Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Group Rationality Diary, March 22 to April 4

6 Post author: gjm 23 March 2015 12:17PM

This is the public group rationality diary for March 22 - April 4, 2015. Here's the usual summary of what it's about:

It's a place to record and chat about it if you have done, or are actively doing, things like: 

  • Established a useful new habit
  • Obtained new evidence that made you change your mind about some belief
  • Decided to behave in a different way in some set of situations
  • Optimized some part of a common routine or cached behavior
  • Consciously changed your emotions or affect with respect to something
  • Consciously pursued new valuable information about something that could make a big difference in your life
  • Learned something new about your beliefs, behavior, or life that surprised you
  • Tried doing any of the above and failed

Or anything else interesting which you want to share, so that other people can think about it, and perhaps be inspired to take action themselves. Try to include enough details so that everyone can use each other's experiences to learn about what tends to work out, and what doesn't tend to work out.

Thanks to cata for starting the Group Rationality Diary posts, and to commenters for participating.

Previous diary: February 15-28

Rationality diaries archive

Comments (16)

Comment author: erratio 26 March 2015 02:26:31AM 9 points [-]

Using my boyfriend as a combination Pomodoro, task log, and accountability device is still proving to be quite useful. It helps that he's extremely willing and happy to be used that way :D

Comment author: Gunnar_Zarncke 29 March 2015 08:14:22PM 0 points [-]

I'm 'using' a friend as a sanity check and reasoning validator. I know that I over-adapt to new ideas quickly and he is a natural but friendly skeptic so he has to endure my never ending stream of ideas (my own or second hand).

Using friends to boost you rationality (by using differnt strength of thought could be a pattern that could see more use, or?

Comment author: erratio 07 April 2015 03:05:52PM 2 points [-]

I am of the opinion that most friends could stand to use each other a lot more than they currently do. But it takes a particular kind of friend to be able to say that explicitly without offending them :p

Comment author: CAE_Jones 29 March 2015 01:49:45PM 8 points [-]

Last summer, I installed Leach Block in Firefox.

This morning, I disabled it, because I determined that I'd rather waste time in Firefox than waste time in Internet Explorer.

Comment author: Vaniver 23 March 2015 05:38:03PM *  7 points [-]

I've started consulting the I Ching in the morning. (Today was "difficulty at the beginning," with a side of "abundance," if I remember correctly.)

Reasons why:

  1. I've been trying to read at least the foundational works of Chinese culture (I read the Analects and Dao de Jing a while back, and am reading through the Romance of the Three Kingdoms now), and this is one of them. But simply reading through the structure seems less instructive than actually engaging with the book as intended. (Compare reading the Hail Mary once, versus actually procuring a rosary and going through the motions.)

  2. A while back, someone made a "rationalist horoscope," with the premise that there were a bunch of generally useful things you could remind someone of, and systematically reminding people of those things (and scoring the horoscopes based on helpfulness) is a useful thing to spend attention on. (As I recall, they didn't have that large of a stable of advice, and so fairly quickly had to resort to repeating themselves.)

  3. A similar venture is Ben Franklin's virtues, which I've attempted to track at least once but didn't find particularly helpful. The I Ching operates on a slightly different level--instead of operating on modes of behavior / reactions, it operates on perceptions of situations.

  4. I've kept a diary for at least a year now, but have rarely managed to work a solid reflection component into it and will often go weeks without entering anything. By consulting the I Ching, I'm provided with a constantly changing perspective to apply to my day, and then at the end of the day can record what that made me notice and how I reacted.

I will also note that up until recently I was primarily familiar with the numerological component of this particular form of divination (consult the RNG for wisdom!) but not the philosophical component. That is, the best advice is designed with its particular listener in mind, but there is value in a library of advice (that one learns how and when to apply by trying it through continually changing circumstances). Indeed, one might see this sort of divination, at its heart, as randomized trials applied to life!

(Incidentally, I immediately thought of three parts of my life where the situation was as described by the lot I cast, and a fourth on the drive to work. I also would not have applied one of the pieces of advice in the baseline history, but on considering why not it seemed like the reason was insufficient relative to the potential benefit.)

Comment author: gjm 23 March 2015 07:41:46PM 3 points [-]


  • Generate your random hexagrams with a computer. (You probably already do.)
  • At the end of each day, indicate which of the things the computer told you turned out to be useful (or harmful) that day.
  • Have the computer keep track of which hexagrams are most frequently useful.
    • (Perhaps allowing for some time-varying component -- e.g., varying with day of week or season of year.)
  • Have it generate more-frequently-useful ones more often. (Not to the exclusion of others. Though this might be a self-correcting problem -- if you hear something too often, being reminded may be useful to you less often.)
Comment author: Vaniver 23 March 2015 09:44:37PM 1 point [-]

Generate your random hexagrams with a computer. (You probably already do.)

I considered this, and I probably will if I do it long-term (though, really, it'll need to be a phone app since I don't boot up my home computer until the late afternoon anyway), but for now I'm flipping coins.

I will note, though, that my translation has slightly more structure than just rolling a d64; two hexagrams uniquely specify the relevant reading. (It's rolling 6d8, but doesn't have the full 8^6 potential outcomes; it's actually only 4^6.) This means the Birthday Problem happens over a much longer timescale. I might get the same primary hexagram tomorrow, which is the same general problem/situation/perspective, but the actual advice may change. (Because the fourth line was generated in the 1/4th case instead of the 3/4ths case, I was advised to seek union, which I interpreted by seeking a meeting with a mentor that I wouldn't otherwise have sought, and he in turn advised me to set up other meetings that I otherwise wouldn't have sought.)

