Rationalist Horoscopes: Low-hanging utility generator?
(5:29:23 PM) Adelene: …horoscopes are for people who, like [mutual friend], prefer to have an authority tell them what to do. *blink*
(5:30:18 PM) Alicorn: *blink*
(5:30:50 PM) Adelene: This is an observation that my brain made just now, but it seems to make a fair bit of sense.
(5:31:18 PM) Alicorn: Plausible.
(5:32:21 PM) Adelene: Especially given that horoscopes seem not to actually make predictions about the future: They say ‘X is a good thing to do today’, not ‘X will happen today’.
(5:32:36 PM) Alicorn: *nod*
(5:32:53 PM) Adelene: …rationalist horoscopes?
(5:33:07 PM) Alicorn: like what?
(5:33:33 PM) Adelene: “Focus on granularizing your goals this week.”
(5:33:43 PM) Alicorn: hmm
(5:34:09 PM) Alicorn: divided according to some mechanism like star sign, or no?
(5:34:38 PM) Adelene: The only advantage I see to that is that it may make it more emotionally plausible.
(5:35:06 PM) Adelene: There may be some other advantage to having different people doing different things at any given time tho.
(5:35:23 PM) Alicorn: According to [other friend], birth *season* has empirically interesting effects in a few areas...
(5:36:35 PM) Adelene: I don’t think we can cash that out very well into advice, and anyway I expect that having that close of a similarity with actual horoscopes is likely to provoke a memetic immune response for most people. Could do it based on some kind of personality test tho.
(5:36:50 PM) Alicorn: *nod*
(5:39:08 PM) Adelene: Really, I think the bulk of the utility of such a thing would be in giving people generally-useful cues to work from—having any given day’s horoscope (or whatever we’d call it) be randomly picked from a set of good ones that haven’t been used recently should be just fine.
(5:39:15 PM) Alicorn: *nod*
(5:40:00 PM) Adelene: Maybe pair it with a rationality quote of the day, too.
(5:40:10 PM) Alicorn: Yessss.
I know it’s a silly idea, but it seems like it might be useful. I’ve played with random quote dispensers in the past, and if they have a good list of quotes to start with they can be surprisingly useful, in my experience—the quote might be useless 9 times out of 10, but that tenth time, when it makes you realize that a connection exists that you never would have noticed otherwise, is pretty awesome. Something like a daily horoscope might have a similar effect, but in a more practical way, getting people to consider taking actions in contexts where they wouldn’t usually consider those actions and occasionally finding an unusually good, but not intuitively obvious, match. And that’s on top of any benefits that such a system would have for people who do work better when they’re told what to do.
Thoughts?
I used to write rationalist horoscopes. All of them were “Today your life will be determined by the laws of physics and logic.”
For some reason, they never really caught on.
Most horoscopes work because of the forer effect, which is an application of positive bias. Now, if everything we do with these “horoscopes” is based on this effect, then there would really be no appreciable benefits of this project. However, if one could describe a typical important crisis/situation in a significantly vague manner, you could perhaps get people to recognize an important problem in their life that they have been neglecting. Then, perhaps give some generic advice that, thanks to our brain’s heuristics, brings a solution to mind.
For instance:
Just typing that, I actually recognized a personal issue that’s been generally percolating through my mind recently, but I haven’t explicitly devoted any time to solving the problem. I’m going to take my own advice now.
That’s basically the idea, yes. The forer effect might be used to make things emotionally salient in useful ways, but there definitely should be more to any given horoscope than that.
This is a good idea, but coming up with good horoscopes and judging their quality is hard. What if every time you give a horoscope, you also ask how good the previous one was? So if on one day the horoscope says to pay attention to X, on the next day you’re asked whether you noticed anything important about X (and if so, then more people are told to pay attention to it).
Oh, this gives me an idea on what you could do instead of random groupings: Have some simple learning algorithm divide everyone into some fuzzy set of categories given ONLY the data abut how good they rank previous pieces of advice. Just have some highly ranked standard pieces that new people are given at the start and group people for whom the same ones worked in the same group. Then when a new piece is entered into the system it tests it against all the different groups and tend to give it to those in the same groups as the ones it worked for in the original test.
The “pices” could be different things: quotes, advice, statements about your personality, predictions, whatever.
One thing that’s important to note is that you can be in any number of groups; A “rank maximally good” rock would be placed in ALL groups. The groups would not be anything made to mimic some specific human word but just unsupervisedly lerned from the raw data. In practice the groups might turn you to correlate to things like “optimistic” and “likes quotes” but also things like”guillable” or “wont actually follow the advice just votes based of what sounds good”. Oh, and while I’m saying things like being “in” or “not in” a group it obviusly shuldnt be a binary thing just a probablility of geting pices belonging to that group.
