필사 모드: Self-Directed Learning Is What Grows Me — Turning Career Hunger into Fuel

Opening — the night I shrank before one LeetCode problem

Late one night, there was a day when I had spent thirty minutes failing to solve a single LeetCode medium problem. When I finally opened the solution, it was so simple that I felt even more wretched.

That night I also looked through a few job postings, and half of the tech stack listed in the requirements was made of things I did not know deeply. The thought "is this really all I amount to?" pulled my mood down low.

I am someone who works at LINE and touches a fair amount of code every single day. And yet nights like that still come. The anxiety did not disappear as my experience accumulated. If anything, the higher my standards rose, the more sharply I could see the areas I did not know.

But after chewing on that anxiety for a few days, my perspective shifted a little. That anxiety was proof that I wanted to get better.

A person with no hunger at all feels no anxiety when they look at a job posting. Being anxious meant there was a place I wanted to reach, and that I was feeling the distance between where I stood and where I wanted to be. The problem was never the anxiety itself, but where I let that anxiety flow.

This post is about the direction of that flow. The core claim is simple. In the end, what grew me was not study someone assigned but study I felt a need for and sought out myself. And the fuel for that self-direction is precisely a healthy hunger toward career.

I did not learn only development this way. English and Japanese too, I held onto myself out of a felt need while working at LINE, not because anyone told me to. Even table tennis, which I play as a hobby, improved far faster through the strokes I repeated because I craved "I really want to land this smash" than through the motions a coach told me to drill. Self-direction, it turns out, does not discriminate by field.

Core insight — what stays in the end is the study you sought out

Looking back, the real leaps in my ability always came from self-directed learning.

The training the company assigned and the lectures I sat through out of obligation certainly helped. But what I learned that way faded fast, like knowledge right after an exam ends.

By contrast, the things I dug into at night myself — because I was stuck at work, because I was frustrated — remain vivid even years later. The few days I floundered while implementing concurrent-login blocking by hand grew me more than any formal lecture ever did. The traces of how I agonized alone over where to put the session store, how to detect token collisions — they are still intact in my head today.

Here is one more honest example. A certain security training the company made mandatory had most of its content evaporate the day after I received the completion certificate. By contrast, the experience of personally debugging a single outage at LINE and digging into a network timeout, I can still explain even when no one asks. The same amount of time was spent, yet the amount that stayed differs this much.

Why the difference? Because the kind of motivation differs. Psychology's Self-Determination Theory says that intrinsic motivation — work you chose and did autonomously — produces deeper learning and longer persistence than work imposed from the outside.

Study I started because I felt "I need to know this" is processed by the brain in a fundamentally different way than study someone told me to do. The former is solving my own problem; the latter is sitting an exam set by someone else.

So I put it this way. A company can give me opportunities and problems, but the core engine of growth is one I have to switch on myself. Learning that waits to be spoon-fed has a clear ceiling. Real growth begins the moment I feel the thirst and set out to seek for myself.

Going deeper 1 — the translation that turns anxiety into fuel

The anxiety you feel before job postings and LeetCode is, in itself, just an uncomfortable emotion. The real question is how you choose to translate that emotion.

The same anxiety can flow in two directions. One is self-deprecation. Flowing into "I guess I just can't do it, I guess I have no talent" leaves anxiety to paralyze a person.

The other direction is directional information. Translated into "so this is my weak spot — then let me fill it in here," anxiety becomes a map. It is the very same anxiety, yet depending on the direction of the translation, it becomes either poison or a compass.

What shows this well is Carol Dweck's research on the growth mindset. People who see ability as fixed take a difficulty as "my limit has been exposed" and avoid it, while people who see ability as something that can be grown take the same difficulty as "not done yet" and rise to the challenge.

The key is a single word: "yet." Changing "I can't do this" into "I can't do this yet." Half of the unknown technologies in a job posting are not my limit; they are my next learning list.

I actually write this translation down on paper. When anxious, I open the posting in front of me and write down every item I do not know. Then the vague anxiety turns into a concrete to-do list. The moment emotion becomes information, my mind grows considerably lighter.

