Here’s my pitch for you:
I can run a school where the students spend 2 hours a day — TWO hours — in required academic class work. They get to spend the rest of their time working on passion projects. They’ll be highly motivated. And they’ll rack up all-but-unbelievable standardized test scores.
Depending on your beliefs and experience, you might say:
- “That sounds AMAZING! Where do I sign?” Or:
- “Please — that’s obviously too good to be true. I’m not buying that bridge…”
Well, schools making this claim exist. They’re called the Alpha Schools. And they’re creating quite a stir in the intersection of cognitive science and education.
Here’s the story.
Alpha 101
For a quick introduction to the Alpha, you could start with this article by Chad Aldeman. Although initially a skeptic, Aldeman now sounds cautiously optimistic about the possibilities.
In brief, Alpha schools use AI to gauge a student’s current level of knowledge: say, how much of the state’s 6th-grade math standards does the student already know? AI then builds tailored lessons for that student — reviewing key material, introducing the most pressing topics, and consolidating uncertain prior knowledge.
Each student works with an adult guide, who keeps track of progress and — when necessary — helps refocus on the work at hand. But this guide isn’t teaching; the AI takes care of the instructional work.
After two hours of this coursework — with appropriate breaks — students get to focus on their own interests: surfing, public speaking, outdoor education. According to the Alpha model, the chance to do all these independent activities provides the motivation that students so often lack. They’re willing to do all the personalized AI tutoring because they get dessert at the end of that academic meal. As Aldeman notes, at least one Alpha school offers another kind of dessert: students earn $100 for acing the state math test. To be clear, that’s any grade level. If a seventh grader scores 100 on the 3rd grade state math test, that’s $100 right there.
Other forces amp up the motivation in the Alpha model as well. Because students can feel their own progress, this academic success should reinforce motivation. Self-determination theory tells us that the feeling of competence itself enhances motivation.
These academic and incentive structures might seem…implausible. But Alpha claims impressive results. According to Aldeman’s data, in 2024-25
“the main Alpha campus had, depending on the grade level, 67% to 90% of students meeting their growth targets in math and 65% to 100% of students meeting their targets in English.”
When I ran those numbers by an experienced colleague, she responded with frank astonishment. She said — more or less — that those results are either fraudulent or off-the-charts amazing.
If you want a deeper data dive, check out this link from the principal of an Alpha school.
Not So Fast
The Alpha model has gotten lots of admiring attention — especially because of those standardized test results. At the same time, it’s getting equally vociferous pushback about its focus, its cost, and its viability at scale.
Take, for instance, a recent post by math expert Michael Pershan. He argues, with passion and conviction, that Alpha is focusing on the wrong priorities.
“Alpha is not trying to provide the best, most ambitious math or ELA education possible according to conventional understandings of that term. If they were, they’d keep studying ELA/math in the afternoon. Instead, their goal is to minimize the time spent on core academics while maximizing skills.”
In brief, if a school’s primary selling point is how little time the students spend on academics, then buyers should beware the product.
Other critics worry about the school’s price tag: I’ve seen numbers ranging from $40,000 to $65,000 per year. Alpha’s rejoinder — as I understand it — is that their method will be honed in private schools, but ultimately made available at scale. (I’ve seen the tagline “transforming education for 1 billion kids.”)
Another recent post — this one by Dylan Kane — rejects this altruistic logic. To design a pedagogical model that works for struggling students, he argues, we need to start with struggling students:
“Testing curriculum at Alpha School is like testing a new medication on the healthiest patients first.”
Pedro de Bruyckere makes this point with considerably more gusto:
Two hours of academic content on adaptive apps is not a substitute for education. It’s screen time with better marketing. Sure, highly motivated kids from wealthy, educated families will click through faster and score higher on standardised tests. They already have every advantage in the world. However, try to scale that to the messy, real classrooms where motivation, context, and relationships matter. Here, the whole model falls apart.
These perspectives suggest that the promise to “make Alpha available at scale” will ultimately prove hollow.
Wait Just A Moment…
This deep — occasionally angry — skepticism is counterbalanced by thoughtful, experienced enthusiasm.
For instance, Zach Groshell is nobody’s pushover. He regularly combats education myths with piles of research. His book Just Tell Them has been widely praised for his evidence-informed championing of direct instruction. And: Groshell is all in on Alpha; in fact, he now works for the organization. (I should note: Zach is a good friend.)
In a recent blog post, he offers several rejoinders to common anti-Alpha arguments. For instance:
- Schools that teach children in groups inevitably make compromises: after all, in a classroom with a wide range of skills and experience, “what helps a novice overwhelms an expert, and what challenges an expert leaves a novice behind.” By tailoring the classroom experience precisely to each individual, Alpha could solve this problem.
- If we’re being honest, most schools have some awesome teachers and some perfectly good teachers. But teaching is a remarkably difficult job, and altogether too many teachers lack the skills or training to get it done. The result: children are often at the losing end of an education lottery.
- In many systems, teachers’ professional autonomy lets them work without enough accountability. A student with an ineffective teacher has no meaningful recourse.
In other words: we shouldn’t compare Alpha to an ideal educational system, but to the system we really have. In Groshell’s view, Alpha is obviously superior.
Groshell is not the only highly-respected scholar who’s on board. In his recent post on “Educational Invariants,” Carl Hendrick talks about his “cautious optimism” that Alpha’s model can work in ways that educational apps don’t.
My Opinion
I enter into this debate with a background in skepticism; I even wrote a book on the topic. The Alpha promise sounds too much like promises we’ve heard before:
“A new technology will remake the deeply intricate, deeply human work of education. Just over that horizon, school will be both intrinsically motivating and universally successful.”
However, I’m hesitant to reject compelling evidence.
- If those remarkable test scores are not fraudulent, then they’re remarkable.
- If Zach Groshell and Carl Hendrick and Sarah Cottinghatt — who know the school much better than I — say it merits our respectful attention, I’m inclined to pay respectful attention.
The old quip asks: “I know it works in fact, but does it work in theory?” If Alpha violates the theories I believe but the students actually learn, I would be foolish to dismiss it.
I don’t know that the AI tutors are using good cognitive science principles — retrieval practice, interleaving, working-memory management, and so forth. But it certainly seems plausible that they could.
For all these reasons, my opinion is: let’s watch this space. We should interrogate the data and the hype…and simultaneously be open to the possibility that Alpha Schools could create a genuinely new school model.
