In one message a man said he was very proud of his team for perhaps the most impressive scientific and technical breakthrough of recent decades, and then, in the next breath, said he was prouder of something smaller. His son, born early, had learned how to eat on his own.
I keep coming back to that message, and I still cannot decide what it is doing to me.
Part of it is the plain tenderness, which needs no help from me. The stranger part took longer to name. He runs one of the companies at the center of the software world, a world that has spent two decades learning to watch systems from the outside: to turn the hidden inside of a machine into numbers on a screen, so that when something goes wrong you can find out why. And the thing that displaced a career's worth of pride was the least observable event I can think of. A mouth learning to work.
For a baby born too soon, learning to eat is not a metaphor. It is the coordination of three things that must happen in the right order, many times a minute: suck, swallow, breathe. A swallow interrupts the airway for something between a third and two thirds of a second, in a chest built to breathe forty to sixty times a minute. Get the timing wrong and the baby stalls, the oxygen dips, the heart slows. Nurses watch for it. There is no dashboard that does it for you. You wait, and you count, and one day the small machine of the mouth finds its rhythm, and a father decides that is the most impressive engineering he has seen all year.
He was honest about the reflex, more honest than most of us would be. Describing the first hour after the birth, he said he felt himself fall in love and narrated it while it happened: he was being neurochemically hacked, he thought, and he could notice it happening, and he was fine with it. Even the love arrived as telemetry: he said he could watch it happen even as it happened. I do not say this to mock him. I say it because I recognize it, and because many software parents think exactly like this all day. His husband, by public account, is a software engineer. The world the two of them work in runs on watching systems from the outside.
There is a whole profession trained to meet uncertainty this way. When a large online system might be failing, you do not stare at it and hope. You watch four numbers: how slow it is, how much traffic it has, how many errors it throws, how full it is. You set alerts. You keep people on call, because someone has to be woken. When something breaks you open an incident, name a commander, keep a log, and afterward write down what happened so it cannot happen the same way twice. It is a genuinely good way to run a system that has a right answer. Its practitioners know the costs, too: waking a human is expensive, and if you page people for every flutter, they learn to skim the alerts and then to ignore them.
Now walk that mind into a nursery in 2026, where you can buy the crib with the monitor built into the furniture.
The danger for a parent like this was never that he would hand the baby to a machine, or that he would care too little. It is that he would care in the only grammar his work had taught him: watch everything, alert on the exceptions, treat each cry as an incident with a root cause waiting to be found. Love starts to feel like being on call.
Love starts to feel like being on call.
Almost every one of these instruments is sold as a way to make a frightened parent calmer. The honest evidence keeps pointing the other way.
Years ago a mother in County Kilkenny logged her ten-week-old daughter into an app: which breast, how many minutes, the contents of each nappy, every stretch of sleep, every dose of medicine, her weight as it climbed. She had started tracking to feel less anxious. The volume of information made her more anxious. Her confidence came back, she said, only after she stopped.
The easy explanations do not hold up. The first is that the products are bad. Mostly they are not. Many of them log, track, and alert in the limited ways they advertise; how accurate that is, and how much it means, varies from one product to the next. The trouble is older and larger than any one gadget. In hospital intensive-care units, where the monitors are medical-grade and the people reading them are trained, keeping the alarms meaningful is a discipline of its own, with whole studies devoted to cutting the needless ones down. If trained professionals have to work that hard to keep the alarms honest, the tired parent on the bedroom floor is badly outmatched.
The second is that these parents are anxious people, or lazy ones handing the baby to a machine. That gets it backwards. They are not caring less. They are caring more, and in a particular direction. After a health scare, one mother found herself checking the video monitor every time the baby had been quiet for fifteen minutes, then zooming in, again and again, to make sure she could see the small chest rise. That is vigilance with nowhere to rest.
What is left, once you clear those away, is harder to fix, because it is not in the devices. It is in the reflex.
A newborn cannot be made observable. That is the whole trouble. When a server misbehaves there is, in principle, a cause you can find if you gather enough signals. A newborn's cry has no such stable root. It can mean hunger, a wet diaper, gas, tiredness, too much noise, illness, or nothing you will ever name. Healthy babies cry between one and four hours a day. About one in five develop colic, which the books define, more or less, as crying that no one can explain. There is no query you can run against that.
