Don't Prioritise What to Learn

The first post I’d like to critique is Decide what not to learn by Dylan Fitzgerald, a self-titled “voracious learner/metalearner”. He opens with an unsourced quote that goes “You can learn anything, but you can’t learn everything”. This is of course undeniable, but I struggle to see what relevance it actually has. In return I’d like to offer my own quote, but Mahatma Gandhi:

Live as if you were to die tomorrow. Learn as if you were to live forever.

Learning everything is impossible, but it is only in striving toward that unattainable goal that one can come close to it. Learning more is in-and-of-itself a worthy goal. He then goes on to talk about how “career-adjacent” learning can be valuable, but also “bottomless and neverending”. Once again, a trivially true statement phrased as a problem without any real motivation. His claim of this being especially true for technology workers is even more absurd:

there’s every day a new language, framework, process, or entire field of study promising to make the life of you and those around you somewhere from better to utterly transformed

The strange thing about this claim is that for anyone who has done any serious (or even amateur) programming will realise that the more languages¹ one learns, the easier it is to learn another one. Ideas don’t just come up out of nowhere, requiring us to build entirely new mental system to accommodate them. Instead they are based on other preëxisting things that we can use to “anchor” these new ideas to. The more one learns, the easier learning becomes. Getting a “Wikipedia-depth summary of each topic” is not a problem, as that knowledge allows one to create novel new pathways to new knowledge.

Fitzgerald proposes this as his solution to this “problem”:

If you, like myself, struggle with this, a tactic: write down all the things you want to learn, are in the middle of learning, haven’t quite had time for, know you really would benefit from…you get the idea. Pick the top one. The rest of the list–especially #2 and #3!–is the stuff you absolutely must avoid at all costs. That’s not your “next” list. That’s your “DANGER” list.

This is anathema to me. Fitzgerald’s ideas may be rooted in some sort of technical thinking — where the mind is some sort of robot; studying continuously before moving on to the next thing — but this is not how learning works (at least not in my experience). Learning is instead a fluid, moving process, where one moves from topic to topic as their interest waxes and wanes. We create the aforementioned pathways by learning new things that support each other. Knowledge is like a brick wall; much stronger if standing in a lattice of other bricks than if placed on top of each other one after another.

The second post is Neel Nanda’s Post 34: Learning how to learn. I agree much more with this post that the former one, but still think it falls into the same sort of thinking about learning as Fitzgerland’s. This is perhaps surprising, as the two are really diametrically opposed in their advice. Nanda’s post mostly focuses on how to learn through teaching, introspection, and spaced repetition, but it also does have a section on what he calls the 80/20 Rule, a personal spin on the Pareto principle². In it he says that you should find out how to “80% of the value from 20% of the effort”, and then do that. He expands on this idea in a post titled The world is full of wasted motion.

In summary, one should try and reduce (in both the literal and metaphorical sense) ideas and try and obtain a “big picture” view, to see what really matters. This reminds me a lot of advice that one should “go back to first principles”, axiomatic ideas that all other assumptions are built upon. Learning these are most important, and so learning the axiomatic basics and “key ideas” (the reduced knowledge) in a given field is what is most valuable. I agree with all of this in principle, but in practice I think it falls apart.

Stepping back and looking at the “big picture” is a common technique for trying to distance yourself from one’s obsession with whatever is in front of them, and to re-prioritise toward what is really important. For this purpose it is very useful; but as a general attitude for learning I find it almost as harmful as Fitzgerald’s suggestion.

The main obstacle to long-term learning, in my experience, is almost never time or “wasting” it on learning the wrong thing. Instead, it is ambition and interest. This may very well be a natural instinct guarding against what Nanda warns about — I usually have no interest in learning about something that is not useful³.

Just as it is problematic to try and avoid merely getting “Wikipedia-depth [knowledge] of each topic”, it is equally problematic to aim for it. Learning should be an intuitive process that we are prompted to naturally, and trying to “choose” what to learn strategically will almost always be a mistake. What (and when) will be useful is almost impossible to predict in my experience. One will always encounter newfound situations where one’s knowledge of an obscure topic will become relevant.

That learning should be fluid and intuitive does not however mean the rejection of structured, deliberate practice. Techniques like spaced repetition can significantly speed up long-term memorization and do not need to be discarded. But what you choose to learn through such things does not have to be equally rigid. Perhaps this is what modern schooling gets wrong? It sees that intensive, rigid learning works, but wrongfully applies that to the curriculum. Just because the metaphors (fluid vs structured) are antipodes does not need to mean the same in practice.

There are of course certain situations that in microcosm appear worth focusing on. If you are a material sciences researcher, it might be more important to read up on crystal structures than to study the work of Shakespeare. But this is a consequence of our society’s technical view; a material science researcher should only be focusing on material science — everything else is unnecessary. But in reality humans are multifaceted individuals who not only may be interested in multiple things, but they also interact with other people interested completely different things. Having a broad base of knowledge, or an unnecessarily deep knowledge of something niche, allows us to see new perspectives, but they also deepen human connection and makes us able to empathize with others.