A real example of turning a job posting into a weakness map

Spoken only in the abstract, none of this lands, so let me transcribe exactly the work I once did with an actual backend senior posting.

The requirements section of the posting had sentences roughly like these. Experience handling large-scale traffic, designing message-queue-based asynchronous architecture, database index tuning, guaranteeing consistency in distributed systems, monitoring with observability tools.

I honestly scored myself next to each item. Traffic handling, I have done in practice so I felt confident; message queues, I have used but my design was at the level of following what others built; index tuning, I knew only from books and had barely measured anything myself; distributed consistency, I knew the terms but my depth was shallow; observability, I had seen dashboards but never set one up myself.

Once I had written all this out, the vague "I am inadequate" turned into the sentence "this quarter I will dig into index tuning and distributed consistency, just these two." A single lump of anxiety was decomposed into five prioritized items. Decomposed anxiety can no longer crush me. It simply becomes work to handle.

What matters here is not trying to fill every item at once. Of the five, I chose only the two that appeared most often and where my weakness ran deepest as that quarter's goal.

The rest I pushed to the next quarter. Reducing my appetite was, paradoxically, the secret to finishing. Rather than trying to do everything and accomplishing nothing, digging two things all the way down left me with far more.

And one more thing. This weakness map is worth making even when you are not trying to change jobs right now. A posting is the most honest signal of what the market wants, and just reading that signal regularly keeps my learning direction from drifting out of step with the market.

Going deeper 2 — lack of skill makes a person petty

Let me say one slightly uncomfortable thing. In my experience, a lack of skill makes a person petty.

When your fundamentals are weak, your mind has no slack. Even when a colleague suggests a better approach, you cannot see it as a chance to learn; you feel as if your deficiency has been exposed, and you turn defensive.

When a new technology appears, you feel threat before curiosity. When a junior asks a good question, you fudge your way through it for fear of not knowing the answer. The weaker the skill, the smaller the person, and a person made small turns small-minded without even realizing it.

I have been there too. When a code review suggested an approach better than my implementation, even while my head knew it was correct, what came out of my mouth first was an excuse like "that doesn't fit our situation."

Looking back later, that defensiveness was not a problem of technique but a problem of confidence. Had I known that area well enough, I would have simply received the good suggestion with joy. In the end, the root of small-mindedness was fear, and the root of that fear was a lack of skill.

Conversely, when your fundamentals are solid, your mind gains slack. Hearing a better approach, you say "oh, that is better" and accept it. You can say you do not know when you do not.

You see a junior's growth as joy rather than threat. Skill does not merely make you good at the work; it makes a person generous. That is why I see fundamentals not as mere technique but as the foundation of the mind.

I see something similar at the table tennis hall. The more solid a person's fundamentals are, the more they review a lost game with a smile and praise the opponent's good shots.

By contrast, the more someone has thin fundamentals and has won by luck, the more they blame the racket and blame the ball over losing a single point. That the slack of skill shapes a person's attitude held true the same way in front of code and in front of the table.

So what counts as fundamentals? I will lay it out as a list in the next section, but in one sentence: not the flashy latest framework, but the things that age slowly. Trends change fast, but fundamentals age slowly.

The reason I solve LeetCode, too, is less for the algorithms themselves and more for that thinking muscle of decomposing a problem and recalling a data structure. It is not about memorizing interview tricks, but about etching into my body the habit of breaking any problem into small pieces and choosing the right tool.

Fundamentals cannot be crammed. Like a marathon, there is no choice but to build them steadily, even if slowly. The person who keeps going — one problem a day, even thirty minutes daily — ends up going the furthest.

What exactly are fundamentals — a list of the things that age slowly

We hear "build your fundamentals" all the time, but rarely have we actually laid out, concretely, what that even means. So let me write out the list of "things that age slowly" as I think of them.

First, data structures and algorithms. The ability to sense, by intuition, when arrays, hash maps, trees, and graphs are each fast and when slow, and to gauge time and space complexity. The library name in a particular language may change in a few years, but the fact that a hash map looks up in average constant time will not change.

