Most of the work was already done.
And yet, I still did not move.
I kept casually checking my phone, let the night pass, and only started working around 1 a.m., once I was already sleepy. It was the behavioral equivalent of pulling an all-nighter and then opening the textbook: “All right, time to study.” The order was clearly broken.
The strange part was that once I started, I could keep going. The remaining work was not huge. Put things in boxes. Get them into a state a mover could handle. For writing, do the final proofreading and decisions. The problem was less about ability and more about the first step feeling absurdly heavy.
That night made one pattern very clear: deadline pressure can be a genuine strength, but the same mechanism means that by the time the engine finally starts, there may be almost no fuel left.
1. At 1 a.m., the human suddenly booted up
It was not that nothing had been done before midnight.
Most of the preparation was already finished. Rationally, there was only “a little more” left.
But as long as there was time, the phone kept winning. I did not have to do it now. Five minutes later was still possible. A little later was still possible.
Then it became 1 a.m., and sleepiness arrived.
If I keep doing this, I will go to bed without finishing anything.
Suddenly, movement became possible.
It did not feel like motivation falling from the sky. It felt more like the option to postpone had disappeared.
There was no excellent supervisor inside me. The clock simply appointed itself supervisor.
2. The final few percent can be the hardest part to start
The funny thing is that there was no giant task left.
Most of it was finished. What remained was packing, gathering, checking, and handing things over. Each individual action was small.
Still, starting was difficult.
That shows why “less work” does not always mean “easier to begin.”
Even a tiny task can feel heavy if it requires a context switch: put down the phone, stand up, pick something up, decide where it goes, look at the next object.
Here, I will call the friction created by switching, deciding, and beginning before the actual work even starts the “initiation cost.”
A task that takes five minutes can sit untouched for an hour if the initiation cost is high.
Then the five-minute task wakes up in the middle of the night.
An unfinished task with a fully nocturnal lifestyle.
3. A deadline may not create motivation; it may destroy the barrier to starting
Right before a deadline, the number of choices collapses.
Earlier, the branches were “do it now,” “do it later,” “do it tomorrow,” or “do something else first.” Near the deadline, the tree is brutally pruned down to “do it now.”
That reduces the decisions needed before starting.
So concentration can suddenly appear.
This is useful. Deadline-powered bursts can absolutely get things finished.
But it may be misleading to say that the deadline increases ability.
In this case, a better description is that the escape routes and alternatives vanished, which made starting easier.
A deadline is less like a motivation injection and more like a forced boot button.
It works when pressed. The problem is that it often gets pressed far too late.
4. Deadline turbo is powerful, but the fuel tank may already be empty
This is the painful part.
I can move when the deadline is close. That means things often get finished in the end.
But by the time I start moving, I am sleepy.
Attention and physical energy are lower than they were earlier in the day. Startup performance is at maximum while the battery icon is red.
The engine started. There is no gas.
That is the “study only after pulling an all-nighter” operating mode.
If you measure only short-term completion, this system can look surprisingly good. The task does get done.
But once comfort, sleep, and next-day capacity are included, the price becomes obvious.
Deadline turbo is excellent emergency equipment. Using it as the normal engine wears out the human component first.
5. Outsource bulk execution and keep the final judgment
Trying to fix the entire tendency at once makes the problem unnecessarily heavy.
“Become someone who always starts immediately” or “just follow the plan calmly” sounds nice. If that were easy, there would be no 1 a.m. boot sequence.
A more useful move was division of labor.
Move as much repetitive execution as possible to another person or system.
Keep the final decisions that are hard to substitute.
For writing, someone else can produce the bulk draft. I can then read it and decide: this is wrong, keep this, rewrite that.
For moving, the physical transport and final packing steps can go to the mover when the service allows it. I keep the decisions about what goes and what stays.
That is not laziness.
It is allocating a scarce resource—my own attention—to judgments that are difficult to replace.
6. For moving, “ready for someone else to continue” can beat “perfectly packed”
When tired, it is strangely tempting to complete every last detail personally.
Close every box neatly. Tape everything. Categorize perfectly. Produce a finished artifact with your own hands.
But if the mover can do those steps, the objective changes.
The real target becomes getting the situation into a state where the next person can continue the work.
If “this is what is inside; please close and tape it” is enough, that may be enough.
There is little value in hoarding low-value steps.
Once the goal changes, the finish line moves dramatically closer.
“Pack everything perfectly” is far away.
“Put it in a box and make it handoff-ready” is close.
At that point it is barely moving prep anymore.
It is preprocessing for a human API.
7. Writing worked the same way: delegate drafting, keep proofreading and decisions
The same division appeared in article work.
Proofreading, noticing something feels wrong, and making editorial decisions can continue for a long time.
Generating a large amount of prose from a blank page is different. The startup cost is much higher.
So the bulk drafting can go to someone else while the final proofreading, revisions, and accept-or-reject decisions stay with me.
That split turned out to be highly practical.
The important question is not “Did I personally do every step?”
It is “Did my judgment enter at the places where my judgment was actually necessary?”
There is no law saying mass generation and final judgment must be performed by the same person.
If an irreplaceable final task will always remain, it makes sense not to spend all the fuel before reaching it.
8. The fatal weakness of artificial intelligence: it does not give me two extra arms
Late at night, what I eventually wanted was not more advice.
“I need someone from the AI side to come over here and pack these boxes.”
That was the actual request.
A chat system can help make decisions. It can order priorities. It can help choose what to discard.
But objects on the floor do not teleport into boxes through a chat window.
Intelligence can be added. Two extra arms cannot.
At that point, the quality threshold drops even further.
Chiikawa is fine. Just put it in the box.
No advanced reasoning required.
Pick up. Put in. Next.
In that moment, the lack of physical automation feels more urgent than the future of artificial intelligence.
Standing in front of a cutting-edge language model, the human still wants a box-packing assistant.
9. Conclusion: design the startup conditions and division of labor instead of trying to become a different person
The night revealed a simple pattern.
Deadline-powered bursts are genuinely useful.
But by the time they activate, the human battery may already be nearly empty.
So “I can always do it at the last minute” is not a complete success metric. It can mean entering every boss fight at five percent battery.
A more practical strategy is to redesign the work instead of trying to transform yourself into a permanently prompt, perfectly disciplined person.
Outsource bulk execution. Give physical tasks to people who can physically do them. Delegate mass drafting. Keep the decisions, selections, and final checks that truly require you.
And for the last five percent, the goal does not always need to be “finish everything perfectly.” Sometimes “make it possible for the next person to continue” is enough.
If none of that helps at 1 a.m., the remaining desire is extremely simple.
Call Chiikawa.
Put it in the box.
That is all.
