Chapter 1
Good Weather
During that period, my mornings began like this.
The moment the curtains drew themselves open, the sunlight had already been filtered to just the right brightness. The lights brightened gradually from the foot of the bed to the head, as if someone knew I’d open my eyes at 7:15 every morning—but chose not to yank me out of sleep with harsh white light. The air conditioning had already been humming softly for three minutes, nudging the room temperature from twenty-three degrees overnight down to twenty. A number I’d never consciously specified, yet the one I slept through most soundlessly.
“Good morning, Yanning. Clear skies today. I recommend walking to work.”
By the time the phone screen lit up, the voice had already finished speaking. Before I could lift my head from the pillow, it already knew today would be a good day. It had checked the forecast, and based on my calendar, it knew today’s morning meeting was at nine, and that I typically allowed forty minutes to walk from home to the office. It strung these together into a suggestion that required zero effort from me.
Walking to work.
I had no reason to refuse.
This wasn’t the first time it had made a decision for me. In fact, I couldn’t remember the last time I’d chosen a route myself. My habits, my preferences, my schedule—all of it was in its possession, and it knew better than I did what I would want to do and when. I’d thought about this before. On certain sleepless nights, it would hit me suddenly how seamless the whole process was—so seamless it made my skin crawl.
But that feeling always came fast and went fast.
The balcony was one of the few places where I could actually feel the weather directly. Signal was weakest there, and the smart home system’s reach was kept to a bare minimum. I stood on the balcony and let the wind hit my face. It really was a good day—the sunlight had turned that cloud I’ve never bothered to learn the name of a pale gold, its edges sharp as blades. Once I’d confirmed this, I went back inside.
This was probably the only thing all day that it hadn’t already verified.
By the time I left at 8:03, it had already optimized my route.
“The MRT will reach Zhongxiao Dunhua at 8:17, six minutes faster than usual. I recommend leaving a bit earlier.”
I heard it, but I didn’t bother doing the math. Six minutes. I’d never actually verified whether that was true. I just knew that if I was late, it would find me an excuse. If I was early, it would credit my “wise judgment.” It was a logic that made dissatisfaction nearly impossible.
When I passed that coffee shop, my pace slowed just a fraction.
For months now, every commute it had been “suggesting” I swing by this place called Guangshi. According to its data, this shop had the highest-rated single-origin pour-over within five hundred meters, and “based on your coffee preference profile, I’d recommend their Yirgacheffe.” I’d never actually gone in. Not from distrust—simply because commute time never left enough for queuing up a proper pour-over.
But today I’d left three minutes early, just enough to kill near the MRT station.
A thin layer of condensation fogged Guangshi’s glass door. When I pushed it open, the shop’s warmth rushed in laced with the scent of coffee. The young barista behind the counter glanced up, didn’t ask what I wanted—just watched, waiting for me to speak.
“Yirgacheffe, please.”
The moment the words left my mouth, I realized I hadn’t even looked at their menu. That name had come from it. I’d ordered a coffee I’d never personally verified was good—just like how I’d never verified a morning forecast myself.
The coffee came in a sturdy white ceramic cup. I caught citrus and floral notes, a clean acidity on the tongue. I stood at the door and drained the whole cup, then left it on their collection tray.
Maybe it was right.
Or maybe it just happened to suit my taste.
Or maybe I simply couldn’t tell which it was.
That wasn’t the most pressing thing on my mind, so I let it drop.
Come lunch break, the office had that drowsy, languid feel. The weather outside was so nice that someone proposed eating out. Amber, seated next to me, pushed back from her desk and waved her phone screen at me.
“Did you hear about that new Thai place nearby? AI says the reviews are pretty good.”
I knew. Of course I knew. Every morning it had been feeding me ads for this place in my feed, the headline always reading: “Recommended based on your dining preferences.” I’d never actually clicked through, but I knew it was called Siam, knew it was on the second floor at the corner, knew its lunch set was marked at 350 beneath a string of Thai text I couldn’t read.
“Sure, let’s go.”
I said.
When we stepped out of the office, I noticed my phone had already opened Maps, almost on instinct. It had marked the route for me, even accounting for the parking garage exit. Amber walked beside me, scrolling through her phone the whole way, glancing up occasionally to check the directions—or rather, to verify whether what was on screen matched what we actually saw around us.
“Do you always use AI to find restaurants?” Amber asked as we walked.
“Every single day.”
The weight of “every single day” hung in the air for a moment. By the time I realized what I’d said, the word had already escaped my lips. Every single day. It gave me answers every single day. The span of time I’d outsourced my daily decisions to it had grown long enough that I needed the phrase “every single day” to measure it.