I was thinking something similar. Kind of like non-mutually-exclusive, dynamically-assigned star signs based on what you find useful.
That does also suggest that you could use the system prescriptively instead of simply descriptively. If it places you in the “talented slacker” category, and you’d rather be in the “fastidiously disciplined” group, you could opt to receive the Fastidiously Disciplined horoscope, and try to change your working habits to facilitate the Fastidiously Disciplined advice.
It’s not obvious to me how to do this and still gather information from the user without interfering with the scores for their chosen category. Having them guess after the fact how well the “slacker” horoscope would have worked for them seems clearly sub-optimal, especially since there’s an obvious pressure for them to say that it wouldn’t’ve.
I bet it would be useful to sort people by “what do you most want to improve about yourself?” It seems every LWer has at least one thing (and some, many).
People who choose “nothing” would end up getting horoscopes centering around Dunning-Kruger, confirmation bias, etc.
Yea. Only problem is the groups wouldn’t be labelled since they were autonomously discovered and and thus finding it would be a bit hard.
Well, they wouldn’t be labelled with meaningful English titles, but you could give them arbitrary names for ease of reference. A bit like actual star signs, only empirically informed.
yea, that’d probably be a good idea.
Good idea, but I’m not sure if it can be worked into an rss feed, and if not it’ll be tricky to implement. I plan on asking a more tech-savvy friend about it later tonight, though.
If we did do this, it seems to me that it’d need at least two dimensions: Likelihood of the advice working, and usefulness of the advice when it does work. The second might actually be more important than the first, for this purpose—if something works consistently, it seems like it’d be better to establish it as a habit rather than being intermittently reminded of it and otherwise ignoring it. (And we could certainly publish a list of suggested habits if the data points that way for enough things.)
I would say that the whole thing, setting up the server, RSS feeds, and some basic statistics gathering, excluding the database of horoscopes themselves, is a one- or two-weekend project for a reasonably experienced web programmer. Javascript and forms won’t work in all RSS readers, but links will. So an entry in the RSS feed would look like this:
And you do some server-side magic at the destination to gather stats about who clicks on which links, filter out duplicates, and decide what to put in the feed next.
Seems like it might be worthwhile to flesh out, perhaps develop as an app / plugin / whatever the kids are up to these days. May need to be rebranded....
Is there any consensus on ‘personality types’, e.g. is Myers-Briggs useful at all? That could be one way to divide the horoscopes up.
(You could also divide it based on blood type or something but give the same advice to everyone—if you’re obvious / sarcastic enough about it.)
I suspect it’d be pretty easy to set up a tumblr account, or several, for the actual distribution of randomized suggestions. Those come with rss feeds, so that should make just about everybody happy on that end. (I have no idea how tumblr interacts with smartphones, but I expect that they can handle rss feeds—correct me if I’m wrong, though.)
We’d also need a way to have more suggestions added, though, and have those moderated in some way. That gets a bit trickier, but may be able to be semi-automated—perhaps any LW post that gets more than a certain number of upvotes could be flaged to have someone check it out and condense it into a few sentences if it’s appropriate.
Rebranding is the big hurdle, though, and not something I’m useful at.
Hmm… there could be a LW post (similar to the “rationality quotes”) devoted to the purpose. Tumblr also has an “ask” field which could be used for sending in suggestions as well… though there’d have to be someone on wheat / chaff duty, and expecting a good deal of chaff.
As for rebranding...
It could be “__ Horoscopes” where the blank is filled with something communicating “secular” (though the concept isn’t quite right) or “arbitrary” (jokes could be made about the relative arbitrariness of regular horoscopes...) or even “actually useful” (which I kinda like). I thought I had another idea, but it’s gone poof, so I’ll let you know if I come up with anything brilliant.
P.S. Do tumblr rss feeds work properly for you? I’m subscribed to ~one but it spams me with updates when all that’s changing is the number of comments/reposts. Am I doin’ it wrong?
Having the system require ongoing maintenance in order to have any output at all is a weakness. Also, having hand-submitted ‘horoscopes’ posted in order, as tumblr’s submission function does by default, removes the randomness element and makes it much harder to include older concepts. It might work, but it wouldn’t be all that much better than just keeping an eye on the new posts here, I expect.
My idea is to have tumblr just serve as an output system, and have a separate program that procedurally generates horoscopes from a list of sentences and then use email publishing or something to post them. That way it only needs attention when new things are going to be added to the list, rather than every few days.