Second, the basics of the operating system. The difference between a process and a thread, the cost of context switching, how memory is managed, why locks and concurrency are hard. Knowing this lets you form hypotheses instead of vague guesses in front of a question like "why did this server slow down?"

Third, networking. The difference between TCP and UDP, handshakes and timeouts, how HTTP operates on top of them, what DNS resolves for you. A good number of the outages I went through at LINE were ultimately solved by an understanding of the network layer.

Fourth, how databases work. Why an index makes reads fast and writes slow, what a transaction isolation level guarantees, how a query planner makes its decisions. You can only solve the real problem when you know what is happening behind the query an ORM generates for you.

Fifth, the ability to write clean code. The habit of naming well, splitting functions small, revealing intent, and protecting things with tests. This is a skill that transcends any particular language, and far from depreciating over time, its value only rises.

Do you see what this list has in common? None of it is tied to a particular company, a particular framework, or a particular trend. So once you build it properly, it follows you even when you switch companies or change languages. This is exactly why I spend the most time on fundamentals.

Of course, this is by no means a call to ignore the latest technology. It is a matter of proportion. I spend about half of my learning time on the slowly-aging fundamentals and the other half on the tools my current work needs. Atop the solid root of fundamentals, the leaves of the latest tools grow far faster.

To add one more thing, the more solid your fundamentals, the faster you learn new technology itself. Even seeing a new framework, you see the kernel: "ah, this is ultimately that data structure wrapped this way," "this concurrency model is ultimately a variation on locks." The gap between someone who memorizes from scratch every time and someone who just sticks a new label onto a principle they already know only widens as time passes.

So investing in fundamentals is not merely a conservative choice to preserve old things. It is in fact the most aggressive investment for absorbing the new fastest. I confirmed this with my own body while introducing several new technologies at LINE.

Designing a system to secure time — environment, not willpower

The point where self-directed learning most often collapses is, surprisingly, not willpower but time. More precisely, it collapses because we try to make time by leaning on willpower alone.

I once resolved every single day, "I will study after work today." The result was dismal. By the time work ended, my willpower was already at the bottom, and whatever remained had been spent by meetings and code reviews. Every night, only guilt piled up.

What I realized is simple. Willpower is volatile. After making decisions all day, the evening me is far weaker than the morning me. So you have to move learning to the time slot where willpower is strong, and lock that time in as an environment.

I moved my learning block to the morning, before work. I set my alarm forty minutes earlier than usual, and the night before, I leave the problem to solve or the chapter to read already open on my screen. When I wake up, I do not need to deliberate about what to do; I just sit down and begin. I drove the decision cost to zero.

The core of environment design is reducing friction. Make it easy to start and hard to get distracted. I leave my phone in another room and turn off messenger notifications during learning time. Just removing a few small frictions noticeably raised my execution rate.

One more thing: a sure thirty minutes beats a perfect one hour. If it looks like I cannot make a full hour, I end up not doing it at all, but thirty minutes carries no burden and so it keeps happening. I deliberately set the goal small. "One problem today." "One section today." A threshold low enough to clear every day carries you further than a high one you clear only occasionally.

To sum up: saying you have no time usually means you have no system. Do not blame your willpower; design the environment so that willpower is not needed.

What my weekly learning block actually looks like

So as not to end in abstraction, let me share a week I actually run. Not a perfect model, but a realistic version an ordinary working person can keep turning.

On weekday mornings I fix thirty minutes before work. Monday, Wednesday, Friday is one algorithm problem; Tuesday and Thursday is one section of a fundamentals book. Nailing the type to the day of the week means I never have to deliberate about "what should I do today."

The ten minutes at the tail end of lunch is a short review. I jot down what I saw that morning in a single line. This small note becomes the raw material for my weekend wrap-up.

On Saturday mornings I bundle what I learned that week into one short piece of writing. It is the time for output. And Sunday I leave empty. I play table tennis or just rest. Recovery, too, has to go on the schedule to actually be kept.

Across this whole table, no single day's burden is ever large. The key is not one big swing but a structure where small blocks connect without breaking. A plan small enough to be kept every day always beats a plan so grand it collapses in three days.