And the ground you are trying to measure will not hold still. Healthy infants have occasional drops in oxygen below eighty percent that come to nothing. Put a sock on that foot that alarms when the number falls, and you have built an alarm that can fire on a well baby. From there the cascade is easy to picture, and the doctors who study these devices have described it: a bit of motion or a dip that means nothing sets the alarm off, and a healthy baby can end up in an emergency room with blood tests, an x-ray, a night admitted, because a number said something the body never did.
None of it buys the safety the parents actually want. Sharing a room with the baby, but not a bed, lowers the risk of sudden infant death syndrome by as much as half. The home monitors show no such benefit; the pediatricians' own guidance notes, almost in passing, that in one study the parents who used them reported more depression, not less. About thirty-five hundred American babies still die suddenly in their sleep each year, and the machine on the sock is not the thing standing between a family and that number. The baby on its back, a firm flat surface, a crib with nothing soft in it: those do more than the dashboard does. The dashboard mostly watches you.
I want to be fair to the other side, because it is stronger than the gadget skeptics allow, and because I feel its pull.
A feeding log is not vanity when it is shared memory between two exhausted adults, when it answers the question of when she last ate without waking the other to ask. Sleep does strange things to remembering; one parent wrote that she could forget whether she had fed the baby "quite literally two seconds" after doing it. For a baby with a real cardiac or breathing problem, a monitor with limits a doctor set, on a baby a doctor chose to watch, is medicine, brought home so the family can sleep in its own beds. And there are hours when the house is least social and the need is sharpest, when there is no grandmother down the hall and no clinic open, and a tool is better than the dark. One of the Japanese cry apps reports that it is especially popular in the middle of the night, when it is hardest to ask anyone for help. With a different app, a mother said it helped her when she could not work out why her son was crying, his stomach full of gas. The kind of ordinary trouble that a more practiced hand might one day learn to read from across the room.
Stay with that second app a moment, though.
A young father, the one who built it, held his phone up to his own three-month-old, and a few seconds later the screen said I'm hungry, with a little gauge showing how sure it was. Three hours had passed since the morning feed. The baby was fed, and fell asleep in his mother's arms. Everything about it worked. And a machine had just named, out loud, a thing the room itself was in the middle of teaching him. Three hours since a feed and a certain cry: that is a fact a parent's own hands learn, in the only way it can be learned, which is by getting it wrong and then right. The gauge does not shorten that apprenticeship. It offers to skip it. A kinder tool would teach the pattern, then let you need it less.
A log that remembers what you cannot is one thing. A gauge that reads the baby for you, so you never quite have to learn to, is another. That is the difference the whole question turns on, and it is thin. The two often look identical, and are sometimes the same device an hour apart. The first night home with their daughter, a couple kept the app running loud on a tablet in their room. When she cried, they heard her first, and then heard the cry again, an instant later, as an echo through the app. On the second night the app threw a red alert, and they ran to find her fine and breathing. They suspected the swaddle and the breathing band had made the reading hard. The monitor had not replaced the room. It had added a second, louder alarm to it, and then been wrong.
The reflex to instrument is not coldness. It is what competence does when it is afraid. If you have spent years being the person who can fix the thing, being handed a small creature you cannot fix is its own kind of vertigo. A dashboard is a place to put your hands. Logging the feed, checking the number, reading the sleep report over coffee: these are things you can do well, and doing them well feels like caring, and it is so close to caring that you can miss the gap for months. The gap is the part that cannot be done well, only done. Sitting in the dark, uncertain, awake, holding a baby who needs no report run and no fault found, only to be held by someone who has stopped trying to make her legible.
I keep going back to the father, and to that one sentence.
The thing that beat his life's work was the one thing no instrument could have delivered. Not a number he drove up and to the right. A mouth learning to time a swallow against a breath, a fraction of a second at a time, on no schedule anyone could set. He, of all people, comes from the world that watches systems from the outside, and this was a system that would not open to that kind of watching. It could only be waited on. And he said that was the thing he was proudest of.
And in the same season, he told other new parents which crib to buy, and the one he named has the monitor built in, and the small company that made it watched its traffic jump by a third within a day.
I do not think he is confused. I think both things are true in him at once, the way they are becoming true in more of us: the same person who marveled that a mouth learning to eat had outrun his life's work, reaching, all the same, for one more thing that watches. In the first photograph a newborn hand had closed around an adult finger, and there is no number for that either. We are going to spend the next few years deciding, one dark hour at a time, whether we can bear to let some things stay unmeasured. I am not sure we can. I am not sure he is, either.