On HackerNews, this article was linked. The general idea is that companies are studying what people like to read, to help authors produce books that people like to read.

Now, for me, when I look at this idea, I see some down sides, but I certainly see some benefits as well.

Almost none of the commenters on NYTimes seemed to see any benefit whatsoever to studying reader behaviour. There were a few who saw the downsides as more mild than the other commenters. But most of the commenters basically saw this technology as some sort of 1984-esque idea that will turn all books into uninteresting, unimaginative pieces of paper that would better serve as a door stopper than as something for literary consumption. Out of 50 comments that I've read, only one person has said something along the lines of, 'This technology can possibly offer something to help authors improve their books'.

Is this just technophobia? Or am I missing something, and this really is a horrible, evil technology that should be avoided at all costs? [That's a rhetorical question -- I'd be surprised if even one LWian held that position]

I guess what I'm asking is, what are the psychological roots for the almost-unanimous aversion to this attempt at gathering and using information about what people want?

New to LessWrong?

New Comment
35 comments, sorted by Click to highlight new comments since: Today at 9:57 AM

I think what they're afraid of is something like the sickness occupying Hollywood, where they're very good at following a couple formulas and very bad at doing anything that'll actually amuse the sort of people who watch a lot of movies and have gotten bored of the formulas. Used in moderation this is a good thing, but marketing culture is very bad at taking profitable ideas in moderation.

I agree. My first thought was that too much measuring leads to convergence on a not-very-interesting standard.

From hanging out with humanities people and technical people, I'm heavily inclined to put this down to straight-up Hansonian status competition. Humanities people and technical people tend to downplay the value of each other's fields. It provides a gratifying, group-unifying status boost, and brings few consequences because the victims are on the other side of campus. In general this is fueled by subtle misunderstandings between groups.

In this case, humanities people will be overestimating the degree to which the techie's studying literature are simplifying things - for instance they might think that the tech people are actually saying their algorithms can capture the complexity of human reaction to literature, instead of just providing useful approximations. Thus, it feels like an insult to them, and a chance to pounce on the hated enemy, and in the process affirm their allegiance to the ingroup, who after all are the people who really matter. Artificial intelligence, neuropsychology and economics all get similiar reactions.

I would have attributed most of the reaction to people being sick of being spied upon.

[-][anonymous]10y110

High-end IQ people do stoo-pidt things. Like being so supportive of reading that they want less reading to happen rather than the 'wrong' kind of reading. Another example: ignoring that more people are writing for pleasure than ever before because it's 'just' texting and facebook.

The support of reading and writing is often as an alternative to more mindless uses of time. If reading and writing becomes as mindless as TV (via Twitter, text, etc.), then I see no good reason to advocate it any longer.

I've noticed some interesting status stuff around reading. "Reading" is used as a shorthand for "reading (mostly long form) fairly high status material". The large amount of reading which is normal for gamers doesn't count as reading, and neither does the huge amount of reading and writing that's involved in fanfiction.

Did you know that novels used to be low status?

And, of course, valorizing reading leaves out that there are many mediocre and worse books, and a huge amount of high quality movies, television, and graphic novels. Some graphic novels have a lot of text, too.

Heading further out into connotation-land, it's possible that having nothing but the words (voice?) in your head is considered more intellectual and higher status.

I don't know that I see it that way. Some of it is just plain old discipline and focused effort, which I don't have. I do much better reading articles, short essays, blog posts, etc. than I do reading books. Therefore, I covet the ability of people who can sit down and absorb a textbook or classic novel. I'd imagine they can often channel that ability toward other useful tasks, too.

Also, more than high-volume readers, I covet the abilities of prolific, interesting writers. And, it is my hypothesis that most good writers do a good bit of reading -- more than the average person, I'd say.

Most good (performers of a task) do a lot more (studying of that task) than the average person, yes.

How on earth can composing written material be mindless? I mean, YouTube comments come close sometimes, but even making up stupid Facebook memes demands a level of skill with the English language that 3/4 of people didn't have a century ago.

For that matter, how is reading any less mindless than TV? I tend to prefer reading, because it tends to be more lengthy and detailed, but the fundamental experience of sitting in a chair as a story enters through your eyeballs is basically identical.

