Speedrun the School Day

The high-stakes, AI-driven bet on the two-hour school day

If you were to visit an Alpha School classroom at 9AM, the first thing you’d notice is the silence.

There would be no teacher lecturing at the front of the classroom; no worksheets being reviewed; no textbooks being flipped through.

You’d see children—headphones on, heads down—siloed in their own academic worlds. Peer closer, and you’d see software that adapts in real-time to their cursor movements, feeding them math and reading modules tailored to their exact proficiency.

This is 2-Hour Learning, the central dogma of Alpha School. The premise is radical in its simplicity: speedrun learning with AI in the morning, and free up time for activities in the afternoon.

After all, traditional school is a maelstrom of inefficiencies, social distraction, and “teaching to the middle.” By stripping away the lecture and handing instruction over to adaptive AI, Alpha claims a student can master a full day’s worth of academic material in just two hyper-focused hours.

After that, the day livens up: afternoons are spent on “project-based ‘life skills’ like entrepreneurship and public speaking.” Children train for 5Ks and develop business ideas. According to Alpha’s website, “the core of the system is solving for motivation, which is 90% of the problem, by giving kids their time back.”

That’s the elevator pitch. Some are tantalized by Alpha’s tech-obsessed approach; others find it viscerally repulsive.

The truth, as usual, is more complicated than its advocates or critics would have you believe.

A Big Tech Fever Dream

The first big criticism against Alpha—as undeniable as it is uninteresting— is that this very expensive private school ($40,000/year in Austin, TX) is education for the rich, by the rich.

Here’s the first sentence from a recent NY Times piece on Alpha School:

“In Austin, Texas, where the titans of technology have moved their companies and built mansions, some of their children are also subjects of a new innovation: schooling through artificial intelligence.”

The implication is clear: Alpha is for alphas—a fun school for the children of tech titans who, by the way, sleep in mansions when they aren’t shoveling down AI edu-slop.

Dan Meyer, a former high school math teacher and ed-tech critic, makes this point more explicitly (and to better effect) in a piece on his Substack , in which he argues that Alpha “gets great results, as far as I can tell, not by replacing teachers with AI, but by replacing poor kids with rich kids, by replacing unengaged families with engaged families.”

Meyer raises a crucial point: it’s nearly impossible to measure schools against each other because the variables are never controlled. We know that socioeconomic status is the single biggest predictor of academic success. When you select for families willing to pay $40k for an experimental AI school, you are selecting for high achievement before the kids even log in.

But dismissing Alpha as a playground for the wealthy misses the point. The rich have always had weird schools. I suspect the real reason many want Alpha to fail is that it feels like an Elon/Thiel/Trump fever dream of “tech-libertarian” school choice. I can only imagine how many Cybertrucks make the loop at morning drop off.

It’s hard for education-minded liberals to work with this aesthetic. It doesn’t help that Mackenzie Price, one of the school’s founders, leans hard into the “disruptor” persona.

Mackenzie Price

But the tech-skeptical left and the tech-obsessed right will eventually need to find common ground to fix schooling. And there are better reasons to be skeptical about Alpha than mere “eat-the-rich” rejectionism.

For one, Alpha-lytes seem intent on recreating themselves: miniature, mirror versions of the tech elite. Their “Life Skills Curriculum” includes workshops like Podcast host, TedX speaker, and Spartan Race finisher.

And the Life Skills in question are surprisingly narrow in scope:

• Teamwork + Leadership

• Storytelling +Public Speaking

• Entrepreneurship + Financial Literacy

• Relationship Building + Socialization

• Grit + Hard Work

This is a recipe card for creating CEOs. Alpha doesn’t seem too concerned with creating the next generation of scientists, poets, astronauts, civic servants, and so on.

It’s a natural human impulse to use education in an attempt to manufacture better versions of ourselves. But that temptation should be resisted.

A school—much like an AI model, actually—is optimized to spit out whatever you prompt it to produce. If our values dictate the output, we shouldn’t narrow the scope to “podcast host” or “tech company founder;” we should prompt our schools toward a much wider palette of human potential.

And to get a richer output, you need a better input.

Part of this means prioritizing teachers who possess deep wells of knowledge to impart—rather than just “guides” to manage the workflow.

