🌿  back

20.92 YRS

saving versus sampling

There are two fundamentally different ways of capturing your thoughts in a digital medium for later use. The first one is the saving approach. It consists of taking the time to capture new thoughts which you find interesting there and then. If you come up with an idea which you find valuable in the moment, you immediately jot it down using your favorite piece of thoughtware, setting yourself up for serendipity when later stumbling across it. If you learn of an idea which you judge to be more mediocre, you might simply let it go. Regardless of the threshold of interestingness which plays into whether a thought gets persisted or not, the saving approach is all about intermittently capturing ideas based on implicit heuristics.

At the other end of the spectrum is the sampling approach. This one completely makes away with idea valuation performed by the user, and instead relies on probing one’s thought process at regular intervals of time which are defined in advance. This is closer to what you’d see in a psychological study, where participants undergo what’s called experience sampling – repeated inquiries into their mental state throughout the day. Just like in peer-reviewed studies, the probes into your thought process are also automated and scheduled when using personal thoughtware. The frequency of inquiries could range from once a day up to once every 15 or so minutes.

Choosing the best approach for you is simply a matter of understanding your goals of data collection endeavor. If you want to be able to tap into your interesting ideas at will during key stages of your workflow, then the saving approach might be a better fit. It’s tailored for capturing as much insight as possible from your intellectual activity, shielding it from the haphazard forces of forgetting. In contrast, if you’d rather have a more accurate representation of your thought process at large, you might choose the sampling approach. Rather than serving food for thought at key moments in your work, sampling enables a high-level overview of your informational diet, helping guide tweaks and avoid failure modes in a top-down way. In the conceptarium-ideoscope duo, you’d choose saving if you’re more into the conceptarium, and sampling if it’s the ideoscope you’re really after.

I personally am a fan of the saving approach, a preference which I suspect is shared by most tools for thought enthusiasts out there. That said, it might prove informative to consider the several ways in which sampling seams to fare better. For starters, what of the very heuristics we use when judging whether an idea deserves to outlive forgetting, to defy Ebbinghaus? There’s no way of avoiding the possibility that our in-the-moment intuitions about the value of ideas are suboptimal. For instance, we might get all hyped up about an idea which we’d judge infinitely less valuable a week later. Not to mention the very assumption of ideas having different values, being comparable and rankable. Don’t get me wrong, I really like this way of conceptualizing thought, but we shouldn’t leave assumptions unchallenged for long, lest we incur ideological debt.

Another aspect to consider in the saving-sampling debate is the nature of captures themselves. Our thinking is way faster than even the fastest mainstream methods of inputting thought into a computer (i.e. typing and dictating), so there’s a time cost for capturing an idea, from a few seconds for some marginalia to a minute or so for a coherent note or flashcard. Imagine you’re in a flow of insights, a cascade of powerful ideas which give way to each other. If you were using the saving approach, you’d have to fragment your line of thinking with constant interruptions for persisting the insights. If you were using the sampling approach, chances are you wouldn’t be bothered by any inquiry probe right at that moment, as those are relatively sparse and spread out. The saving approach sometimes tends to interfere with the very process it seeks to promote and nurture.

A final consideration in this debate is related to what I call the saving daemon. Similar to a software daemon doing some routine job as a background process on your computer, the saving daemon is a mental process which evaluates new ideas to determine whether they’re worth persisting or not. Such a daemon naturally makes for an indispensable part of the saving approach, it’s just a techy metaphor for pointing to its idea valuation component. However, the metaphor goes, any background process needs some resources to run on. In the case of the saving daemon, those are mental resources, as the daemon is posing a certain cognitive load on the user. There’s a bit of the brain involved in being on the lookout for insights, so the saving approach implicitly has a toll on its user l, albeit almost negligible. Consequently, a sustainable saving workflow must include the act of temporarily turning off the saving daemon in order to achieve truly restorative leisure. This could mean restricting access to thoughtware tools a few hours a day, even if frictionless is one of their main selling points. Alternatively, you could employ a lighter version of the daemon, slightly reducing precision but significantly improving the working memory footprint. Strong breaks allow us to go faster.

All in all, both approaches have their share of pros and cons, some more salient than others. I feel this debate will become increasingly relevant in the upcoming years, as thought capture will become increasingly commonplace in knowledge work, and as more frictionless tools for achieving exactly that are launched. Until then, make sure to keep your saving daemon in check.

This article is heavily based on a recent conversation with Maier Fenster, an avid proponent of the sampling approach.