I'm not making an argument against reading. Nor did I say it was mindless. I was making a comparison between some of the cheaper sorts of reading and writing, and highly passive entertainment (i.e. TV). The medium is of only relative importance.

Of course, even within TV, there is a spectrum. And there is some funny, novel, useful stuff on Facebook, though it is the exception to mostly drivel.

I'd argue there is generally a much more involved process involved in reading vs. TV. Reading requires the manual cognitive creation of audio and visual stimuli while TV serves it on a platter. It's really not even close.

One other point: Bad memes via Facebook is not something I see as "good", per se. You might say that is often ignorance manifesting itself through basic literacy; irrationality spreads through the use of language.

And there is some funny, novel, useful stuff on Facebook, though it is the exception to mostly drivel.

I think that largely depends on your facebook friends.

I think one of us misunderstood the other. You said you liked reading because it's less mindless than TV, I said TV is not particularly mindless, you talk like I said reading was mindless.

Facebook is not mostly drivel, it's mostly a computerized manifestation of its users' lives. Boring people make boring posts, interesting people make interesting posts, and you keep track of party plans and weddings and deaths.

As for reading forcing you to mentally create audio and visual stimuli, that's never been my experience. I'm an avid reader, and I read the words on the page. Descriptive text is basically lost on me, because I'm simply not imagining what the characters look like. I may possibly be capable of turning a written description of someone's looks into a mental picture of them, but I can't think of a time when I have actually done so.

And no, bad Facebook memes aggravate me. But they're better than no writing at all. Reading and writing may be getting less classy on average, but that's because more people are doing it, and plebs gonna pleb.

I have a hard time believing you sincerely think reading is "basically identical" to TV in terms of the mindlessness of the activity, by any reasonable definition of the word "mindless".

Facebook is not mostly drivel, it's mostly a computerized manifestation of its users' lives.

I've no idea what sort of distinction you are making.

and you keep track of party plans and weddings and deaths.

Like a calendar?

As for reading forcing you to mentally create audio and visual stimuli, that's never been my experience. I'm an avid reader, and I read the words on the page. Descriptive text is basically lost on me, because I'm simply not imagining what the characters look like. I may possibly be capable of turning a written description of someone's looks into a mental picture of them, but I can't think of a time when I have actually done so.

Noted.

And no, bad Facebook memes aggravate me. But they're better than no writing at all. Reading and writing may be getting less classy on average, but that's because more people are doing it, and plebs gonna pleb.

I'm a pleb. And it's ludicrous to say every meme is good. Every meme created exhibits non-zero intelligence and is therefore evidence of skill in language...so... I guess you've technically made some sort of point.

Tap.

I sincerely think that the mediums of TV and books are largely the same in intellectual requirements(subject to necessary levels of literacy, of course). I think the sort of people who read a lot of books and don't watch much TV read intellectual books and come across drivel on TV, which gives them a false impression, but the same thing happens the other way for someone who watches Breaking Bad religiously and only comes across Fifty Shades of Grey as far as books go. The choice of medium is, so far as I can tell, completely irrelevant. Passive entertainment is passive entertainment. Now, if you were comparing either one to computer games I'd say the games win an a heartbeat, but books = TV = radio = movies = plays = listening to your buddies bullshit over a couple drinks.

The distinction I'm making is that Facebook is only drivel if your friends lead drivelly lives. If you have interesting friends, Facebook is an interesting place. And I was unaware that calendars inform me of my friends getting hatched, matched, and dispatched. Here I just thought they told me what date it was.

I think you are wrong in saying that no one claims benefits from it: claiming benefits is practically all the linked article does. (BTW, your link goes to page two of the article. You may want to fix that. [Edit: Fixed.])

The article gave one viewpoint (and left out the other), and so everyone else is trying to give the counterpoint. (Not that I'm saying it's wrong for the article to only give one side: maybe debates work better for transmitting information than balanced pieces. But it certainly is the correct response to try to steelman the other viewpoint when you see an article in favour of one side.)

I don't think that's what they were doing. The commenters (the NY Times commenters, btw, not the Ycombinator commenters) seem to genuinely believe that it is only bad and no good.

"It might be the time to download “1984” from your Scribd or Oyster subscription service. I'm sure they have it."