Skill that compounds — a year of thirty minutes a day

The biggest misunderstanding about self-directed learning is the belief that if results do not show up fast, it is not working. But skill accumulates not in a straight line but as compound interest.

Thirty minutes a day looks laughable. But over a year it adds up to more than about a hundred and eighty hours. That is nearly a month's worth of working hours spent solely on filling my own weaknesses. And it repeats every year.

Moreover, learning is not simple addition. What I learned yesterday becomes the footing for what I learn today. Knowing data structures lets me understand database indexes faster, knowing indexes makes query tuning easier, and that in turn leads to system design. Knowledge lifts itself up. This is compounding.

The trouble is that the early part of compounding is boring. The front of the curve is almost flat, so even a month of work shows nothing. Many people give up at exactly this flat stretch. They stop right before the curve turns steep.

So I count process instead of results. I check not "did my skill grow today?" but "did I do my thirty minutes today?" Results are outside my control, but process is inside my control. Counting only what you can control keeps the fuel going for a long time.

Concretely, I leave a small mark on the calendar. One box per day. After it runs for just a few days, a desire grows in me not to break that streak. It is a way of using the visible thing called continuity as motivation, in place of the invisible thing called skill.

And at the end of the quarter, I pull out the weakness map I wrote at the start of the quarter. By then, some of the items I could not handle have become things I can now explain. The moment of confirming that change with my own eyes was the greatest fuel for surviving the next quarter.

To put it in table-tennis terms: if you hit two hundred of the same forehand drives every day, a week shows no change. But after two or three months, at some point the ball flies differently. It had been changing while the body was not aware. Coding skill and language were exactly the same. Change always comes much later, all at once.

Self-direction that crosses fields — what I learned from studying languages

That self-directed learning is not a principle that works only for development — I felt this most strongly through studying languages.

Working at LINE, moments that require Japanese come often. At first I took the lectures the company supported. I attended, but my ability grew slowly. It was study that followed an assigned curriculum.

The turning point was the moment my own need became clear. On a day when I was frustrated at being unable to say what I wanted to in a meeting with a Japanese colleague, I started writing down and memorizing the expressions I had failed to use that day. Not the lecture curriculum but my own frustration became the textbook.

From then on the pace changed. I prepared in advance the expressions I would use in the next meeting, and afterward I filled in the parts I could not handle. Because learning was pinned to real combat, it was useful every time, and because it was useful, I did not forget it. I improved my English the same way.

The lessons I gained from studying English and Japanese applied to development exactly as they were. First, learning is fastest when it starts from my real need. Second, when you use what you learned immediately in real situations, that itself becomes review. Third, a little every day, even if small, beats an occasional binge.

Whether the field is code or language or table tennis, the principle was one. Learning that starts from a deficiency I felt, not a curriculum someone assigned, is what grows a person. So when I learn something new, I always ask first: where am I most frustrated right now? That is the place that will grow the fastest.

You only make it yours by outputting — the trap of input-centered learning

There is one thing in common among people who say they do self-directed learning yet never improve. They only input and never output.

They watch videos, read books, sit through lectures. In that moment it feels like they understood it all. But when actually told to do it themselves, their hands do not move. Between the feeling of having understood and actually being able to do it, there is a wide river.

I was stuck in this trap for a while too. I sat through dozens of hours of algorithm lectures, yet in front of a new problem I was at a loss. Watching and solving were completely different abilities. After realizing that, I cut down on lectures and increased the time I spent solving myself.

Output takes many forms. Writing code yourself, organizing what you learned into a short piece, explaining it to a colleague. Explaining to someone else is especially the most powerful output. When you try to explain, exactly where you do not know is revealed.

When I want to check whether I have truly understood something, I write it out imagining I am explaining it to someone who does not know it. The point where I get stuck is the hole. Only when I fill that hole does it finally become mine.

So when I lay out my learning time, I consciously balance input and output. If I spent an hour taking something in, I make or write at least as much myself. With input alone you can never cross the river.

The simplest output tool I use is short writing. It does not have to be a grand blog post. Something like "explain in three lines why an index is fast today" is enough.

