The output is the rubric

The most common question we get from parents on the tour, after they've seen the room and met a mentor, is some version of this:

How will we know it's working?

Behind that question is an honest worry, which we respect: parents are measured by school, school is measured by tests, tests come back as numbers, and the numbers are supposed to tell you whether the kid is on track. So when a kid spends two years at a place like Hi, Bot, what's the number that says it was worth it?

Here is our answer. It's not a dodge.

The number is the artifact

The rubric we hold ourselves to is the one we put on the homepage: what did your kid build, and is it on the shelf.

If your kid leaves a cohort with a working AI Pokédex she made for a sibling, an agent she tuned that won a tournament, three paying users for a small product she shipped — we did our job. If she leaves with a stack of certificates and nothing on a screen, we didn't, and you should ask for your money back.

This is not a soft rubric. It is harder than a test. A test you can cram for. A working thing somebody else uses, you cannot cram for. It has to actually work, on a Tuesday, when somebody who is not the kid sits down with it. There is no partial credit.

We picked this rubric on purpose because it's the one that survives contact with the world. When your kid is twenty-four and applying for a job or pitching an investor or asking somebody to give her a chance on a project, nobody is going to ask for her fifth-grade reading score. They are going to ask: show me a thing you made. We are training for that meeting.

A note on test scores, because we get this too

A lot of parents come in worried about scores that have stayed flat. She got a 78th percentile last year. She got a 78th percentile this year. Why isn't she improving?

She probably is. The tests are designed in a way that hides it.

Think of a standardized test as a treadmill that speeds up slightly every year. The material gets harder. The peer group keeps getting better, because everybody is also a year older and a year sharper. A kid who scores the same percentile two years in a row is not standing still — she is keeping pace with a treadmill that sped up.

There are two kinds of tests, and they measure different things. State accountability tests measure whether your kid has mastered the grade-level material. The same scaled score in year two means she mastered harder material — that's progress. National norm-referenced tests measure where she ranks among peers who are also improving. A flat percentile means she is keeping up with a moving group. Both of these can look like stagnation on the parent dashboard and neither one is.

The shape of real learning over years is not a line going up and to the right. It is plateaus, jumps, occasional dips, more plateaus. Brains consolidate during the plateaus. The dip after a jump is normal — the kid is integrating something hard. If you only ever look at the one-year delta, you will misread it constantly.

This is not us telling you to ignore test scores. It is us telling you to interpret them correctly so you don't pull the wrong lever.

What we want you to look for instead

If you want a better signal than a percentile, watch these things:

• Does your kid talk about what she's working on, unprompted, at dinner? A kid who is building something she cares about cannot shut up about it. A kid who is grinding a curriculum can.
• Does she ask harder questions over time? A nine-year-old asking how does the chatbot know that? is a nine-year-old whose questions, six months in, should be sharper — what's the training data, what's a token, why did it get that wrong. The arc of the question is the rubric.
• Does she finish things? Starting is cheap. Shipping is the skill. Count the things on the shelf.
• Does somebody else use what she made? The sibling, the friend, the grandparent, the group chat. One user is worth ten finished projects with no users.

If those four are trending up, the percentile can do whatever the percentile is going to do. Your kid is fine.

What we measure on our side

We track the artifact. Every cohort, every kid, what shipped, who used it. That's the file. When you ask us how your kid is doing, that's what we'll show you. Not a grade. Not a percentile. The list of things she made and the names of the people who used them.

If that's not the kind of report card you want, we are not the right place. If it is, the waitlist is open.