list, with new stories coming in at the top of the page, and percolating down towards the bottom as further stories come in. A blog which can always produce something fresh at the top of the stack can draw in more readers by having more novel content for readers to click on and share. Ryan points out that, unlike newspapers, who have a finite number of column-inches per day and unlike cable news, which has a finite number of hours per day, a blog's appetite for content is functionally infinite. Blogs whose writers produce the fastest win, regardless of the quality of their writing.
The final constraint on blogs is in how they get new readers. In general, people don't subscribe to blogs like they subscribe to newspapers or magazines. Instead, blogs get traffic from links and headlines that are shared on link aggregators (like Reddit) or on social media. The reputation (such as it is) of the blog counts for nothing when stories are passed around as disaggregated headlines, each fighting for the reader's attention on its own. In order to "hook" readers, headlines have to be as provocative as possible. In fact, it's in the blog's interest to make headlines misleading, since a reader that clicks into a page and clicks away in disgust still counts as a page-view, which earns the blog money. And since people don't generally pay attention to the history or credibility of a source when sharing or clicking on links, the blog does not suffer any penalty for wasting the reader's time or attention.
These three constraints (virality, structure, and disaggregation) serve to create a set of entry points that allows a media manipulator such as Ryan Holiday to influence the content and framing of stories that blogs cover. Ryan noticed, contrary to prevailing wisdom, that most original reporting in online media was done by smaller blogs, whose stories were picked up and summarized by larger publications, until they reached mainstream media outlets and entered the "national conversation". Therefore, by influencing small blogs today, one could alter what was in the Washington Post tomorrow.
Ryan created a process, which he termed "trading up the chain" to do just that. First he would observe which large blogs the national media outlets he was targeting drew from. Then he would observe which small blogs those large blogs pulled stories out of. Finally, he would craft a media campaign targeting those smaller blogs, seeding the same provocative story in enough places to ensure that it would get picked up and passed up the chain until it received national coverage.
A concrete example of this is work he did for the movie I Hope They Serve Beer In Hell, starring Tucker Max. He wished to get coverage for the movie in the Washington Post. By observing the sources of the Post's stories carefully, he found out that much of the Post's media coverage originated from stories on Gawker. Going one level further he noticed that Gawker, in turn, pulled its stories from smaller city-focused blogs like MediaBistro and Curbed LA. He then targeted those blogs with a series of provocative actions, such as buying billboards promoting the movie and then vandalizing them himself, calling feminist groups to protest showings of the movie, arranging for provocative advertisements on buses, and other acts designed to go viral on social media. When those blogs inevitably began covering I Hope They Serve Beer In Hell, he ensured that there was enough "chatter" to quickly drive the story up the chain to the Washington Post.
The end result of these tactics was that an otherwise no-name D-List "internet celebrity" like Tucker Max was being interviewed by The Washington Post and late-night TV hosts like Conan O'Brien. Ryan makes the point that much of what we consider to be "organic" content, spreading by "word of mouth" is actually carefully engineered and seeded by professionals like himself to generate coverage for particular celebrities, products, and events.
The important thing to note is that all of these manipulations make the stories that we read less accurate. Moreover, the publishers of these stories bear no cost for misleading us and wasting our time and emotional energy. The toll that the exact on each and every one of us is a pure externality, just as much as smokestack fumes or toxic chemicals going into the water supply.
If the effects of this media manipulation were merely to drive customers to products they wouldn't otherwise buy, Ryan would still probably be out there plying his trade. What caused him to reconsider his profession (and write this book) was the increasing use of these manipulation techniques to spread political ideas, and, in the process, hurt individuals. In the second half of the book, he talks about how sites like Jezebel and Breitbart News use the techniques he pioneered to push product for American Apparel to maximize their own page-views by stoking outrage both among their supporters and their opponents. In his view, much of responsibility for the coarsening and polarization of politics and culture can be laid at the feet of professional manipulators like himself.
Ryan's hope is that by writing this book and exposing the actual techniques that manipulators use, we can inoculate ourselves and make ourselves less susceptible to the sort of media manipulation that he used to carry out. Though things look bleak at the moment, Ryan looks to history to show that our online media ecosystem today is very similar to the "yellow-press" era at the turn of the 20th century. Given that, he has hope that the current state of affairs is not sustainable, and we will eventually craft a stronger, more trustworthy online media, just as the provocative tabloids like the New York Herald and The World eventually gave way to more trustworthy publications like The New York Times and the Wall Street Journal.
Personally, I found this book important because it explained and crystallized many of the troubling trends I had personally observed in online media, and put them in a framework that allowed me to see clearly how they worked and how they were manipulating me. Though I knew that online media was growing ever more provocative and ever less accurate, my prior was that it was the result of blind evolutionary forces, as described in Scott Alexander's post on the same topic, The Toxoplasma of Rage.
While Scott's piece is important and insightful, it's still written from an outsider's perspective. It frames the ever escalating spiral of provocation as the result of groups competing to stoke the most outrage among their own members. Trust Me I'm Lying, by contrast, says that the result of deliberate manipulation by people who deliberately set groups against one another in order to bring attention to the issues that they want attention brought to. When Ryan vandalized his own billboards and organized protests by feminist groups against I Hope They Serve Beer In Hell, he wasn't attempting to send a message about the content of the film. He was merely operating under the adage, "Any publicity is good publicity." Similarly, Ryan's claim is that much of the outrage we find in today's politics and culture is the result of deliberate manipulation by publications who want to drive traffic to their own sites, without necessarily caring about which side "wins".
Where the book is weakest is in its advice about where to go from here. The book ends with a note that people's time and attention are limited, and eventually people will catch on to the fact that they're being manipulated, and will start to demand higher quality reporting, rather than merely quantity. In support of this contention, he cites the evolution of print journalism, which evolved from "yellow press" tabloids to newspapers that are widely considered to be accurate and relatively unbiased. However, he gives few predictions about how this will come to pass, other than noting the current media ecosystem is unsustainable and that unsustainable things cannot be sustained over the long term.
Nevertheless, I found the book insightful, entertaining, and more than slightly horrifying. As a result of the book, I can look at stories like this and better look past the manipulative elements to see how little substance there actually is to the article. As a result, I find that I'm more efficient at extracting the factual content from news articles, and better at identifying and avoiding so-called "fake news". For this reason, I consider Trust Me I'm Lying to be a strong recommendation.
If you're interested in a more detailed outline of the book, I have one on my wiki.