When you try to fill three lines, the part you mistakenly thought you knew gets revealed at once. The process of filling that blank is the real learning. Input gives you the illusion of understanding; output reveals the truth of understanding.

These accumulated short pieces become an asset as a bonus. Months later, when I met the same topic again, there was no faster review material than the notes my past self had organized. Output is not for showing others; it is savings for my future self.

Going deeper 3 — I hold the wheel of my career

The bigger picture of self-directed learning is ownership of your career.

Many people think of career as something the company decides. You do the work the company assigns, take the training the company gives, follow the path the company sets.

But the company's priorities and my growth priorities are not always the same. The company assigns the work it needs right now; it does not take responsibility for my ten-years-from-now. In the end, the only person who will think most seriously about my career is me.

Cal Newport's "So Good They Can't Ignore You" criticizes the common advice "follow your passion," and instead urges you to build rare and valuable skills to create "career capital." Only once that capital accumulates, he says, do autonomy and choices finally arise.

The freedom to choose what to do is granted only atop sufficient skill. Self-directed learning is the surest way to build this capital.

Will Larson, who has long written about engineering leadership, often emphasizes that career growth is not chance but the accumulation of deliberate choices.

The point is that the person who chooses for themselves which projects to take, which technologies to invest in, and which seats to stand in ends up going furthest. The perspective that ownership is not a grand declaration but the sum of small choices helped me greatly.

At LINE too, when it came to deciding what work I would take on, the pace of my growth differed completely between the period when I just received whatever fell to me and the period when I consciously raised my hand to bring in work that could fill my weaknesses. At the same company, in the same seat, a single attitude toward choice made my one-year-later look different.

Holding the wheel is not something grand. It is deciding for yourself "this quarter I will dig deep into databases, my weak point," and moving toward that goal a little at a time, separate from company work. It is learning what you need even when the company does not assign it.

Skill built that way follows me even when I change companies, even when my role shifts. The company you leave; the skill stays.

Environment and job changes — handling them in a healthy way

Whenever I talk about self-directed learning, a question always comes up. "But what if the company environment is just terrible? What if there is nothing to learn?" A fair question. The environment certainly matters.

First, what to admit. Pinning everything on individual willpower is harsh. In an environment with no material for growth, no good colleagues, and no challenging problems, there is a limit no matter how self-directed you are.

Changing the environment — that is, a job change or a team move — is also a bona fide self-directed choice. Enduring blindly is not self-direction. The decision to leave a bad environment, if you made it yourself, is in itself a proactive act.

That said, using a job change as "escape" and using it as "strategy" are different. Moving on after learning everything there is to learn in your current seat, after filling your weak points, in order to reach the next stage — that is strategy.

By contrast, if you blame the environment and move every time it gets hard, every time you hit a wall because your fundamentals are weak, you will hit the same wall at whatever company you go to. The weakness I carry follows me even when I switch companies.

So I see it this way. Do make decisions to change the environment, but before that, honestly distinguish "is this the environment's fault, or my fundamentals' fault?" If it is the environment, moving is the right call; if it is on me, moving will not solve it. Keeping this distinction makes a job-change decision far healthier.

A conversation with a junior

Once, a junior colleague confided in me about agonizing over a job change. The conversation stuck with me, so let me transcribe it as I remember.

The junior said this. "There is nothing to learn here. The work I am assigned is just simple repetition, and I want to move." I asked, "Have you exhausted every way of doing that simple repetition better?"

The junior hesitated for a moment. "That... no." I asked again. "Have you ever tried automating that repetitive task? Have you ever traced the code history to see why it was written this way?" The junior shook their head.

What I meant was not "do not move." It was, even if you move, move after you have wrung out every bit of learning you can squeeze from your current seat. In the same environment, some people produce an automation tool, while others leave behind only complaints and depart.

A few months later the junior automated that simple repetition with a script, and said that the experience became the strongest story in the next company's interview. The time spent pausing the blame on the environment and looking inward became, in the end, the ticket to a better environment.

