AI-Accelerated Academics Make Room for Real-World Skills

Joe Liemandt frames Alpha School’s model around a question: if students can learn core academics faster, what should fill the time conventional schools spend on instruction?

Liemandt says classroom instruction has changed little in a century, even as cognitive science, learning science, and large language models have made it possible to build what he calls a learning engine. He describes the engine’s potential as teaching students ten times faster than the traditional classroom model. Separately, he says Alpha students learn twice as much in two hours of academics a day as they would in six hours of class plus homework. Those are distinct claims: one describes the engine’s potential; the other is how he characterizes Alpha’s academic day.

In Liemandt’s account, the technology is not a new tool to insert into an unchanged school day. The school day itself has to be redesigned around it. Alpha makes three commitments to students, he says: they will love school, learn more in less time, and spend the rest of the day developing capabilities the standard curriculum does not systematically teach. In surveys he cites, more than 90 percent of Alpha students say they love school, and half say they would rather attend than go on vacation. At one campus, he says, the figure is 65 percent.

The afternoons are not meant to be extended recess. Students work on projects with peers and adults, organized around five areas: leadership and teamwork; entrepreneurship and financial literacy; socialization and relationship building; storytelling and public speaking; and grit and hard work. The time freed by faster academic instruction becomes room for skills Liemandt believes matter in a changing economy but are often left to extracurricular activities—or not taught at all.

That makes the central change more demanding than adopting an AI tutor. Liemandt says a school must be willing to redesign its day around students’ motivation and make school an experience they choose to return to. Otherwise, he argues, a learning engine dropped into a conventional schedule will become another EdTech product with limited effect. The test is not whether software can deliver instruction; it is whether the institution will use the time it frees to teach skills students might otherwise never practice.

That redesign also has to hold up as Alpha grows. Liemandt says scaling schools while maintaining quality is difficult, and describes productizing the model—making it understandable and repeatable—as a central part of his work after becoming principal. He says Alpha’s test scores, customer satisfaction, student happiness, and workshop quality have improved as the system has grown. His formulation is that Alpha “gets better as we get bigger”: growth, in his account, is meant to improve the model rather than dilute it. He also says AI helps the school scale quality.

For Liemandt, making the model legible is part of that work. His background is in product and entrepreneurship, and he says Alpha needed to turn its approach into something families could understand and campuses could reproduce. Gagan Biyani describes the shift Alpha has made as one of believability: showing that the approach works well enough for people to ask why school should be organized differently.

Joe Liemandt rejects the idea that a school children love must be easy or unserious. His account is almost the reverse: students enjoy school because they are held to high standards and get to take on difficult work with friends. “Kids want to go do hard things,” he says. The challenge has to be structured so students can see progress and experience success, rather than repeatedly encountering work they cannot yet do.

In academics, Alpha’s learning engine is designed to keep students in what Liemandt calls an 80-to-85-percent-correct range. If a student answers everything correctly, the material is probably too easy; if it becomes too difficult, the student risks disengaging. The system can give a student prerequisite lessons when gaps appear, or move them ahead when the material is already mastered. A student in eighth grade who needs a fifth-grade lesson, he says, can take that lesson and then progress.

The point is not personalization for its own sake. Liemandt describes the engine’s core value as engineering “wins” for students: work should be demanding enough to produce learning but attainable often enough to sustain effort. In his analogy, a video game that lets players fail too often loses them. A classroom teacher managing a group of 30 students cannot as easily keep each child’s work at that level, he argues.

This target also changes what it means for a student to be “behind.” Liemandt says that in a time-based system, a student who falls behind can spend the rest of the year struggling with material built on concepts they have not learned. Alpha’s engine can return that student to the missing lesson, let them master it, and then move on. The goal is not to lower the standard, but to give each student work they can do while they are learning. When students are succeeding at appropriately difficult work, Liemandt says, they become excited by their progress.

Alpha treats life skills as things students should practice and pass, not as soft aspirations. Liemandt offers teamwork as an example. In his view, ordinary group assignments often teach the opposite of teamwork: one student takes over while others sit back. Students remember those projects as frustrating, he says, and learn the wrong pattern for working together.

For one exercise, third graders wear microphones while completing an escape room. AI analyzes the transcript to estimate how much of their language is supportive and team-oriented, rather than focused on fighting over whose idea to follow. Liemandt says less than 20 percent of their language was uplifting before the workshop. Students then take an eight-week workshop and return to the task. To pass the class, he says, they must bring the share of their language that reflects team-oriented behavior up to 85 percent. He also reports that they finish the escape room three times faster when working as a team.