I like “Actually Useful Horoscopes”, if we’re keeping the term ‘horoscopes’ at all.
Yeah, I’ve never had a problem with it… are you sure you’re using the right feed url? It’ll be something like http://example.tumblr.com/rss
All right, this finally spiked my “just fix the rss thing” action potential, so thanks for that.
My reasoning is that “horoscope” is mostly there to anchor the idea to something people “get” already. It could be something like “Rationality Bites”, I guess, (“Sensible Bites”? :) ) but there isn’t really a “space” carved out for that in most peoples’ brains. Maybe something like a secular equivalent to “inspirational thoughts” that do more than just inspire...? That’s another idea that fills a similar role...
Oooh, or “Fortune Cookies”! Add descriptors to taste—“The Fortune Cookies of Highly Effective People” ;)
Yep.
Also, I’m liking the ease with which one can say “my horoscope told me to...”, either to oneself or to others who’re aware of the project. ‘Bites’ doesn’t work too well with that, and neither does ‘fortune cookies’. I guess it’s minor enough that those could win out on the grounds that they don’t make one sound like a nutter.
(We’re using ‘Actually Useful Horoscopes’ for the dev phase, but it’s easy enough to change tumblr account names.)
I suspect that you could openly tell everyone that the groups are randomly assigned, if the grouping has some inertia to it—that is, if your being in the same group as someone today is a moderately strong predictor of your being in the same group as them again tomorrow.
Yeah, you’d want to be open about it, probably even emphasize it. We could even give personality descriptions based on these explicitly random groups, providing a bonus rationality lesson! (“Hey, that does sound like me!”)
The hard part, of course, is getting people to disbelieve once they notice that it seems to work.
Hah, yeah, we don’t want to start an accidental cult or anything. Maybe, as you suggest, switching their groups randomly on occasion would help? :)
my rationalist horoscope is my daily anki deck review of the biases and fallacies flashcards.
So something like Brian Eno’s Oblique Strategies?
I’ve found them useful in the past; interestingly, I get a similar effect with spaced repetition—more than once a quote or factoid has come up and I suddenly see how it fits somewhere or what it’s really saying.
Sounds similar, yeah. The random quote generator I had in mind was actually a bit of code in a ‘pet’ in Second Life, which chose a random quote from a list and displayed it to everyone in the area every 10 minutes or so. The quotes weren’t anything special, just a bunch of things I’d saved because they sounded interesting or useful, but I got the same effect you’re talking about, of every so often having one go *click* somehow. (Also it was quite amusing when it’d insert a particularly relevant quote into the middle of a conversation, which happened perhaps once or twice a week.)
This does not look like horoscope, but like the well known practice of a daily/weekly proverb and/or suggestion. Which is a great idea.
You can make a rationalist calendar out of it, or a service that sends the user a nicely segmented suggestion every time period.
A bit more context-based version is the Force Your Connections trick from the Mind Performance Hacks book. It’s basically how all those movie/computer game/postmodernist paper online title generators work, except the grammar that defines the generated phrases can be relevant to a problem domain. Morphological analysis is apparently the version of this that’s presented in a way in which it sounds justifiable to hire a very expensive consultant to do it for you.
I think people who don’t believe in the supernatural and still practice “magic” do something like this with a Tarot deck.
I did this (minus the practice ‘magic’ part) for a year or two with a Tarot deck. Like any semi-structured meditation technique, it sometimes worked and sometimes didn’t.
I’d probably enjoy something like this, just because it would give me an anchor point on “what should I focus on TODAY”, without having to do a cognitively expensive cost-benefit analysis of all the techniques I’d like to get better at. I tend to waste far too much cognitive effort trying to pick where I should start...
I like this idea, but I would lean towards quotes from the sequences (or occasionally HP:MOR) rather than the ones that you find in the quotes thread, often are imprecise to the extent that I would lose the emotional utility (I think this is the reaction that you’re referring to when you say ‘memetic immune response’.)
That’s part of what I mean by memetic immune response, yeah.
I was thinking of quotes from the quote threads, but not of using all of them—only ones over a certain karma score, and even then they’d need to be vetted to make sure that they’re generally useful and don’t depend on some specific context.
If someone or some group wanted to generate a bunch of quotes from the sequences, that’d work too, but I’m not personally interested enough in that aspect of the idea to put that much effort into it.
Horoscope content ideas, separated for voting, please contribute:
Take the time to evaluate your default responses to questions. They may not be as good as you think they are.