"Surrendering your thoughts: A Haiku

Creepy. Nasty. Yuk. A good way to hasten the Singularity "

"I'm going to find out the top 50 favorite words and then write a book using only those 50 words. Who cares about creativity? It's about the money, kids."

I don't think these comments come out of a desire to just present the other side fairly. I think that this is just, straightforwardly, what they think about the concept of studying reader preferences.

Focusing on producing only the media that you think the public wants will discourage imaginative artists from creating new visions of what the public doesn't know they might like.

The most useful aspect of this service would be to prevent people from writing things that people don't want to read. Anything that stops people wasting their time is nice.

No amount of data mining is going to specify the next mold-breaking instant classic, but hopefully it can quantitatively back up a reduction in vampire romance novels, or as the article points out, extremely boringly titled and seemingly boringly written history books.

things that people don't want to read

vampire romance novels

People do want to read vampire romance novels.

You're not wrong, but there is a vast oversupply of vampire romance novels, to the extent that publishers don't even look at new submissions that fall under that category. So in the interest of matching available talent with market demand, writing vampire romance novels is in almost no one's interest.

Upcoming writers should try to chase trends, then step themselves a foot forwards of that trend. Current trends: interconnection, futorology, threat of conflicts that often don't realize, surveillance, and of course new years day and selfies. : )

The most useful aspect of this service would be to prevent people from writing things that people don't want to read.

Effectively you are saying that you want to censor unpopular stuff and this seems to be an effective way of doing so.

Often times society advances precisely because someone writes something that people don't want to hear.

With technology you have to be careful what you wish for, because it measures what you tell it to measure and optimizes towards that goal.

Often times society advances precisely because someone writes something that people don't want to hear.

Often these are things that some people in power don't want to hear, not what people in general don't want to hear.

I think that part of this is a worry that it will invoke Goodhart's law. There's also the issue of feedback loops; people get used to what's on the market, and the market finds it safer to put out whatever people are used to.

In addition to ygert's good comment I want to mentioned that maybe there is a status thing involved too. Maybe you noticed that in the natural sciences and in engineering there is the meme that something is either mathematical or not worth doing. A more extreme form is to see themselves as the most important workers in the world. Ackknowledging that, yes, studying literature to learn what makes a great piece of literature just like studying engineering to learn what makes a great bridge is a worthwhile and valuable endaveour, will hurt their status in the hacker community. Just go to a site like reddit and see how English majors are bashed on by engineering majors.

I am guilty myself of hating the humanities. Until the exact day I realised what the subject is about and that studying at college does not have to be a purely economic decision.

There was a bit of ambiguity on my part: the commenters I was referring to weren't Hacker News commenters, but the commenters on the original article itself, on NY Times.

Any initial approach of measuring anything (here: book interests) can be attacked on its simplicty. Neccessarily the first measure will have lots of unwanted special cases and it is just too easy to point these out - especially if they are obviously biased toward economic value.

[Trying to simplify] is a goal of dumbing down books to get stupid people to understand them. That's why people cite 1984. Orwell's newspeech is also about dumbing down intellectual discourse.

It is also easy to laud any approach of measuring based on the gain of information/knowledge it entails and transitively the improvements that gains.

What gets measured gets optimized.

It is always more difficult to see where such a measurement really leads. Extrapolating can be done on different complexity levels.

  • The simplest is always to just assume maximization of the measureable (here: maximize fun and by sales monetary gains).

  • The next level is to consider effects of the maximization and balancing effects (it may fall short or it might tip the system; here: dumbing down authorship tips with Hollywood-sickness result).

  • Another direction is to consider improvements to the measurement. What else might be measured to give authors feedback? And what results from maximizing that?

  • The next level is to consider the effect of the public on the process of measurement.

  • Ultimately one can envision to model this as part of a societal dynamic stabilizing on a fixed point. I wonder whether prediction markets can reach such a level.

My prediction:

  • The public will voice concerns (as already seen in the comments). This alone will initially have little effect as long as profits can be generated by the measure and not much actual change is visible.
  • Alternative measures will be introduced by competitors. Some tuned to niches e.g. by measuring not only fun but 'quality'.
  • NGOs will monitor companies' (ab)use of the measure(s).
  • Policy makers address such issues.
  • Consolidation sets in.
  • Development over time is documented.

