My wife and I both went through traditional schools. Good ones, by most measures. We came out literate, credentialed, and employable. By any metric the system worked.

So why does the idea of sending our own kids through the same thing make us so uncomfortable?

We want to start a family soon, and the closer that gets, the harder it becomes to ignore a question most parents either agonize over or sleepwalk through: where do you send your kids to school? Not which school… What kind of school.

Last week I published an essay called 16,000 Hours about the structural problems with American education. My main argument was that a child spends roughly 16,000 hours in compulsory schooling but the core material requires only a fraction of that time. The rest is crowd management dressed up as learning. Thirty minutes of instruction stretched to fill an hour. Review sessions for kids who didn't get it, while the kids who did sit idle. Worksheets designed to keep forty kids quiet, not to teach them anything.

The system was built during the Industrial Revolution to produce literate factory workers and assimilate immigrants. It has barely changed since.

That essay explored the problem. This one is about what's actually available if you decide the default isn't good enough. I've spent the last several months reading the research, comparing models, and trying to understand what works, what doesn't, and what the tradeoffs are.

A caveat before we start: every alternative model carries a selection bias. Families who actively choose something different are often more engaged than families who accept the default. That engagement probably matters more than any curriculum. Keep that in mind as you read the data. It doesn't invalidate the results, but it should make you appropriately skeptical of any single model claiming to be a silver bullet.

---

Established Alternatives

Montessori

Montessori is the most established alternative with the deepest body of research behind it. The model was developed by Maria Montessori over a century ago and is built around self-directed learning. Students choose their own activities from a structured set of hands-on materials. Classrooms are mixed-age. There are no grades, no homework, no lectures. The teacher observes, guides, and intervenes when necessary. The child drives the process.

The data is strong. A 2023 meta-analysis published in the Campbell Systematic Reviews examined 32 studies across 132,000 data points and found that Montessori students outperformed traditional students on virtually every measure. By sixth grade, Montessori students were roughly a full school year ahead in language, math, and general academic ability.

The nonacademic gains were even more striking. Montessori students showed significantly stronger executive function (the ability to self-regulate, plan, and hold information in working memory) than their traditional peers. The largest gap of all was in how students experienced school itself. Montessori kids simply liked being there more. Across 132,000 data points, that was the single most consistent advantage.

Effects were strongest at the preschool and elementary level, and stronger in private Montessori than public implementations. That distinction matters because the Montessori name isn't trademarked. Anyone can put it on a school sign. Fidelity to the actual method varies enormously, and the research suggests fidelity is what produces results.

There are roughly 3,500 Montessori schools in the US, with tuition ranging from free (in public magnet programs) to $20,000+ for private. It's the most accessible of the proven alternatives.

The concern is what happens after elementary school. The research thins out considerably at the middle and high school level, and parents regularly report that the transition to traditional schooling can be rough. A child who has spent eight years directing their own learning may struggle when suddenly asked to sit in rows, follow a fixed syllabus, and perform under timed exam conditions. The very qualities the model cultivates (autonomy, self-direction, intrinsic motivation) can become liabilities inside a system that rewards compliance.

Waldorf

Waldorf takes a different bet. Where Montessori trusts the child to lead, Waldorf trusts the developmental sequence. The curriculum is arts-integrated and staged to match what the school’s founder, Rudolf Steiner, believed children are ready for at each age. Formal reading instruction is delayed until six to seven years old. The same teacher stays with a class for multiple years. Screens are absent from the classroom. Students knit, do woodworking, paint, and perform music alongside their academic work.

The academic results tell an interesting story. A study of California's Waldorf-inspired charter schools found that Waldorf students significantly underperformed both charter and traditional public school peers in every subject between third and sixth grade. Then something shifted. By seventh grade, Waldorf students caught up. By eighth grade, they surpassed both comparison groups by nearly five percentage points in ELA and math. The researchers described it as a "slower academic build-up" producing stronger advanced performance. The tortoise and the hare.

Where Waldorf consistently shines is in the nonacademic outcomes. Studies show higher motivation, greater enthusiasm for learning, lower aggression, stronger moral reasoning, and more positive attitudes toward school. A 2012 German study found Waldorf students significantly more enthusiastic about learning and more likely to feel individually known by their teachers. A longitudinal survey found over 80% of Waldorf alumni rated their education as the most important influence in developing their sense of responsibility for others.

The tension with Waldorf is real, and it cuts two ways. The delayed reading approach has more support than most parents expect. A longitudinal study by Sebastian Suggate, published in the Journal of Research in Reading, found that while early readers appear to outperform late readers in the short term, the gap disappears entirely by age 11; and late readers often surpass their peers in comprehension and motivation. Finland delays formal reading instruction until age 7 (the same age as Waldorf) and consistently outperforms almost every other country in literacy. The data suggests that the anxiety parents feel about a non-reading seven-year-old may be exactly backwards.

