'Getting Empirical' with my rationality practice
I've recently been inspired to make my rationality practice more empirical.
Concretely, I envision to start a more rigorous process of coming up with and testing hypotheses about matters related to rationality, my personal productivity, well-being, social life and possibly more. By making this process more explicit, I expect to be able to gather more information, quicker, and make the pseudo-experimentation I've been doing (largely in my head) until now more systematic, more effective, more explicit and better documented.
Why? And what to I hope to get out of make my rationality practice more empirical?
First of all, there is the direct value of generating and testing hypothesis and interventions which will hopefully result in me resolving some of my bugs sooner (rather than later), and, ideally, in a more sustainable manner. Moreover, I pursue second, more indirect/meta goal. Namely, I want to reliably build up certain mental habits that I think are very useful. This is basically what I call 'improving the way I think' and, while I already know how to do these things, what I really I want to get to here is the process of internalizing these standards/mental moves to a degree that they become actual habits, reliable and don't require conscious effort anymore. Concretely, I want:
Empirical mindset: get in the habit (always, rather than sometimes) of coming up with falsifiable hypothesis ex ante, making my assumptions explicit, actively looking for ways the experiment could be biased and reasons my hypothesis might be wrong, actually going back to the experiment ex post and consciously process and update on its results).
Noticing and original exploration: This is still a little ill-defined and I don’t feel like I have capture everything I’m interested here yet. It's vaguely about training my mind to detect more bugs, more reliably (even the “not ideal but sufficient”-type bugs that I often struggle to detect reliably - write on this at some point) and incite myself to be more creative, original and proactive when it comes to thinking of and trying interventions, and generating hypothesis and models.
In practice: my ‘rationality experiments’
Some first thoughts (non-exhaustive) on how I want the process to go and what sort of parameters I might want to track in any experiment: (I expect this list to be significantly improved, refined and extended overtime. This might also be helpful to make experiments increasingly low effort while keeping a relatively high quality by systemizing the process as good as possible.)
Writing up my hypothesis in advanced and making assumptions explicit; ideally, I'd model the issue (using Yeb or on my remarkable) to refine my hypothesis and draw out as many of the underlying assumptions, aliefs as possible before the experiment itslef - I think this would likely increase the informational value I can get from running the experiment
Defining what indicators I'm interested in (eg. what would 'success' look like, what are things that might arise but that I am actually not interested in, …) and defining how I want to track them;
Thinking about obvious ways the experiment could be biased, invalidated or otherwise screwed. (I.e. how am I to interpret and contextualize the results?)
Defining a clear time window (start and end date) for the experiment; not only is this necessary for being able to properly evaluate the experiment, most of all, I think this is a quite reliable way to increase my dedication to 'actually try' and put me in the right mind space, instead of half-heartedly tinkering around.
Increase my commitment, rigor (social accountability) and documentation (external memory) via posting the set up and underlying thoughts of the experiment here.
Discuss with a third party (mostly my friends) to reevaluate the set up of the experiment and check for obvious obstacles or blind spots; and set up accountabilities with them if necessary/useful.
Set up everything else that is necessary to run a fair experiment (eg. check whether I have actually enough time and mindspace to sufficiently commit to what I've planned)
This type of experiments are obviously limited in many ways and I'm vigilant of not over-updating on any of the results. Most importantly, they are experiments about how I function (usually, n=1=me), and their external validity will in most cases be very low. At most, I'd hope that the experiments and write ups of my mental models and how they change might serve others as inspiration and/or prompts for their own hypothesis generation. In other words, I am hopeful my write ups will ask useful questions, rather than give final, conclusive answers.
Some things I want to run experiments on? (a loose list of very rough ideas)
Overwhelm (model building, phenomenology; test interventions)
Distraction, flinching and craving (test interventions)
Focusing (experiment with usage, ‘personalize’)
TAPs (experiment with usage, get better at)
Noticing (test interventions to get better at it; model building/phenomenology (are they substantial differences depending on what I try to be noticing)
Memorization skills (test techniques; overtraining)
Status and human interaction (test, explore ideas from Improv)
Energy balance and Total Energy expansion (test hypothesis, come up with interventions)
Productivity (test interventions; supposedly strongly related to the above, including things like working structure (focus work, light work, writing, reading, ...), meditation, sports, food, social interactions, thinking time, applications, (daily/weekly/monthly) checklists, etc.)
Team rationality (?)
Ways to learn certain skills
...
The next experiment I will run (and first under this concrete mental umbrella) will be about noticing when I’m flinching away from certain (non-embraced) thoughts; when I’m not able to do deep focus work but get distracted by too many thoughts.
Particularly, I’ve installed an accountability system with my co-workers to notice this mind state and to it by getting up and doing deliberate focusing. This serves to purposes: a) testing focusing as an intervention to respond to this ‘mind-restlessness’ and b) I expect the information I get from focusing on the source of mind-restlessness to improve my mental model of why my mind starts distracting itself, and hopefully illuminate my understanding of its causal patterns.
However, I will write about the details of the experiment, my underlying hypothesis and assumptions in more detail in a separate post.
Geneva, January 2019 - Thanks, Michael, for living your inner child with me.