29 January 2022

The under-discussed benefits of learning how to learn

Some skills are valuable in their own right, usually because they're rare and valuable. Some skills enhance other skills. I've heard compelling communication and basic time management, for example, called "multiplier skills."

The ultimate multiplier skill is learning how to learn.

I'm grateful to live in a time and place with excellent examples and resources on learning how to learn. The Bay Area is filled with thinkers from Eliezar Yudkowsky to Paul Graham to Naval that value epistemology and adaptability. And Silicon Valley espouses the idea that it's not the strongest, fastest, or even smartest that survive, but those that adapt the best. 

I've also found academic work on how to learn valuable. It's been exhilarating to experiment with spaced repetition, energy-steady diets, power law distribution of resources, tacit knowledge acquisition techniques, and deep work. I hope to post about some of my favorite tactics and resources in the future.

But despite all the talk and resources about learning, I wonder if the most exciting aspect of learning how to learn goes under-discussed: the positive feedback loop between motivation and ability.

Many of us develop terrible learning habits in school. We focus on knowledge deficits rather than celebrating development, we divorce learning from curiosity, and we associate learning with stressful competition usually governed by Goodhart's Law. For the lucky, our demotivation and misunderstanding is temporary. For most, it's permanent.

Even if we were to emerge from childhood unscathed, we'd still be unlikely to be aware of all the things we're not even trying to learn, but could learn to do extraordinarily well. Failure of imagination and learned helplessness collude to make invisible the alternative realities with the most outsized outcomes. The danger isn't regret, but our lack of regret.

Up against this backdrop, learning how to learn is all the more valuable and difficult. The first time you succeed in learning something truly intimidating you may wonder if it's a fluke.  You might caveat your success -- you did it slowly, or imperfectly.

But with each rep, the idea that you can learn anything starts to shift from something you want to believe to something you actually believe. And with the belief, comes new motivation, which can lead to more practice, which can increase your speed and further motivation, and so on. This is especially true when you succeed where you once failed.

This is very exciting. But almost none of the value is in the intellectual idea; it has to be visceral. The magic is in experiences that make the belief subconscious. 

I've gotten better at intentionally creating experiences to shape my subconscious, but one of my most influential experience happened unintentionally. Shortly after college, I began working long hours for my dream job at a prestigious management consulting firm. Months later I was blindsided by a devastating performance review that said I was error-prone, unmotivated, offensive, and unresponsive to feedback! 

I indignantly plotted a startup and prepared to leave, but a scary thought prevented me from quitting: what if my manager was right? Was I risking an important lesson just to maintain the comfort of a confident self-image? Too new to the domain to assess my own confidence vs. competence, I decided to suspend judgment for at least six months, and assume in the interim that my review was gospel, that I had failed, and that I needed to humbly learn consulting all over again. 

The emotional surrender was painful, but I found just enough self-compassion to motivate routine experiments and reflection. I started tracking my time religiously. Every Monday morning for a year I spent half hour selecting 3 new efficiency experiments and reviewing last week's results. Slowly, my efforts paid off: I rebounded and was promoted, even as I decreased my weekly hours by 20%. I left the firm with a reputation for being efficient, rigorous, and coachable.

I've written elsewhere about what I did after consulting: I went from first line of code to Google offer in 8 months.

For me, the most rewarding parts of these experiences were not the promotion or job offer, but learning to feel that the impossible was possible. I started to associate a calmness, even excitement, with the cognitive dissonance of learning, rather than discomfort or anxiety. Growth mindset became more subconscious.

I remember my first SWE interview. Hearing the prompt felt like getting punched in the stomach. I thought,"I don't even understand the question, and I have to solve this puzzle in an hour!?" A younger me -- a version of me that clung to A's in school -- would have choked and never recovered. But in the weeks leading up to the interview I had lots of reps solving coding puzzles that first seemed impossible, and my next thought was, "The tougher they seem the sweeter they solve." My subconscious had started associating confusion and dissonance with impending delight. 