The exercise makes the assessment concrete: students do the task, review how they communicated, practice a different way of working, and have to meet a threshold. The measure is not just whether the group escapes, but how members speak to and support one another along the way. Liemandt says the activity is fun for students while teaching a skill parents want them to have.

Alpha carries that training into later grades. Liemandt says fourth- and fifth-graders must pass a Wharton MBA teamwork simulation based on climbing Mount Everest, which he says half of MBAs fail. He presents the progression as evidence that students can learn these skills much younger than schools typically expect, provided they have time to practice them. With less of the day taken up by academic instruction, schools can make teamwork and other life skills part of the curriculum rather than relying on after-school sports or occasional group assignments.

Gagan Biyani says employers he hears from are less concerned about whether graduates know a particular technical skill than whether they have maturity, people skills, systems thinking, and the ability to take responsibility for unstructured work. Those capabilities, he argues, come from working with other people, building complex things, dealing with rejection, and learning to structure one’s own time.

Joe Liemandt agrees, describing leadership, teamwork, agency, ownership, and the ability to receive feedback as human traits that become more important as AI makes some forms of intelligence easier to access. Students can develop those traits through projects, he says, but not simply by sitting in class.

The same logic informs the discussion of university, though Liemandt cautions against copying a K–12 model directly into higher education. He says Alpha’s own high school is already very different from its K–8 program, and that university requires another distinct design. He does not expect students to stop going to college; he sees an opportunity to reinvent it.

Liemandt recounts asking more than 70 interns from MIT, Stanford, and the University of Texas whether they believed their universities were preparing them for the world after graduation. One student raised a hand. The students, he says, were worried that entry-level jobs would not be there by the time they graduated and that AI would do work graduates once expected to do.

The interns spent the summer working on a proposal for an educational language model that could be made available broadly, including as a small model running on a device. They researched learning science and worked about 80 hours a week, Liemandt says. They were not handed a textbook assignment; they had to find and study the material needed to develop the idea. Liemandt says they used a data structure called a “brain lift” and worked through learning-science concepts relevant to their proposed system.

Their work combined study with the task of building something. At the end, they told him this was what they wished university classes could be: a group working together on a difficult project with meaningful potential impact. They wanted to develop an educational tool that could make learning more broadly available, and the research was part of figuring out how. The point, in Liemandt’s telling, is not that the students avoided academic work, but that the work was organized around a project they wanted to pursue.

Biyani’s point is that motivated students can learn what they need in order to build. At 18, he says, students may be more driven by making something people use, earning revenue, or getting a job than by completing coursework for its own sake. A university could use those motivations to organize learning rather than treating building as an extracurricular activity.

Liemandt’s own university experience supports the distinction. He dropped out of Stanford, but says the valuable parts were the people he met, encounters with working entrepreneurs and technologists, and building things with friends. He recalls talks by Steve Jobs, Bill Joy, and Scott McNealy, and the chance to spend time with people doing the work. Large lecture classes, by contrast, were the part he found least valuable.

He also says Alpha’s first graduating class, after entering university, reported earning 4.0 grades while finding large introductory lecture classes a waste of time. Some wrote editorials criticizing the format, he says, while valuing smaller seminars, clubs, and other activities. For Liemandt, the university is a bundle of experiences; the question is whether its formal academic component is organized around the parts students find useful.

Biyani sees the same problem from the institutional side: in his view, the academic experience is often the part the university formally provides, while much of the rest of the value is created by students. The challenge, as he and Liemandt describe it, is to make building, relationships, and preparation for real work central to the institution rather than leaving them outside the curriculum. The design for college-age learners has to start from the work, relationships, and decisions they will soon encounter—not simply reproduce Alpha’s afternoon projects.

Joe Liemandt says he has spoken with tens of thousands of K–12 parents and has been struck by how many families with Ivy League or Stanford backgrounds are looking for an alternative. Some still want a direct route into an elite university. Others worry that the path they followed will not prepare their children for the world those children will enter.

Liemandt offers the expansion of high-school education in the early twentieth century as an analogy for the choice families face. He says that in 1910 most Americans had only an eighth-grade education, many worked in agriculture, and the arrival of tractors made clear that existing work was changing. In his account, towns voted to raise taxes for high schools even though residents had never attended one themselves. His point is that families may support an educational model they have not experienced when they believe the old path will not prepare their children for the future.

The analogy reflects the choice Liemandt sees parents making now: whether to trust the route they know, or consider a different model for a world they expect to change. Parents of older students may feel committed to the existing path, he says, while parents of younger children have more time to question whether it will prepare them. He expects a new university such as the Horowitz Andreessen Academy to face the same two-part task: students may see the appeal quickly, while parents need to be shown why the alternative is worth choosing.