Whenever I recall this conversation, I think this. The question that distinguishes environment from self is not for others but for oneself. Only the person who asks it honestly avoids meeting the same wall twice.

A checklist to distinguish the environment's fault from mine

It is easy to say, but in the thick of hardship the distinction blurs. So I keep it laid out as a table and pull it out whenever I hit a wall.

| Question | Closer to the environment's fault | Closer to my fault |

| --- | --- | --- |

| Does the same difficulty repeat across several companies | No, only this company is like this | Yes, the same wall everywhere |

| Is there a channel to learn from good colleagues | The channel itself is blocked | The channel exists but I did not use it |

| Are challenging problems given | Only simple repetition is assigned | Challenges exist but I avoided them |

| Is there room to improve if I try | Structurally blocked | My effort has simply been lacking |

| Will leaving solve this problem | Yes, change the environment and it resolves | No, the weakness follows me |

If the marks cluster in the right column, there is a strong chance you will meet the same problem even after moving. In that case, filling the weakness first is the proper order. If the marks cluster in the left column, then it is the environment, not your will, that is the problem, and moving is the right call.

The real value of this checklist is not that it gives you the answer, but that it briefly pauses a judgment swept up by emotion. Throwing in a resignation while angry and deciding after answering five questions yield different results a year later.

Comparison — passive vs. proactive career

| Aspect | Passive career | Proactive career |

| --- | --- | --- |

| Learning motivation | Because the company assigned it | Because I felt a need |

| Coping with anxiety | Paralyzed by self-deprecation | Translated into a map of weak spots |

| Fundamentals | Surface only, chasing trends | Slow and solid |

| Job changes | Escape when it gets hard | Strategic choice |

| The wheel | The company holds it | I hold it |

| What remains | Vanishes when you leave the company | Follows you anywhere |

Looking at this table again, every row points in one direction in the end. Who holds the wheel. Learning motivation, coping with anxiety, judging a job change — all of it came down to whether or not I grip the steering wheel.

Comparing in a healthy way — growing while avoiding burnout

Self-directed learning has a shadow. Used wrongly, it becomes endless self-abuse, and endless self-abuse breaks a person down.

The most common trap is comparison. On social media there is always someone better than me. One is an open-source maintainer, another moved to a better company, another draws attention with a side project. Place them beside the present me and, no matter how hard I try, I always look lacking.

Comparison stimulates in the short term. But in the long run it gnaws away at the fuel. Because it has no end. Catch up to one person and another appears, and the bar of comparison keeps fleeing upward. Running toward an unreachable goal eventually exhausts you.

So I changed my comparison target. Not others, but yesterday's me. If I solved today a problem that yesterday's me could not, I have grown. This standard is reachable, wholly mine, and gives a small achievement every day.

About burnout I want to speak carefully. I am not a medical professional, and this is neither a diagnosis nor a prescription.

That said, one thing organizational psychologist Christina Maslach's burnout research often points out is worth remembering. Burnout comes not to lazy people but rather to those who immersed themselves too long, without recovery. In other words, it can mean that the hardest-working person may be the one most at risk.

The lesson this perspective gave me is simple. Rest is not laziness but a condition of continuity. To keep thirty minutes a day going for a year, paradoxically, you also have to design days of rest. I leave one day a week completely empty of learning and play table tennis or just rest. That one day protects the other six.

One more thing: I try not to use anxiety as my only fuel. Anxiety is good for starting the engine, but as fuel for a long run it burns a person up.

So I keep a positive goal alongside the anxiety. Not "because I might fall behind" but "because I want to build this" — I collect such reasons one by one. Using fear and aspiration together lets you run much longer.

How to practice — turning hunger into routine

1. **Secure time first**: Willpower evaporates. Fix the same time daily (say, thirty minutes before work) as a learning block. Environment beats willpower.

2. **Make a weakness map**: Each quarter, open three or four job postings and write down everything you do not know. That becomes your quarterly learning goal.

3. **Pick one strand of fundamentals**: Each quarter pick one area of fundamentals (data structures, DB, networks, and so on) and dig steadily. Trying everything at once finishes nothing.