But that aggregate picture hides a genuine risk. Research on the science of reading shows that 20-30% of children need explicit, systematic phonics instruction to become fluent readers, and children who aren't reading by the end of third grade face significantly higher odds of never reading fluently. Traditional Waldorf schools have historically resisted structured phonics programs, and some parents of struggling readers report hitting a wall when they try to get early intervention. Several Waldorf schools have begun integrating structured literacy alongside their traditional methods (the Lake Champlain Waldorf School in Vermont is one example), but this adaptation is far from universal. If your child happens to fall in that 20-30%, the delay could cost them.

Montessori and Waldorf arrive at the same conclusion from different directions: slowing down early pays dividends later. Montessori lets the child set the pace. Waldorf holds the child back from abstraction until they're developmentally ready for it. Both produce students who underperform on early standardized metrics and outperform on later ones. Both produce students who actually like school.

The factory model's obsession with early measurable outcomes (reading by five, multiplication by seven, test scores by eight) may have the entire sequence backwards. We're optimizing for the wrong end of the timeline.

---

The New Guard

Alpha School

Alpha School (@AlphaSchoolATX) is the model getting the most attention right now, and it most directly validates the argument I made in 16,000 Hours.

Students spend two hours each morning learning core academics through an AI-driven platform that adapts in real time to what they know and what they don't, what engages them and what doesn't. There are no teachers in the traditional sense. "Guides" earning six-figure salaries provide motivational and emotional support rather than academic instruction. Afternoons are devoted to life skills workshops, entrepreneurship, public speaking, and passion projects. There is no homework.

The results are hard to dismiss. Alpha students score at the 99th percentile on NWEA MAP testing in every grade from kindergarten through eleventh, in every subject tested. Their senior class SAT average is 1535. Their freshmen average 1410, outscoring most schools' juniors and seniors. But the most remarkable finding is how fast they're still improving. In educational statistics, high-achieving students typically show slower growth (there's less room to gain when you're already near the ceiling).

Alpha students are learning at 1.7 to 5 times the national average despite already being at the 99th percentile. That combination (elite achievement and elite growth simultaneously) is the rarest outcome in American K-12 education. In 2025, Alpha graduated its first class of seniors. Eleven of twelve went on to four-year universities; the other founded a company and raised VC funding.

Joe Liemandt (@jliemandt), Alpha's principal, published a detailed breakdown of these numbers in a mid-year report card thread.

The obvious caveat is cost. Tuition starts at $40,000 in Austin and climbs to $75,000 in San Francisco. Parents paying that kind of money for an experimental school are not a random sample of the population. They're among the most engaged, most resourced families in the country. You cannot extrapolate Alpha's results to the general public and call it a day.

But that's not really the point. The key thing that Alpha demonstrates is that excellent academic outcomes can be achieved in two hours of focused instruction per day. Everything else in the school day (the remaining six hours of athletics, projects, and life skills) is a bonus, not a prerequisite.

Now consider what the public system spends. The U.S. averages roughly $17,000 per pupil per year; in New York, it's closer to $30,000. That's $857 billion annually across the K-12 system. And the returns? NAEP scores in math and reading are worse today than they were in 2003. Public schools are spending more per student than at any point in history and getting less for it.

Alpha School suggests that the bottleneck was never money. It was time allocation. If two hours of well-designed instruction can outperform seven hours of traditional schooling, the question for public education is “what on earth are we spending all of this time and money on"? The answer to that could make public school not just more effective, but cheaper for taxpayers.

Acton Academy

Acton Academy makes a different bet on the same underlying insight. Founded in 2009 by Jeff and Laura Sandefer, Acton organizes education around the "hero's journey." Students (called "eagles") set their own goals, design their own learning paths, and are held accountable not by teachers but by their peers. Socratic discussions replace lectures. Real-world apprenticeships replace theoretical career prep. Adults are guides who ask questions rather than provide answers.

Students at the Austin flagship have been advancing as much as three grade levels per year. Half of their sixth graders exceed state twelfth-grade standards. Graduates leave with portfolios and seven years of apprenticeship experience that colleges are increasingly valuing over traditional transcripts. One graduate was hired directly out of high school by a Silicon Valley company. Another was asked by a college professor to lead multiple classes in goal-setting and project management as a freshman.

What makes Acton structurally interesting is the franchise model. There are now over 300 Acton Academies worldwide, with tuition around $8,600 to $9,300; roughly what taxpayers already spend per public school student in many states. That makes it one of the more accessible options on this list.

Franchise quality varies, though, and significantly. The model requires exceptional guides and deeply committed parents. Acton also takes its own standards seriously enough that 10-15% of students don't make it and get voted out by their peers. You can read that as harsh or as honest, but a system that enforces its own standards is at minimum a system that believes in something.

The gap worth flagging is evidence. Montessori has a century of research and a 32-study meta-analysis. Waldorf has longitudinal data stretching back decades. Acton has testimonials and internal metrics. The "three grade levels per year" claim comes from the school's own reporting; there's no independent, peer-reviewed study confirming it across the network. That is not to say that the results aren't real. It means we're taking them on faith in a way we don't have to with the established alternatives.

