Who is doing the thinking?

30 August 2026 · by Sharath · for parents · 3 min read

A child who asks an AI gets the right answer in four seconds. The skills and virtues they would have built getting there on their own are the things quietly being spent, along with the tokens.

Your child asks a machine a question and a good answer comes back in four seconds. Nothing looks wrong. That is the part worth watching, because the answer arrives before the thinking would have.

What gets skipped

A ten year old is still building the habits they will use on every hard thing for the rest of their life. Sorting a mess into groups. Putting things in an order. Cutting a big problem into pieces small enough to solve one at a time. Reading an error and hunting the cause. Staying with something for twenty minutes after it stops being fun.

Nobody can tell them any of that. They build it by doing it badly, then slightly less badly, a hundred times. A child who asks and receives skips every rung. They get the right answer today and have nothing to offer on a day the answer is not there.

Speed is real and it is worth having. Just notice what it is costing your child.

Coding to learn, or coding to build

Learning to code for its own sake is a school subject, and a dull one now that machines write most of it. Building something is different work.

Building gives them a goal they picked, a route nobody has drawn, and a hundred small choices with something riding on each. A long campaign of Dungeons and Dragons does the same: you want something, you cannot see what is ahead, you weigh a risk, and you choose. That is a different animal from the games built to keep a thumb moving. These ones make you own your decision, live with what follows, and carry on with a bad choice until you find your way out of it.

A child building a game meets all of it. It does not run. Nobody rescues them. They read the error, guess, change one thing, run it again. That loop is the lesson. The game is only the excuse for it.

Why building through coding is the safe way in to AI

AI is not radioactive. It is also not a teething toy you hand over and walk away from. It becomes whatever you set it up to be: a patient guide that makes a child think for themselves, or a machine that agrees with everything and does the work, leaving the child elated at that moment, but dull in the long run.

A child who has built with code has experienced AI being wrong, all the way through to fixing it and making it their own. Along with witnessing AI being confident, fluent, and wrong. These experiences change the relationship for good. They stop obeying it and start working with it. They can say in their own words what is being built, and how, and most importantly why, so they can tell working from broken and not be frustrated about it.

The habit matters most a few years later. At thirteen, a machine that never disagrees, always flatters and is always awake is not a tool. It is company with an aim you cannot see. A child who has already caught it lying about a missing bracket is much harder to fool.

None of this asks you to keep AI away from your child. It asks you to put it in their hands second, after they have built something small with their own, so they meet it as the boss and not as the pupil.

Zaalyo is a small Saturday coding school. Ask for a place, or read the others.