4. **Marathon pace**: For LeetCode or a book, set a small daily portion and do it every day. A little every day lasts longer than weekend binges.

5. **Turn learning into output**: Leave what you learned as a short piece of writing or explain it to a colleague. Only by outputting does it become yours.

6. **Periodic review**: At quarter's end, revisit the weakness map and check what was filled and what remains. You have to see progress with your own eyes to keep the fuel going.

7. **Design rest days**: Leave one day a week empty of learning. Recovery is not laziness but an investment for continuity.

8. **Compare against yesterday's self**: Cut off comparison with others and measure progress only against yesterday's me.

A quarter-level example

| Point | Action |

| --- | --- |

| Start of quarter | Draft weakness map from job postings; pick one fundamentals area |

| Daily | 30 min of learning at a fixed time (one problem or one book section) |

| Weekly | Organize one thing learned that week into writing; rest one day |

| End of quarter | Update weakness map, review progress, set next quarter's goal |

The table is simple, but the key is the unit of a quarter. A day is too short for change to show, and a year is too long for the plan to stay sharp. Three months was, in my experience, the most fitting cadence — long enough to dig into one area properly while still confirming progress with your eyes.

Pitfalls — when you misread self-direction

A good principle, used wrongly, harms a person. Self-directed learning is no exception. Let me gather the traps I have fallen into myself or seen up close. Knowing them in advance makes them easier to avoid.

- **Degenerating into self-exploitation**: If self-direction becomes endless self-flogging, you break down in burnout. Maslach's burnout research warns of the danger of chronic overload without recovery. Rest is not laziness but a condition of continuity.

- **Using anxiety as the only fuel**: Anxiety can start the engine but is unfit as fuel for a long run. Running on anxiety alone wears you out. Also fill up with the fuel of a positive goal — a direction you want to go.

- **All the environment's fault, or all my fault**: Both are traps. Distinguish honestly. If it is the environment, change it; if it is on you, fill it.

- **Chasing trends only**: Chasing only new frameworks leaves fundamentals empty. Invest more in slowly aging fundamentals than in flashy new arrivals.

- **Fueling on comparison with others**: Comparison stimulates short-term but erodes you long-term. Make your comparison target yesterday's self.

- **Relying on willpower alone**: Willpower is weak. Replace it with systems like time blocks and environment design.

A first-quarter guide for those just starting

Even if all of this sounds good, it can be daunting to know where to actually start. So let me write out, for someone going from doing nothing at all, the order I recommend for spending the first three months.

In the first week, set just one thing: the time of your learning block. You do not yet have to decide what to study. Just build the habit of sitting at your desk at the same time every day. The first week's goal is not learning but attendance.

In the second week, open three or four job postings and make a weakness map. Then pick exactly one area from it as this quarter's topic. The key is to pick only one, without greed.

For the following ten weeks, dig into that one area a little each day. If it is algorithms, one problem a day; if it is a book, one section a day. Every Saturday, bundle what you learned that week into a short piece, and on Sunday, rest.

In the final week, pull out the weakness map you first made. Cross out the items you filled, and choose next quarter's topic. Turn this one cycle properly just once, and from then on your body remembers the rhythm.

Grand resolutions do not last three days. But this small cycle turns four times in a year. The me of a year later, having filled four weaknesses, becomes a completely different person from the me of last year who only repeated resolutions.

FAQ

**Q. Work alone overwhelms me — and now self-directed learning too?**

Do not set it grandly. Thirty minutes a day is enough. The key is not quantity but the persistence of "every day." Securing time as a fixed block is the first step.

**Q. My anxiety is so severe I cannot do anything.**

Write the anxiety on paper and convert it into concrete tasks. Once vague anxiety becomes a list, it becomes manageable. If you are too overwhelmed, start with the smallest single thing.

**Q. If the environment is truly bad, isn't a job change the answer?**

A job change is a legitimate self-directed choice. But before moving, honestly distinguish "is it the environment or my fundamentals?" If it is the environment, moving is the right call.

