Weekly Head Voices #180: EnsembleYou and The Describers.

Dear friends,

Welcome to this, the 180th edition of the Weekly Head Voices, in which I perform retrospection on the week from Monday September 30 to Sunday October 6, 2019.

To be more specific, this is mostly indirect retrospection, meaning that I think about and attempt to describe the thoughts I had about topics and events that might or might not have occurred in the period of time that this post is ostensibly looking back on.

TwiFRaF progress report.

Almost a month ago, all the way back in WHV #177, I announced that I would be cutting out TwiFRaF (Twitter, Facebook, Reddit and Friends) from my information diet.

This has been going quite well.

Subjectively, I’ve been spending substantially more time on more constructive activities. Also important, my work-day focus has improved; there seems to be more of it to go around.

However… A part of the vacuum left by TwiFRaF was being sneakily filled by YouTube recommender riding.

You know, load the YouTube main page, SEE ALL THOSE SUPER INTERESTING RANDOM CLIPS WHICH THE YOUTUBE AI KNOWS WILL GET YOU (in my case movie trailers, movie reviews and analyses, comics, running, HOW TO GET CHISELED (no idea how that got in there), Saturday Night Live clips, running really terribly far, some programming or computer thing) and middle-click on about five of ‘em.

Watch the clips, reload the main page, repeat!

As entertaining as this was, most of it couldn’t really count as a constructive use of my time, and so, at the start of October, TwiFraF was summarily upgraded to TwiFraF-Yo.

For the past week, I have observed a further uptick in the quantity and quality of my focus.

The Describers.

Thanks to an excellent recommendation by friend PK (he of the introducing of the rider and the elephant, I’m currently reading Stumbling on Happiness by the Harvard psychologist Daniel Gilbert.

I’m only about a third of the way through, but so far this tome is filled to the brim with knowledge about the thinking and feeling parts of the human machine.

(As an aside, the topic of the story I wrote in WHV #174 which no-one noticed, you know the one about your experience of reality being mostly synthesised and hallucinated in real-time, was discussed at length by Gilbert (the topic, not my story of course). According to Gilbert, this is what the philosopher Immanuel Kant introduced into the world under the label (Transcendental) Idealism. Knowing this, I feel like slightly less of a barbarian now.)

What I would like to focus on here though, is The Describers.

Gilbert talks about their research on focalism, the human tendency to place too much weight on one piece of information when making a prediction or a decision.

In their 2000 paper, they discuss an experiment where they compared two groups of college students (psychology research, also the important parts, is positively riddled with experiments on college students) with regard to their ability to predict the future emotional impact of a negative event, in this case, losing some important (just run with it, please) football match to some rival college.

The difference between the two groups is that they asked the one group, which they named The Describers (capitals and emphasis mine, I like all of the story-telling possibilities), to describe the events of a typical day (see p823 in the paper, this was a structured questionnaire), before predicting the impact of that future negative event.

In the end, the non-describers over-estimated the impact of the loss on their happiness, whilst the describers predicted much more accurately.

Although the main reason I’m writing about this here is that I’ve really taken a liking to the name The Describers, it will not have escaped you, astute readers, that this is yet another convoluted reason to write stuff down.

If the research above can be believed, and here I have to divulge that my life (debating) partner is still highly sceptical, journalling could contribute positively to the quality of your prediction and decision-making.

At the very least, the hypothesis that taking the time to re-process and document one’s thoughts, observations and reasoning could contribute positively to one’s judgement seems eminently plausible to me.


In machine learning, ensemble methods combine multiple different classifiers or algorithms, which in many cases comes down to simply averaging their output predictions, in order to obtain better accuracy than any of the constituent classifiers.

As I was chatting with a mathematical friend from work the other day, I rambled on and on (sorry mathematical friend) right into the idea of EnsembleYou™.

It was actualy the unsurprising admission that the in-the-moment me is really not very good at predictions, decisions or anything else.

However, continuously writing everything down is one practical way of averaging out the observations and decisions of a whole bunch of instantaneous mes, or yous.

(More specifically, this was about the fact that the “Vitals” section of my monthly journal files, containing lists of reminders for a good life, is copied from month to month, evolving slowly as I add, remove and incrementally improve its contents.)

Writing is the multi-purpose thread that runs backwards and forwards through time to connect the various different versions of me into a slightly less faulty entity.

I am calling this specific composite entity EnsembleCharl, but because I would like to give the idea to you, that will go through life as EnsembleYou™!

Friendly Wishes.

Friends(es), I wish you effortless connectivity, and great wisdom in your multiplicity.


comments powered by Disqus