Possible alternative: Today, you may not give a default answer to any question more meaningful than “how are you?”.
Today, indulge your curiosity. Find the answers to three interesting questions not directly related to your current projects. (Bonus points for finding answers to questions that can’t be answered via Google or Wikipedia, like the name of that cute person you see in the library every so often.)
Take some time to optimize the things you pay attention to for signal to noise ratio. Remove boring RSS feeds from your reader; find out how to stop receiving junk mail or getting calls from telemarketers; get rid of objects that clutter your living space or work space; perhaps even have a heart-to-heart with that friend you’ve been growing away from.
Today is a good day for starting new things. Choose something new that you’ve been putting off, and spend at least an hour working on it.
If you think about it carefully, the universe has recently shown you several examples of people working together successfully. What common principle can you see at work in these events?
(Horoscope idea based on this study.)
What do you want? What are you doing to get it?
Think about how your ideal self would handle the challenges you encounter. What would you do if you were smarter?
Alternate version, because imagining what you would do if you were smarter is fairly impossible:
Think about how your ideal self would handle the challenges you encounter. What would you do if you always followed your goals?
Actually, unless my memory is playing tricks on me, there’ve been reports from people who’ve tried the ‘what would I do if I were smarter’ trick, and had good results with it. It sounds like it shouldn’t work, definitely, but it seems that in practice it does anyway, probably by cashing out to “what would I do if I wasn’t stuck” or something.
Here’s an example from no less than Claude Shannon.
There seems to be a general trick to expecting more of yourself.
I can get more willpower by thinking “I don’t have enough strength, but I put myself in the hands of the Lord, who’ll give me his strength”—and I discovered that trick long after I became an atheist! I suspect the trick is to enforce the expectation of undepleting willpower, which makes it true.
It might be a rigged self-test. Where else can we test if the trick works?
The being stuck thing sounds like it matters a lot here. Thinking hard is a lot of work to begin with, and most habits people end up stuck with replace needing to think with rote responses that can be broken with a change of mindset.
A follow-up on this might be how well it works on people in different situations. You need to have some idea on what useful smartness is like, and someone who has grown up in very disadvantaged conditions might just not have enough exposure to that to form useful models. Someone with mental issues might not be able to break their habits of behavior to get benefit from the thing. People who are already pushing themselves hard cognitively, like math grad students, pro chess players or stock traders might get less out of it, though the exercise could still help them come up with a new perspective on things.
I wonder if there’s a generalizable attitude here. This reminds me a bit about the pedagogical advice that you should always tell children that their academic success is because they worked hard, not because they are talented, since otherwise they’ll model themselves as having some comfortable level of talent and stop pushing themselves to whatever actual limits their achievement might have.
It doesn’t seem to be just a question of working harder. You can still get stuck to thinking that you’re as smart as you think you are and therefore can keep working like you’ve always worked and do a bunch of ineffective hard work. Thinking what a smarter person would do also makes you question the quality of your metacognition.
Well, it was already pointed out that getting feedback on the usefulness of various horoscopes is a good idea, and coding has commenced with that as part of the plan, so I think the thing to do is actually try it and see how it works.
I had no idea. That’s really cool.
People do not, by default, apply the full extent of their intellectual capabilities to most problems. This being the case, to imagine what you would do if you were smarter requires only extending yourself to meet the smarter person’s default, background, unconscious level of smart thinking. In some cases this will be possible via the act of explicit attention and deliberate focus.
Yesterday you did something for which you probably want to apologize. What was it? Can you do something today to make it right?
Your subconscious is trying to tell you something! Get a piece of paper and, for each of the previous three days, write down 3 things that happened as they leap into your mind. Put the paper down and come back to it several hours later and write down the theme that connects these nine events and what lesson you should learn from it.
That thing you’re hesitating over trying, the likes of which you’ve never done before, is probably easier than you think it is. You’re likely overestimating the difficulty because you lack any reference for it. Bear this in mind when reconsidering whether or not to try it.
Question your questions. Are you asking the right things?
Meta-content: “Try to be more specific and concrete”. :)
(what I’m saying is that most of these could use a “for example...” or similar).
I think that’s going to be tricky to do in some cases without interfering with the near-universal applicability of the advice. That said, I’m going to give it a try with the suggestions I’ve already made.
See if you can understand the functioning of an object well enough to create it from raw materials or scrap. Start with something simple, like paper fasteners.
*contemplates how to make staples out of paperclips*
Making staples is the superior option only if you are racist or sexist. If you’re not, it’s more helpful for your rationality to make paperclips from other scrap. After all, you probably want a sense of “completeness”—that you’ve made something fully functional. And when you attach that “home-made” paperclip to your first sufficiently-slim stack of paper, you can feel good in having done all of it yourself. (Racists and sexists aren’t interested in that feeling.)