Someone later told me, "It barely sounds like you were interviewing in order to get a job. Solving the puzzle was your drug. You were in it for the hit!" He might have been right. Stanford neuroscientist Andrew Huberman talks about the cyclical relationship between dopamine and norepinephrine during learning and neuroplasticity. Norepinephrine levels grow as we experience cognitive dissonance, a defining experience in any difficult learning process. Too much norepinephrine will trigger resignation. But as you perceive yourself crossing certain milestones, dopamine -- the reward hormone involved in both motivation and addiction -- is released and norepinephrine thresholds are reset. The holy grail of learning how to learn may be to tinker with our own learning-reward conditioning so that we literally become addicted to learning.

Every once in a while, you meet someone who has so much agency they appear to think they could run through walls. I heard a well-known entrepreneur say over dinner recently that they truly believe they could solve any problem, tackle any challenge, but that it's extremely hard to hire people who seem to share this trait, and she can't teach it. She concluded people must just be born with it, and it's rare. 

I disagree. I think it's much more likely that through design or good fortune, she's racked up enough experience that her subconscious has more agency than most. I believe this because I've watched my sense of agency change over time, and this is a very exciting observation, because it means my sense of agency is malleable -- I can plot ways today to have more subconscious agency tomorrow.

There's one final reason I think we undervalue learning how to learn. By making it easier to start flywheels that dramatically change our motivations and abilities, we also learning how to learn also gives us more control over our future personal preferences.

Shortly after becoming a software engineer I enrolled in a grueling 15-20 hours of night classes. I graduated a year later just as I was starting to burn out. I took an extended vacation... and surprised myself a month later by considering a PhD! As an undergrad I couldn't fathom doing a PhD, but a few learning journeys later, the idea of a PhD had become fun. I've had somewhat similar experiences in learning to enjoy getting in shape, dating, and meditation.

I find it fun to think about the world we'd live in if school and social norms celebrated tricks to learn, from epistemology to memory management to neural annealing. If we could all learn Kung Fu the way Neo learned in the Matrix, the real question wouldn't be "what do you want to learn?" but "who do you want to become?"

27 January 2022


Welcome! I'm Stephen.

I'm a grateful meditator, effective altruist, and software engineer. 

I'm starting to blog to imitate some of the people I most admire. Many of them find their blogging habit to encourage new thoughts, better flush-out ideas, and connect to likeminded people.

Things I have or hope to write about:

  • hard-won lessons and failures
  • behavior-changing insights
  • changes of mind
If you share my excitement about something or have thoughts, I'd enjoy hearing from you @zerfas33 on Twitter.


Why I blog

I'm starting this blog for a few reasons.

Writing improves thinking. Nothing is so good for cracking the illusion of explanatory depth as explaining something out loud or on paper. And I'll get to test it publicly.

Writing encourages more thinking. When I sit down to journal a few thoughts, I almost always write more thoughts than I sat down to write. Many of these are more uniquely my my ideas. That's both powerful and fun.

Blogs scale well. I'll be able to point to some of my best ideas and works as needed in the future. And just as code I write can run while I sleep, so too can my content be read. Naval calls code and media "permissionless leverage."

Like a consistent journal, a blog might serve as a paper trail for the evolution of my thoughts. By capturing my thoughts over time, a blog may both help maintain a sense for how malleable my beliefs are, and in retrospect, highlight what catalyzed my best changes of mind.

A blog could improve my retention. I think there are diminishing marginal returns to content consumption, especially without sufficient creation or applied learning. If I read loads of stuff but don't talk about it, use it to shape a decision, or commit it to spaced repetition, months or years later it might be as if I never read it at all.

I'm also persuaded by Paul Graham's heuristic that when trying to choose an obsession that matters, one that involves creating something someone else consumes is a good place to start. Perhaps writing about any of my obsessions will help me chase passions in ways that will be more valuable to others.

A blog may improve my writing; I hope to get better with practice. 

I wonder if a blog could affect my thinking well beyond the blog. Perhaps when I learn something new I might think "What here is important enough to write about?" or "How might I explain this to someone else?"