Actually one can always build such a prediction by assuming the topic develops sufficiently to be come worthy of study. The direction in time is mostly math (can it be done) -> technology (how is it done effectively) -> economics (how to generate value from it) -> sociology (how does is affect people) -> politics (aggregate people affects) -> philosophy (reflection and rationalization) -> historics (posthoc documentation)

Damn. I can't write short comments.

Is this just technophobia? Or am I missing something, and this really is a horrible, evil technology that should be avoided at all costs? [That's a rhetorical question -- I'd be surprised if even one LWian held that position]

You seem to underrate the amount of contra-contrasrism on Lesswrong. For most subjects you will find someone on Lesswrong that can take the other side of the debate.

I guess what I'm asking is, what are the psychological roots for the almost-unanimous aversion to this attempt at gathering and using information about what people want?

Take a look at the discussion about Transcendence. You will find a sentiment that a film isn't supposed to be judged by the amount of fun that reader have in the cinema but that the film is supposed to be judged on the way it frames the societal debate.

Pop culture is culture that optimised on short-term enjoyment and pleasing the audience. Serious fiction is supposed to be optimised for more noble goals.

Effective technology that focuses on giving people what they want by looking at what people read is technology that engages in what we label on Lesswrong as "wireheading".

Ugly things happen when you start to pay all journalists by the click instead of focusing on accurate reporting.

Nassim Taleb said on the subject: "Truck driver who read books don't want to read books that are written for truck drivers."

I guess I don't imagine the idea always being used to that degree. I can imagine someone writing a new classic novel and they turn in their first draft of their next draft to their publisher, and their publisher says something like, "This sentence structure...studies have shown that it's a bit too complicated for most readers to parse on the first read, and they can take 3 or 4 times reading it before they understand what you were trying to say. Try to simplify it or break it up into multiple sentences."

I mean, that's not the only example. That's a rather mild example of how this sort of data would come into play, but I guess the examples I think of are less, 'Shelf full of Twilight novels' and more 'Same variety of books we have now, written with structure that's more in tune with how people read and think.'

I want a shelf full of Twilight as little as the next guy. But I also see that this sort of data can be used in helpful ways as well, not just used to produce the next mind-numbing teen fantasy.

I can imagine someone writing a new classic novel and they turn in their first draft of their next draft to their publisher, and their publisher says something like, "This sentence structure...studies have shown that it's a bit too complicated for most readers to parse on the first read, and they can take 3 or 4 times reading it before they understand what you were trying to say. Try to simplify it or break it up into multiple sentences."

That's a goal of dumbing down books to get stupid people to understand them. That's why people cite 1984. Orwell's newspeech is also about dumbing down intellectual discourse.

Is a book provides the reader an intellectual challenge that's not supposed to be a bad thing. Authors of serious fictions do have a license to provide their readers an intellectual challenge.

I don't have a problem with an author making a stylistic choices to use simple language, but I wouldn't want to create a complex system that enforces a dumbing down of literature.

That's a rather mild example of how this sort of data would come into play, but I guess the examples I think of are less, 'Shelf full of Twilight novels' and more 'Same variety of books we have now, written with structure that's more in tune with how people read and think.'

That depends on the quality of the computer algorithm. If you have an algorithm that can tell a publisher the percentage that a specific book is going to be the next Twilight, that publishers might start making publishing decisions based on that number.

There are monetary pressures.

In the New York Times journalists have to suddenly care about readers interests via hard numbers. If you want to read about what damage that dynamic did to journalism read Ryan Holiday's "Trust me, I"m lying".

[-][anonymous]10y10

#

[This comment is no longer endorsed by its author]Reply

When is a mass audience ever going to be interested in a comprehensive historical-fiction of the American steel industry?

That sounds like something James A. Michener would have written. So, roughly 1947 to 1995 or thereabouts?

Has your grandfather looked into self-publishing?

Especially since he can't expect to make any money off it even if he gets a regular publisher, given the niche... There is zero reason for him (or his kids, I guess, to be more likely) to not just put it up on the Internet and provide an EPUB for people to read on their Kindles etc.