In contrast, if you chose staples as your project, how would you put them to use, to test their functionality? You would need a stapler. And unless you do this entire project again, but for a stapler—something more complex than a paperclip—as a SUBSTEP to your first achievement, you just can’t get that same feeling of accomplishment when you apply your first “home-made” staple. Rather, you will have to “live on someone else’s strength”—specifically, whoever made the stapler.
Also, paperclips are re-usable and make great gifts, if the recipient likes paperclips.
I suspect that someone could get that feeling from using homemade staples in a pre-made stapler, just the same as they could get that feeling from using a homemade paperclip on pre-made paper. (Paper’s a lot easier to make by hand than a stapler, though.)
On the other hand, hand-making a staple to fit in a standard stapler is a lot harder than hand-making a useable paperclip.
This makes no sense. I would prefer see less silly comments about racism. (This is different from the actual novel argument about the benefit of homemade paperclips over staples. That kind of creativity is mildly entertaining.)
Are you saying humans should make staples???
Are you trying to manipulate humans on this website into making a decision by associating the other option with racists or sexists?
Or is there some intrinsic relation between Clippies and Staplies, unknown to me, that makes helping Staplies over Clippies a form of racism or sexism?
Focus on granularizing your goals. When something you want to do seems impossible, try to break it down into a series of smaller steps.
One of the things that you’re told today, which you’ve heard a thousand times before, will be false. Find it.
Today, spend time thinking about possible failure modes of your plans. What are the most likely things that could go wrong, and how will you handle it if they do?
Keep an eye out for good ideas in unusual places today.
An etherpad document for horoscope theme ideas and partially-formed horoscopes has been started here.
The time for a lucid appraisal of your own abilities is prior to action, not in the middle of it. Once you find yourself engaged in real-time application of some skill or other, act as if your mastery of that skill isn’t at issue at all, rather than let yourself be distracted by assessments of the likelihood of failure, because they are likely to be self-fulfilling prophecies. (src)
Make a specific commitment today that you’re confident you can follow through on. Write it down and post it somewhere where you’ll see it regularly until it’s done.
Today, be paranoid. Don’t assume that the people around you are trustworthy, or sane, without evidence. (Evidence gained before today is, of course, admissable.)
Take some time to test your beliefs today. Devise and carry out at least one experiment.
Take the time today to put some important information in a more intuitive format. For example, you might make a pie chart or other visual representation of how you spent your money last month, or how you spent your time yesterday.
Take a minute to think which sentence or paragraph you spoke or wrote yesterday was the MOST effective in achieving your goals and which sentence or paragraph was the least effective. If the results have not come in, mark a note to your future self to do the same.
Are there any notes from your past self about analysis of previous sentences? Check if results have come in.
I think I work better when told what to do; therefore I am interested in this.
If you’re going to divide it up in a way to make it more emotionally available, maybe separate by Myers-Briggs category?
Well given the low reliability and low validity (see here for a bit of an overview) i’m not sure that would work to well
Splitting LessWrong by MB type isn’t going to be a even distribution. The Hamming distance from INTJ isn’t very great for most of us...
J? So that is what makes some lesswrong folks so damn annoying at times. I maxed out the ‘P’ rating. Technically equal T and F resolves to F too so in raw hamming that gives me a 2.
I didn’t figure this was supposed to be for Less Wrong specifically, although getting it to spread might be difficult.
That seems like the most likely way of dividing it, to me. I’m not entirely comfortable with endorsing that model (I don’t know enough about it, basically), but it looks plausible at least. (I’m not entirely sure we’re going to divide it up at all, at this point, but the code’s written in such a way as to at least allow it.)
I suppose it’d be interesting to allow the different M-B groups to optimize the quotes that the groups get separately, rather than having all the votes go into the same database—I don’t think it’d be all that hard to set that up, and it’d give us some interesting data on M-B.
I don’t think there is good reason to split according to M-B groups. If you want to split according to personality, the Big 5 factors are the current consensus. Even then though, I don’t know how advice would be specialized to low agreeableness individuals, for instance.
The broad categories come to my mind are
Epistemic
Instrumental/anti-akrasia/motivational
Social
Creativity/seeing with fresh eyes
Quotations
You could either bundle these with the instructions to pick a single one to focus on or if you have room for customization, let people pick whether they want a high or low amount of each in their stream.
The current sequence reruns might be filling a function similar to this.