Guides on the Side

Another substantive concern is Alpha’s staffing model. Alpha hires “guides,” not teachers.

While many are certified educators, the model explicitly deprioritizes subject matter expertise. In a recent interview, Price was asked how a parent should feel about their high schooler taking physics from a guide who…doesn’t know physics.

“It’s okay that our guides aren’t experts in physics or math,” Price said. “They’re experts in their students.”

That is a comically bad answer.

It relies on the assumption that “skills” can be taught in a vacuum, independent of content. But as education journalist Natalie Wexler argues in her excellent book The Knowledge Gap, you can’t separate something like reading comprehension from background knowledge.

You can’t “skill” your way through a complex physics problem or a historical text; you need a knowledgeable human to help you construct the framework of understanding.

You need to actually know things.

And deep knowledge is anchored in lived experience—rich, sensory encounters in physical space.

Cognitive science tells us that memory formation relies on novelty and context. A chatbot offers neither; it provides a flat, uniform interface where every lesson feels exactly the same.

Perhaps a future iteration of Alpha’s platform will solve for this lack of texture, but it isn’t here yet. So for now, if the software fails to engage and the adult in the room is just a cheerleader, who helps the student when they hit a wall?

But…What If It Works?

I quibble with Alpha’s execution. It feels a bit slick: light on the humanities, thin on substance, lacking in soul.

But so are plenty of schools. Here’s the thing: what if the basic mechanism at the heart of Alpha actually works? What if personalized algorithmic learning is better, faster, and more scalable than the old armory of textbooks and lectures?

Research on “Intelligent Tutoring Systems“ (ITS) has been ongoing for decades. Meta-analyses generally show that while AI tutors aren’t yet as good as human tutors, they are often significantly more effective than traditional large-group instruction.

If Alpha can perfect this software, the implications for the developing world are staggering.

Founder Joe Liemandt has been explicit about this. The $40,000 tuition is essentially R&D funding for the real goal: cheap tablets that can deliver a world-class K-12 education to any child on Earth in two hours a day.

If you can bypass the global teacher shortage and teach numeracy and literacy via cheap hardware, you have solved one of humanity’s oldest problems.

The Verdict

So does Alpha work? Surprise: we don’t know yet.

One of the most comprehensive deep dives into Alpha comes from a finalist in the Astral Codex Ten 2025 review contest. Written by an anonymous parent who moved their family across the country to Austin specifically to test the model for a year, the review offers a rare look behind the PR curtain.

The author confirms the core promise: the academic acceleration is real. Their children, attending the “Gifted and Talented” version of Alpha, mastered material roughly three times faster than their age-matched peers. Students at Alpha show approximately 2.3 to 2.6 times the national average academic growth on MAP (Measures of Academic Progress) assessments.

However, the author clarifies that the “2-hour” claim is slightly misleading—it’s closer to three hours of work spread over a morning—and that the AI isn’t a proprietary generative chatbot, but rather a “turbocharged spreadsheet checklist with a spaced-repetition algorithm.” Alpha’s “custom AI platform” is currently just a dashboard of pre-existing apps like IXL and Khan Academy.

Crucially, the success this parent is seeing isn’t just due to the software; it relies heavily on a robust system of financial incentives (school currency) and the aforementioned “guides” who keep the kids on track. In other words, Alpha has mastered reward systems: achievements, points, and other incentives are ridiculously effective. This will come as no surprise to long-time readers of Quests!

Ultimately, the author concludes that while the model is expensive and definitely weird, it successfully solves the problem of efficiency, gifting children years of free time back to their childhoods.

So, yes, Alpha might dramatically oversell the AI magic. But they’re on to something.

And I’m open to the possibility that software can eventually deliver richer, more memorable experiences in the pursuit of academic mastery. Hopefully, this will happen alongside—and not replace—human teachers who actually know physics.

The upside is enormous. So, I hesitatingly applaud Alpha’s mad-science experiment—even for its faults.

Now it’s time for you to choose.

Path A: Speedrun the School Day with AI

Individualized mastery is the only way forward. Let’s explore how to make this accessible to everyone, not just the rich.

Path B: Stick with Human Teachers

Education is relational and needs to be grounded in physical space. Let’s explore models that double down on human mentorship and analog learning.

As always, thanks for reading!