Alpha School uses AI to compress academics into two hours and fills the rest with life skills. Acton Academy uses peer accountability and Socratic discussion to make the student the driver of their own education. Both reject the assumption that learning requires seven hours of adult-led instruction. Both produce students who outperform the traditional model by wide margins. And both, in different ways, point toward something genuinely exciting: if the core insight holds (that focused, personalized instruction beats bulk seat time), then the architecture of the factory school isn't just outdated. It's optional.

The question is how to make that option available to everyone, not just the families who can afford to opt out.

---

Homeschooling and Microschools

The fastest-growing segment of American education isn't a school at all.

Homeschooling now accounts for roughly 3.7 million students in the US (about 6.73% of school-age children). According to Johns Hopkins' Homeschool Research Lab, growth in 2024-25 averaged 4.9% nationally—nearly triple the pre-pandemic rate. 36% of reporting states recorded their highest homeschool enrollment ever, surpassing even pandemic peaks. The researchers at Johns Hopkins are clear: rising homeschool rates cannot be explained away by the pandemic. We are experiencing a structural shift in how American families think about education.

The research data, while complicated by the selection bias I mentioned earlier, is consistently positive. Homeschooled students score 15 to 25 percentile points above public school students on standardized tests. 78% of peer-reviewed studies show homeschoolers performing statistically better than institutional peers. College acceptance rates run around 87% for homeschoolers versus 68% for public school graduates.

The most interesting development here isn't the solo family at the kitchen table, though. The category has outgrown that image entirely. 29% of homeschool families now participate in hybrid models: co-ops, microschools, and learning cooperatives. These are small, often parent-founded communities of 5 to 30 students that combine the personalization of homeschooling with the structure and socialization of a traditional school. The National Microschooling Center's 2025 report found microschools operating in all fifty states, with 84% serving elementary-aged children and a growing number expanding into middle and high school.

Microschools solve the two biggest criticisms of homeschooling (socialization and parental burnout) while preserving what makes it work (personalization and parental agency). They also address the biggest structural problem with traditional school: scale. A microschool with 15 students and two adults can do things a classroom of 40 simply cannot.

The cost of homeschooling itself is low ($700 to $1,800 per student annually), but it requires at least one parent with serious time flexibility, which is its own form of expense. Quality depends entirely on parental capability and commitment. Regulatory oversight also varies wildly by state.

Colorado, where I live, has relatively light requirements, which cuts both ways. On one hand, freedom for capable families. On the other, no safety net for kids whose parents aren't up to the task.

But the trajectory is clear. Homeschooling is no longer a fringe decision made by religious families or off-grid idealists. It's a rational response to a system that isn't keeping up. And the hybrid models (co-ops, microschools, learning pods) are quietly building the infrastructure that could make personalized education scalable without requiring a stay-at-home parent or a $40,000 tuition check. If any model eventually closes the access gap, it will probably look less like a school and more like this.

---

What the Data Tells Us

I've spent months sitting with the question of what kind of school to choose and I still haven't decided what my wife and I will do when the time comes. My goal with this essay is not to sell you on a model, but to share what I have found.

Across every alternative I studied, four patterns hold:

Personalization works. Every model that outperforms the factory school shares one structural feature: it adapts to the individual student rather than forcing the student to adapt to the group. This is exactly what Benjamin Bloom proved in 1984 when he found that one-on-one tutored students outperformed 98% of classroom peers. The alternative models have each found a different way to approximate that effect.

The mentor beats the lecturer. Whether it's a Montessori guide, a Waldorf class teacher, an Acton eagle, or an engaged homeschool parent, the relationship between teacher and student is the mechanism that produces results. The ancient Greeks understood this. Socrates didn't deliver lectures. He asked questions. The factory model abandoned that insight for efficiency, and we've been paying for it since.

Early metrics are not destiny. Waldorf's slow build, Montessori's child-led pacing, and Alpha's compressed academic day all suggest that trusting the developmental process produces better long-term outcomes than optimizing for early standardized performance. The parents most likely to choose these models are the ones comfortable enough to ignore the scoreboard for a few years. That comfort is itself a privilege.

Cost remains the binding constraint. The best alternatives are either expensive (Alpha, private Montessori) or demand enormous parental investment of time and energy (homeschooling, co-ops). Technology may eventually close that gap. AI tutoring could make Alpha-style personalization available for a fraction of the cost. But right now, the families with the most options are the ones who need them least. That's a problem none of these models have solved, and it's the one that should trouble us most.

I don't have a recommendation. What I have is a conviction that the choice matters more than most parents realize. Choosing the traditional schooling model by inertia is still a choice; you're just making it passively. Those 16,000 hours are a finite resource. Once they're spent, they're gone.

The least you owe your kids is to make that choice deliberately.