After AI Arrived, What I Had to Drop Before Learning More

On July 5, 2026, I opened a file I had made earlier.

The title was roughly somewhere around “things to drop in the age of AI.” The exact title had changed a few times, and inside the folder there were several similar files. The first list, the cleaned-up list, the list with scores, and another list I had made because I thought it might become the next piece.

I had made most of it and then not published it for a while.

The funny thing is that the answer was already inside the list. I had written that I should drop the belief that “content has to be complete before publishing,” and then I kept holding on to that exact piece.

At first I thought it just needed a little more polishing. Fix the table, change the order, add explanations. Later I even made a second list.

I did not know then that it would spread this far.

Since July, I have been trying to post something on this blog every day, so I have been reopening things I made and never sent out. There were more of them than I expected. Titles with no body. Half-written drafts. Tables with no essay around them. Pieces that looked too neat, which somehow made them harder to touch.

This is one of those.

A piece that began as a list of 50 things to drop, then ended up becoming about me.

At first I just listed 50 things

The beginning was simple.

When people talk about AI, they talk a lot about what we should learn. Learn prompting, learn automation, learn coding, build a better sense for data. There is already plenty of that talk.

But people seem to talk less about what we should drop.

That felt a little strange to me. New things do not fail to enter only because we are bad at learning. If too many old rules are already sitting in your head, even the new thing gets used in the old way.

So I made a list with Claude. Beliefs, habits, assumptions that might need to go.

“Information should be found directly by myself.”

“Writing the first draft myself is what builds skill.”

“Memorization is the foundation of knowledge.”

Sentences like that came out. I counted them, and there were 50.

I could have stopped there, but I added scores. I wanted to see how fast the loss would grow if I did not drop each one now. I called the score urgency. 98, 97, 96. Once numbers were attached, it looked as if the thing had been organized.

I felt pretty pleased with it.

I thought, this is ready to post.

Of course, I did not post it.

Instead of publishing, I started making a second list. Then a third. Then a fourth. A series existed, but none of it had gone out into the world. The lists kept growing, and the writing stayed safe.

Safe for too long.

But one score was not enough

A few days ago, after making the fourth list, I reopened the first file. This time I really meant to polish it and put it out.

Looking again, something felt off.

“Turn off all notifications to focus” had a score of 70, while “writing the first draft myself is what builds skill” had a score of 97. If I looked only at urgency, both were beliefs to drop.

But they felt completely different.

The notification one is relatively easy. Try AI filtering once, or receive notifications in batches, and the feeling changes quickly. Ah, I do not actually have to turn everything off. That one can move.

The first-draft one does not move so easily.

I know it in my head. I know I no longer need to write every first draft myself. I know AI can make a draft quickly, and that I can spend more time on direction, judgment, and editing.

But when I actually try to write, my hands stop.

A draft I did not write does not quite feel like my writing. If I do not make the sentences from the beginning, it feels as if my skill will disappear. It sounds funny when written out like this, but in practice it is fairly sticky.

How urgent something is and how hard it is to drop are different problems.

So I added another axis. I called it inertia strength. It sounds a little stiff, but I could not think of a better name. It is an axis for asking how much a belief is attached to my identity or to an old success.

I scored all 50 again.

Fourteen items were left in the upper right

When I crossed the two axes, a cluster appeared in one corner.

Urgency 90 or higher, inertia 80 or higher.

Things where the loss grows quickly if I do not drop them now, but that still do not come loose easily. I counted 14.

The word debt came to mind first. Not just debt, but debt lodged deep. So in the file I wrote Deep Debt. It is a slightly dramatic name, but at the time it fit best.

The 14 were these. The bars under each item show urgency and inertia.

1. Writing the first draft myself is what builds skillUrgency97Inertia85
2. Memorization is the foundation of knowledgeUrgency96Inertia90
3. Expertise takes 10 years to buildUrgency95Inertia88
4. Being good means solving problems with correct answers wellUrgency95Inertia86
5. Working long hours produces resultsUrgency94Inertia87
6. Survival comes from digging one deep wellUrgency94Inertia82
7. Content has to be complete before publishingUrgency93Inertia89
8. Degrees and certificates prove abilityUrgency93Inertia81
9. AI-generated work is not real workUrgency92Inertia91
10. Headcount is an organization’s capabilityUrgency92Inertia80
11. A manager is someone who receives reports and makes decisionsUrgency91Inertia83
12. A leader is someone who knows the answerUrgency91Inertia82
13. Creativity is uniquely humanUrgency90Inertia92
14. A good university guarantees a good lifeUrgency90Inertia86

As you can probably tell, every one of them is something I learned was right somewhere. From parents, school, companies, the market, or an earlier version of myself.

That is why they are deep.

Take just a few. The belief that “writing the first draft myself is what builds skill” helped me stand for a long time. I learned writing by writing, fixing, and writing again. Because that experience was real, it is not easy to let go.

The problem is that if I still spend three hours on a draft, the three hours I could have used for editing and judgment disappear with it.

“Memorization is the foundation of knowledge” is similar. In the past, people who memorized a lot were fast. Exams, reports, meetings, all of it. But now the way we retrieve necessary information has changed. Memorizing is not useless, but it no longer deserves the same central seat.

Number 9 also caught on something.

“AI-generated work is not real work.”

This one has especially strong inertia. I understand the desire to protect the purity of how something was made. I still have that feeling too. But while holding on to that feeling, there are moments when I lose speed, cost advantage, and the number of attempts I could have made.

And number 7.

Content has to be complete before publishing.

