The other day I needed to remember how to make something I had not made in a long time. Once, that sort of question would have meant reaching for a cookbook or calling someone who knew. Now I can type a few words into my phone and have an answer before I have found the right pan.
There is nothing particularly troubling about that. I do it all the time. Technology has saved me countless hours of searching for information, filling out forms, correcting mistakes, and doing repetitive work nobody will someday remember fondly. Artificial intelligence is already helping people translate languages, navigate disabilities, learn new skills, and find information that once required a well-stocked library and a free afternoon.
Still, I wonder about the phone call that doesn’t get made. If I call someone to ask how she makes her soup, I may get the same instructions a computer could give me, probably less precise ones. She may tell me to add enough salt “until it tastes right,” which is not especially helpful if I am trying to measure things. She may forget an ingredient and call back ten minutes later. Somewhere between the onions and the stock, though, we may start talking about something else. I called for a recipe and received a person.
That distinction matters because not everything we do can be measured by the quality of the finished product. A computer can help me compose a beautiful condolence note, perhaps more eloquent than anything I could write on my own. But part of the meaning of a condolence note is knowing that someone sat down, thought about us, struggled for words, and bothered to send them. The struggling may be part of the gift.
This does not mean difficulty is automatically good. We have a strange tendency to romanticize inconvenience once we no longer have to endure it. Nobody needs to rediscover the character-building benefits of standing in a bank line for 45 minutes. If a machine can eliminate drudgery, open a door for someone who otherwise could not enter, or give us more time for the people we love, then making life easier may be precisely the point.
The harder question is deciding which difficulties are merely obstacles and which are part of what gives an activity meaning. Learning to play an instrument badly before playing it well is not wasted time. Neither is wrestling with a difficult idea, making a meal that doesn’t quite work, finding our own words for an apology, or sitting beside a grieving friend when we have absolutely no idea what to say. Sometimes the important thing isn’t producing the right response. It is giving another person our attention and remaining there when there is nothing to fix.
Artificial intelligence makes this question more urgent because it is becoming remarkably good at producing things that once required considerable human effort. That is an extraordinary achievement, and I don’t think we need to fear it. But the better our machines become at performing human tasks, the more carefully we may need to ask what the task was for in the first place.
Was the purpose of writing the birthday message to produce 50 cheerful words, or to remember the birthday? Was education about getting the answer, or becoming someone capable of understanding it? Was the recipe only about getting dinner onto the table? Sometimes it was. Tuesday night is Tuesday night, and hungry people need to eat. But sometimes the long way around carries something the shortcut cannot.
I don’t think the question before us is whether we should use artificial intelligence. We will, and in many places we should. The more interesting question is what we will do with the time and effort it gives back to us. When a machine makes something easier, does the shortcut leave us with more room for one another?
Because sometimes the recipe is just a recipe. And sometimes it is a reason to call Mom.
This is stronger. In particular, “I called for a recipe and received a person” is the center of gravity. Everything after it is really an exploration of what that sentence means.
I also like the ending more after tightening it. It doesn’t answer the headline with a rule about what AI should or shouldn’t do. It gives the reader a discernment question and then returns to the kitchen. That’s exactly the Threshold move.
And beneath the surface, without ever saying so, the whole column now carries the grammar we’ve been developing: output is not participation; task is not responsibility; assistance can enlarge human presence or replace an occasion for it; the measure is ultimately what happens between persons. That makes it an excellent third sample beside the memory and civility pieces.