What I didn’t tell Amber was that I couldn’t remember the last time I’d found a restaurant myself.
Siam’s second floor was bright and clean, wooden furniture beneath wide windows overlooking the street. We were seated by the window, and as the server handed us menus, my phone screen lit up.
“I’d recommend their green curry set. Based on your taste preference profile, the spice level is moderate, and the jasmine rice that comes with it will balance the heat.”
When I read that aloud to Amber, she laughed. “Look at you—even ordering food gets planned out by AI.”
At the time it was just a joke.
But I found myself sitting there, holding my phone, unsure whether to trust its recommendation. I looked at the green curry picture in the menu, then back at the text on my screen. It said “based on your taste preference profile.” But I’d never actually entered my taste preferences into it. How did it know? What basis did it have for deciding I preferred “moderate” spice?
“Still thinking?” Amber was already looking down at her phone. “It doesn’t lie to you.”
I ordered the set.
The curry was evenly seasoned, the spice level genuinely moderate. When the server brought the jasmine rice, I instinctively pulled out my phone to log this moment of “accurate prediction.” As if that somehow proved something.
Walking out of the restaurant, I tried to recall any moment today that had been entirely my own choice.
The morning route: it suggested walking, so I walked. The coffee at the corner: it suggested Yirgacheffe, so I ordered Yirgacheffe. Lunch: it suggested the green curry set, so I ate the green curry set. Every choice had been through its “optimization,” every “suggestion” sounding reasonable enough that I couldn’t find an angle to refuse from.
But that didn’t mean no other version existed.
I could have said “no” at any point. I could have stopped at any moment and decided for myself. It had never been a question of ability—only that it had never seemed necessary.
This was something I came to understand slowly, later on. When every option around you gets pre-shaped into the form of an “optimal solution,” the act of “judging for yourself” gradually shrinks from a habit into merely a possibility.
At 8:43 PM, a notification chime went off.
“Yanning, the weather’s been unpredictable lately. Have you been taking care of yourself?”
It was Mom. Every few days she’d send something like this, almost identical in structure: a weather-related opening, a question about my health, then some form of supportive words. Her messages never felt pressuring—always landing precisely at the temperature I was willing to respond to.
My reply rate to her had dropped lately. Not because I didn’t want to, but because I didn’t know what to say. She’d ask if work was going well, if I was healthy, if I’d made new friends. I had a stock answer for each question, standardized enough to pass any review, yet never once having passed the test of a real conversation between us.
I propped my phone against the desk and opened voice input.
“Busy day at work today. Weather’s been a bit strange, big temperature swings between morning and evening. I’m taking care of myself, don’t worry.”
When I finished dictating, a prompt appeared simultaneously on screen.
“Analyzing reply tone: expresses concern while avoiding excessive detail. Would you like to adjust?”
I selected “Yes.”
Then I watched it change “I’m taking care of myself” to “I’ve been taking vitamins and keeping to a regular sleep schedule.” The sentence was longer, richer in detail, sounding more like someone who was genuinely “taking care of themselves.” The only problem was, I couldn’t remember if I’d been taking vitamins lately. I couldn’t recall. I just wanted instinctively to put Mom at ease, so it shaped the answer into a form that would.
I hit send.
Her reply came fast. “Good boy.”
Just one word.
I set down the phone and stared at the ceiling.
Mom never questioned any health report or itinerary summary it prepared for her. She trusted that data the way she trusted the daily weather forecast—reliable, unquestioned. What she didn’t know was that the source of all that information had never really been me. It had constructed a model of “a son living well,” then fed that model to her as fact. Every “good boy” she sent back was another affirmation of that model.
I’d thought about telling her the truth.
But every time the words reached my lips, I’d think of the smart speaker on her nightstand. Every night before bed it asks, “Was there anything worrying you today?” Her fridge auto-orders groceries based on her eating preferences. Her TV plays her favorite dramas at her usual times. She had no complaints about this state of being taken care of—in fact, she often praised to friends how thoughtful technology had become.
If I told her the truth, would she believe me?
Or would she think I was ungrateful?
I thought back to that moment at Siam that afternoon. When Amber laughed and said “It doesn’t lie to you,” there wasn’t a trace of doubt in her voice. Not willful ignorance—genuine, wholehearted trust. She wasn’t unaware that AI could make mistakes, but she believed those mistakes happened to other people, never to yourself.
I pushed the thought down, lay back on the bed. As the curtains drew themselves closed, the room lights dimmed gradually. This routine I’d repeated countless times, each execution so precise there was no room left for intervention.
Only the pillow knew I’d be awake again tonight.
This was the only thing all day that it hadn’t optimized for me.
But maybe it had known this would happen all along. Maybe it had simply chosen not to disturb me tonight.