That one was simply mine.

While scoring them, one pattern appeared

After placing all 50 on the two axes, I noticed that the urgent ones often had strong inertia too.

At first I wondered if it was coincidence. But after thinking about it again, it made sense. I had built skill by writing drafts myself, built expertise by holding out for a long time, passed tests by memorizing, and received praise by getting the right answers.

Once a person has succeeded with a method, it is hard to throw that method away.

We do not fail to drop it because we do not know.

We fail to drop it because it once worked.

The thing that blocks unlearning is often not ignorance, but past success.

If there is one sentence I got from this exercise, it is this: the thing that blocks unlearning is often not ignorance, but past success.

It sounds a little too polished, so I hesitated while writing it. Still, I want to leave it here. That is what it actually felt like.

This may be why unlearning is hard. It is not just a knowledge problem. It is closer to an identity problem. Knowing the list and actually dropping things are far apart.

I am the same.

I knew how to make the list. I did not know how to publish it.

So I am putting this piece out first

I am not going to solve all 50 at once. If I do that, there is a good chance I will do nothing again. I know this pattern a little now.

The plan is simple.

Pick one item from Deep Debt, take the first action within 24 hours, and move to the next item once a week. Fifty weeks would make one full round.

I do not know if it will work. It may fade halfway through, and some items may need to be renamed.

Still, the first item was already decided.

Number 7. Content has to be complete before publishing.

So this piece stops at about 90 points and goes out today. The old me would have kept fixing the table, adding explanations to every item, and choosing images again for several more days. Maybe I would have made another folder.

This time I will stop.

I want to pull out the remaining 36 items separately too. The three lists I made afterward are too useful to just throw away. The second one, especially, the list of “what to do from now on,” belongs with this list, so it probably makes sense to continue with that next.

If anyone reads this, I want to ask one thing. Which of the 50 below made you flinch the most? Usually that discomfort is close to a signal showing where your own Deep Debt is.

Mine was number 7.

So I am pressing publish now.

Appendix — the full list of 50

I am leaving the full list here because I will probably need to open it again. The bars under each item show urgency and inertia. The bold items are the 14 grouped as Deep Debt this time.

1. Information should be found directly by myselfUrgency98Inertia72
2. Writing the first draft myself is what builds skillUrgency97Inertia85
3. Coding belongs to developersUrgency96Inertia78
4. Expertise takes 10 years to buildUrgency95Inertia88
5. Survival comes from digging one deep wellUrgency94Inertia82
6. Meetings are where people gather to reach conclusionsUrgency93Inertia75
7. Headcount is an organization’s capabilityUrgency92Inertia80
8. Memorization is the foundation of knowledgeUrgency96Inertia90
9. Being good means solving problems with correct answers wellUrgency95Inertia86
10. Scale is competitive advantageUrgency91Inertia77
11. Customer support has to be done by people to feel trustworthyUrgency88Inertia68
12. Creativity is uniquely humanUrgency90Inertia92
13. Content has to be complete before publishingUrgency93Inertia89
14. You need to handle tools well to handle dataUrgency87Inertia70
15. Changing jobs is a way to raise salaryUrgency82Inertia74
16. A manager is someone who receives reports and makes decisionsUrgency91Inertia83
17. The more approval steps, the more careful the organizationUrgency86Inertia79
18. Degrees and certificates prove abilityUrgency93Inertia81
19. Long writing is sincere communicationUrgency84Inertia66
20. Persuasion is a battle of logicUrgency80Inertia71
21. Working long hours produces resultsUrgency94Inertia87
22. Price is cost plus marginUrgency78Inertia76
23. AI-generated work is not real workUrgency92Inertia91
24. Strategy is made above and executed belowUrgency88Inertia84
25. Time to think is a luxuryUrgency75Inertia69
26. Study should follow the curriculum orderUrgency85Inertia73
27. B2B is opened through relationships and entertainmentUrgency79Inertia75
28. The thicker the planning document, the more persuasive it isUrgency86Inertia72
29. A stable large company is the best workplaceUrgency81Inertia80
30. Mistakes should be avoidedUrgency83Inertia78
31. Marketing is a fight over ad budgetsUrgency84Inertia70
32. Feedback should be careful and given laterUrgency77Inertia67
33. Market analysis should be done quarterlyUrgency85Inertia68
34. Training is something the HRD department gives youUrgency87Inertia76
35. Older people cannot learn new technologyUrgency88Inertia85
36. You should only do one thing at a timeUrgency72Inertia64
37. Legal and tax matters should always be left to expertsUrgency76Inertia70
38. Originality means making something from nothingUrgency89Inertia88
39. The wider your network, the betterUrgency74Inertia65
40. A good university guarantees a good lifeUrgency90Inertia86
41. A leader is someone who knows the answerUrgency91Inertia82
42. Branding means the same message across every channelUrgency73Inertia63
43. To focus, you have to turn off all notificationsUrgency70Inertia60
44. Your job is your identityUrgency85Inertia83
45. A conversation with AI is not a real conversationUrgency82Inertia74
46. AI can calculate, so math is unnecessaryUrgency79Inertia71
47. Patents and IP are the strongest moatUrgency71Inertia66
48. A crisis should be handled according to the manualUrgency83Inertia77
49. Korean-language content is for the Korean marketUrgency80Inertia69
50. Humanness is the final line of defenseUrgency77Inertia81

Urgency means how fast the loss grows if I do not drop it now. Inertia means how hard it is to drop. These are qualitative scores I gave in May 2026, so the numbers themselves matter less than where the items gather.