Austin Scholar #217: Why Alpha students stopped asking, “When am I ever going to use this?”

This week from Austin Scholar...

1. Why Alpha students stopped asking, “When am I ever going to use this?”

2. Why traditional school can’t solve the relevance problems

3. What it looks like when relevance is engineered in the program

My sister is graduating from Alpha High this Friday!! I’m so beyond proud of her and am the luckiest person in the world to be her older sister. She’s been working on an incredible valedictorian speech, so I’ll make sure to give y’all the highlights next week. Stay tuned for that newsletter where I’ll highlight the graduating class of Alpha 2026!!

Every teacher has heard the question, when am I ever going to use this, at least a hundred times. The standard answers come in three framings:

• “You’ll need it someday.”

• “It builds discipline.”

• “Because it’s on the test.”

All three are real answers and none of them are good ones, because none of them address what the kid actually said.

What the kid actually said is: I cannot find a single thread connecting this material to anything I care about, anything I recognize, or anything I will plausibly do, and I am asking you to give me one. The student has correctly identified that the lesson, as designed, will not stick and is asking for the missing connective tissue before they invest the effort.

And when the adult cannot supply that connective tissue, it makes sense that the kid stops trying. A long line of cognitive research, which we’ll get to in a second, says that material learned without a felt connection to the learner mostly evaporates. The learning science name for the missing ingredient is relevance.

And on that idea of learning science, let me dive into the four incredible papers that inspired the approach to relevance.

The dominant model of how interest develops in learners is Hidi and Renninger’s four-phase model: a learner moves from triggered situational interest (something in the material catches their attention), to maintained situational interest (they keep coming back), to emerging individual interest (they start re-engaging on their own), to well-developed individual interest (it becomes part of who they are). The model’s most important claim, for our purposes, is the first one: interest is not a fixed trait kids walk in with. It is a developmental sequence, and the entire sequence is gated on phase one, where something in the material itself has to trigger the kid. The kid who looks “unmotivated” in your classroom is, in nine cases out of ten, a kid whose teacher never got past that phase.

In a now-classic line of experiments by Hulleman and Harackiewicz and their collaborators, students asked to write brief reflections on how course material connected to their own lives showed measurable gains in both interest and performance. The effects were largest precisely for the students who were struggling most. This is the cleanest evidence we have that relevance is a manipulable variable that, when increased, produces real learning gains in real classrooms.

Deci and Ryan’s self-determination theory argues that intrinsic motivation requires three things: autonomy (some control over what you do), competence (a sense you can actually do the thing), and relatedness (a sense it connects to people, places, or things you care about). Relevance is the key to unlocking the relatedness aspect.

Finally, a one-line nod to Brown, Collins, and Duguid’s 1989 paper on situated cognition: knowledge learned out of context tends to stay there. The chain rule learned only as “the chain rule” lives in a sealed compartment in the brain labeled “the chain rule,” and refuses to come out when a physics problem actually needs it. Pulling knowledge into the contexts a learner cares about is how knowledge becomes transferable in the first place.

Synthesize all four of these papers and the picture is unambiguous: kids learn when the material connects to them. The reason we have kept ignoring this is not that the research is contested. It is that connecting every kid to every piece of material, every day, has historically been logistically impossible.

That’s the part that recently changed.

Before I get to what changed, I want to make a point that I think gets skipped a lot in this conversation. The relevance failure in traditional school is not a teacher failure. I have had brilliant, devoted teachers who deeply believed in making material relevant and could not actually do it for me, because the structure of their job made it impossible.

The scaling problem. One teacher, thirty kids, one lesson plan. Even a teacher who believes in personalized relevance can hold the cognitive load of customizing for maybe two or three students at a time. The other twenty-seven get the generic version. Multiply across 180 instructional days and the median student has spent the entire year being taught material framed for someone else. This is the same shape as Bloom’s 1984 “2 Sigma” problem — the gap between what works one-on-one and what survives the ratio. Bloom was talking about mastery; the same logic applies to relevance. The thing that works at one-on-one doesn’t survive the ratio.

The textbook problem. Curriculum gets written for an imagined median student who does not exist. The reading passage is about baseball when half the class doesn’t watch baseball. The word problem is about a train when no one in the room has been on one. The historical example is the one in the textbook because it’s the one in the textbook. Publishing economics demands a single artifact that ships to every classroom in the country, so the artifact is calibrated to no one. The teacher’s options are to stick to the materials and bore most of the class, or to spend their evenings rebuilding the curriculum for the kids in front of them (which a small number of heroic teachers actually do, until they burn out).

The IOU problem. Because individualized relevance is structurally impossible, teachers fall back on deferred relevance: you’ll use this someday. Kids correctly intuit that “someday” is an IOU they cannot cash. They watch the adults around them and notice, accurately, that almost no one is actually using the polynomial long division they were promised they’d need. Motivation collapses. Compliance fills the gap, for a while, until it doesn’t, and the kid checks out.

None of this is anyone’s fault. It is what happens when you try to deliver individualized relevance to thirty kids with one adult, one textbook, and forty-five minutes. The bandwidth doesn’t exist. It never has.

Which is why the only honest version of the relevance problem, until very recently, was: we know it matters, we know how it works, and we cannot do it at scale. The thing that ended that sentence is AI. The tool that didn’t exist now exists. And it is most clearly visible, right now, at Alpha which has rebuilt the curriculum on the assumption that the tool is here.