**Q. Isn't it better to learn practical tech instead of building fundamentals?**

Both are needed, but it is a matter of proportion. Practical tech ages fast; fundamentals age slowly. In the long run, fundamentals return at a higher compounding rate. I split my learning time in half.

**Q. I really cannot make even thirty minutes a day. This is genuinely an impossible stretch.**

Such a stretch comes for everyone. When it does, shrink the amount further. Ten minutes a day, even half of one problem, is fine. What matters is not breaking the chain. The difference between zero and ten minutes is bigger than the difference between ten minutes and one hour. As long as the thread of the habit stays alive, you ramp back up quickly when room opens.

**Q. Comparing myself with others makes me feel so pathetic. How do I stop?**

Move your comparison target to yesterday's self. Another person's output is the compressed result of their several years, but what is visible to me is only the final scene. The process is invisible. Comparing my present against an invisible process can only ever feel pathetic. Yesterday's me is the only comparison target whose process I fully know.

**Q. Even as I grow, the anxiety never disappears. Is that normal?**

Even now, with experience accumulated, I am the same. As skill grows, standards rise alongside, so the unknown areas actually come into sharper view. The anxiety not disappearing is closer to a signal that you have not stopped. Rather than trying to eliminate the anxiety, build the habit of translating it into a to-do list every time.

One-page summary — this post in five lines

If I distill only what I want to leave in your head after reading, it is this.

First, anxiety is not the enemy but directional information. Let it flow into self-deprecation and it is poison; translate it into a weakness map and it is a compass.

Second, what grew me was not study someone assigned but study I sought out myself. As Self-Determination Theory says, autonomous motivation stays the deepest.

Third, fundamentals age slowly. Data structures, operating systems, networks, databases, clean code. Invest more in the root than in the trend.

Fourth, it is system, not willpower. Lock time in as environment, do a little every day, confirm through output, and design days of rest.

Fifth, your comparison target is not others but yesterday's self. That is how you go far, without breaking down.

Closing — hunger is nothing to be ashamed of

Ever since that night I shrank before one LeetCode problem, I decided to treat anxiety differently.

Instead of blaming myself whenever anxious, I look at the direction that anxiety points to. I write down what I do not know, make it a quarterly goal, and fill it a little each day.

The anxiety still visits. Even now, with experience accumulated, it does. But now I know that it is a signal that there is still a place I want to reach. The day the anxiety disappears will probably be the day I stopped.

Hunger toward career is nothing to be ashamed of. The wish to get better, the wish to be recognized, the wish to do better work.

Let that hunger flow into self-deprecation and it becomes poison, but let it flow into self-directed learning and it becomes the strongest fuel. The same emotion can paralyze a person or carry them far. You only have to change the direction.

And skill built that way, no one can take from you. Even if I leave the company, even if my role changes, even if the market shakes, what I sought out and built myself follows me.

In the end, what grew me was not study someone assigned but study I sought out. Development, English, Japanese, even table tennis — all of it was the same.

If you are anxious today, write that anxiety on paper and turn it into a one-line task. It does not need to be grand. Just one line will do. That one line makes the you of a year from now.

References

- Deci, E. L., & Ryan, R. M. — Self-Determination Theory overview. [https://selfdeterminationtheory.org/theory/](https://selfdeterminationtheory.org/theory/)

- Carol Dweck — Growth Mindset, "The Power of Yet" TED. [https://www.ted.com/talks/carol_dweck_the_power_of_believing_that_you_can_improve](https://www.ted.com/talks/carol_dweck_the_power_of_believing_that_you_can_improve)

- Cal Newport — So Good They Can't Ignore You. [https://calnewport.com/so-good-they-cant-ignore-you/](https://calnewport.com/so-good-they-cant-ignore-you/)

- Will Larson (lethain) — writing on career and technical growth. [https://lethain.com/](https://lethain.com/)

- Maslach, C. — Burnout research overview. [https://pubmed.ncbi.nlm.nih.gov/27265691/](https://pubmed.ncbi.nlm.nih.gov/27265691/)

- HBR — Take Control of Your Career. [https://hbr.org/2021/01/take-control-of-your-career](https://hbr.org/2021/01/take-control-of-your-career)