So it may be the case that I can figure out that the fourth line should be more yang than yin, or more likely to be reversed than 1/4, or so on, but I think if I model at the level of individual hexagrams it'll be a long, long time before I've shifted the priors enough to have an effect. (Of course, one could set up a hierarchical model that takes all this into account, and with more people using such a system the time required for data collection would decrease.)

Comment author: gjm 23 March 2015 09:59:25PM 1 point [-]

two hexagrams

Ah, OK, that would indeed make a difference. My apologies for my ignorance. Still, if you have a primary and a secondary hexagram, maybe it suffices to look at the "effectiveness" of the primary hexagram only.

Comment author: adamzerner 24 March 2015 02:41:09AM *  4 points [-]

Keeping logs (using the "Seinfeld Productivity Technique") has worked well for me. I've done push ups and achilles exercises almost every day for a month. I plan on adding ab exercises, "go to sleep early", and "no junk food" for next month. I tried adding them in the last two days, but have been unsuccessful.

Comment author: Epictetus 17 April 2015 11:56:37PM 3 points [-]

My exercise regimen has turned into a habit. I've been getting strange feelings in my muscles, as though they want to be exercised. Skipping a day makes me feel wrong. Not in a rational sense, but rather I feel on an intuitive level as though something that should have happened didn't. In other words, I've been going to the gym because my body tells me to and it feels weird not to go.

I have had this experience before (~10 years ago). However, the last time I injured myself and ended up taking enough time off for the habit to break. Hopefully it works better this time around.

This gave me the idea to do something similar with my thesis work: to form a sufficiently strong habit of working that failure to do so just feels unnatural.

Comment author: Gunnar_Zarncke 20 April 2015 09:47:27PM 1 point [-]
  • Established a useful new habit

Smile more. Yes I know it helps more if you are an optimist and I'm just a realist - but a positive one to first derivative.

  • Obtained new evidence that made you change your mind about some belief

  • Decided to behave in a different way in some set of situations

Used have made a habit of writing very precise meeting minutes and even copied that into the info mail. But nobody read it. I still write minutes but only send links so when people follow the link they arrive at all the information. I also plan to focus on one single most important point in my mails to help focus people better on the key points.

  • Tried doing any of the above and failed

Tried a haircut I always wanted to try and failed. It was more flashy than planned. But this way at least I got quick feedback.

Comment author: hamnox 16 April 2015 11:50:29PM *  1 point [-]

Things I've been up to:

Keeping a composition notebook as a lab notebook for life. The very first entry is "Keeping a Research Journal", where I'm recording metadata about this attempt and my self-confidence in doing important science subtasks before and after.

Most recently, I did some random-sampling with tagtime to help me locate convenient straps to bootstrap trigger-actions with. At each ping, I wrote down my current actions, what in my internal experience or environment triggered me to be doing that, and maybe a few things I thought I 'ought' to be working on instead. It was a more frequent sampling rate than I've ever done before (~10min between pings), and the most significant conclusion to pop out was that getting interrupted by random pings all the time really, really stressed me out. This highly skewed my samples towards anxious internet browsing.

Now I've got:

  • Think about opening social media -> open mood survey
  • Feel prideful -> reinforce the moment with msg to bf or tumblr
  • Feel bored -> put on some music (feeling 'bored' is the often the first sign that I'm about to go into overload)
  • Feel confused about next actions -> Take a walk outside
  • Notice it's after 10 am -> make tortilla toast (simple enough to make while starving, in case I forgot to eat)

I'm deliberately not adding things until I've got a firm handle on the composition and effects of my baseline habits. My history with self-improvement is a long sad tale of accidentally tearing apart vital building blocks to fuel the production of shiny paperweights.

Comment author: RowanE 28 April 2015 01:20:44PM 0 points [-]

Okay, this is a stupid bit of pedantry but, the word "prideful" is different in meaning from the word "proud", in that it's specifically negative and synonymous with arrogant, and doesn't refer to e.g. when you're proud of a job well done.

Comment author: Gunnar_Zarncke 29 March 2015 08:21:05PM 1 point [-]

Using a (stochastic) time tagger has not worked for me. Apparently I'm not compulsively checking my phone enough. I miss lots of pings.

At first I thought I could just build enough habit. Didn't work.

Then I thought just enough data would solve it. Until I noticed that there are times where I don't reliably track as realiably as in other times (e.g. at home when the phone is out of earshot).

Then I thought I could salvage pings by smartly analysing them - evaluating tags relativ to the hour they come it thus compensating high-miss times. Until I realized that they correlate with the type of tag.

Then I gave up.

The last use of the ping right now is to remind me of writing this post.

Comment author: IlyaShpitser 16 April 2015 09:54:25PM *  0 points [-]

There is an interesting missing data problem here. Not necessarily unsolvable either. A lot of causal inference stuff is precisely dealing with this kind of data.

Comment author: hamnox 16 April 2015 09:51:43PM 0 points [-]

Sorry it didn't work out for you.

It seems like such a brilliant idea, and it doesn't work for me either. I still use it occasionally when there's something important I want to randomly self-sample for a few days, but I quickly train myself to ignore pings and alarms after a while.