The obstacle is that parents are being asked to choose a model whose results they cannot yet see. Liemandt says the first families at a new school have to act before much outcome data exists. He remembers Alpha holding large information sessions while struggling to find even 25 families for a new campus. New York opened with roughly 10 or 12 students, he says; the campus later filled to about 150. Once parents could see early students thriving, demand grew. His shorthand is that people need someone to go first.

That experience shapes his advice to families considering a new institution. A new school will have hiccups, he says, and it is important that the families who enroll genuinely want what the school offers. Alpha is clear with prospective families about its approach, including its use of AI and screen time. Liemandt says that people who do not want those things should not come. The aim is to attract families who want the model and build a community around them, rather than trying to persuade every skeptic.

Gagan Biyani notes that a college-age student can try a one-year program and return to a conventional college later, making that decision different from leaving a K–12 school at a critical point. Liemandt agrees that trying a year does not necessarily close off other options. He also argues that a student’s strong desire to attend should matter to parents: even if the experiment does not work out, he says, the student will have learned from making the choice and living with it.

Liemandt says his own family took time to make such a decision. His children were in Catholic school, as he had been, and he was skeptical when Alpha’s co-founder invited him to consider the new school. He thought a school operating out of a garage sounded strange. It took him two years to move his children. He says he later regretted the delay, while his children still tease him for not acting sooner.

The story also informs his account of motivation. Liemandt says he was frequently in trouble at school, skipped classes, did the minimum needed to get into Stanford, and then dropped out. His father publicly called him a moron. Later, after Liemandt had started a company and was working extremely long hours, his father told him he had changed his mind. The difference, Liemandt says, was that his son was now doing work he loved and was fully engaged.

He sees a similar change in students who arrive at Alpha and quickly take on difficult work they want to do. Students who appear lazy or disengaged in one setting may work intensely when they have agency and a reason to care. Liemandt and Biyani’s concern is that strong grades alone do not guarantee that a student can direct themselves once external structure falls away.

Joe Liemandt separates learning knowledge from applying it. AI, he says, allows students to acquire expertise more quickly; it also gives them new capacity to build. Students can learn the prerequisites for a project while working on that project, rather than spending years preparing before they begin.

He points to Alpha students who are rebuilding the school’s Timeback software stack and becoming knowledgeable about learning science in the process. He expects students to create more of the software they learn from, including learning tools built by peers. As an example, he describes an Alpha student who developed an AP World History curriculum around a GeoGuessr-style game, with historical events and a snarky AI element the students found engaging. Liemandt says all the students who took the student-built curriculum earned a 5 on the AP exam.

He contrasts the student’s work with what the adults at Alpha could build. The game drew students into the material, he says, and the school’s adult team would not have created the same thing. Liemandt expects students to produce more learning software themselves, with peers learning through tools designed by other students.

Gagan Biyani argues that students should not choose a field primarily by asking which one will be safest from AI. They should ask what interests them. An AI-native person working in any field, he says, can use new tools to do work that was not previously possible. He contrasts that with a defensive search for a supposedly protected career and says the students he wants at the Horowitz Andreessen Academy should be willing to pursue an interest and figure out how to build in it.

Biyani uses his own work in education as an example. He says education has historically been a difficult field for founders, but that he remains proud of choosing it and finds the work fulfilling. His argument is that reaching the top one percent of a field one cares about matters more than selecting a field simply because it seems insulated from change.

Liemandt describes the corresponding educational goal as raising students who see themselves as able to shape the future, rather than waiting passively to learn what AI will do to them. Alpha students, he says, are optimistic about their ability to use AI and build. The responsibility of educators is to help them become capable of making the future they want.

Joe Liemandt expects AI to contribute to an “explosion of entrepreneurship.” He argues that large companies may become smaller because of AI, while the tools give entrepreneurs new ways to build. Yet schools rarely teach entrepreneurship directly, he says, so Alpha has developed a sequence that begins in kindergarten.

The early lessons cover earning, saving, spending, investing, and donating. Alpha uses a school currency called Alphas to introduce those ideas. Liemandt’s premise is that entrepreneurship is not a sudden flash of inspiration but a set of skills students can learn and practice.

At the high-school level, Alpha has opened Founders School for students who want to spend their time building businesses. Its guarantee is deliberately concrete: students will start a business that earns $1 million in profit while they are in high school, or their tuition will be returned. Liemandt says the school’s process is scaffolded, with students first learning to earn $100, then $1,000, $10,000, $100,000, and ultimately $1 million.

The first group had been at it only a few months when Liemandt spoke, and he said some students were moving through the milestones quickly enough that he wondered whether they would reach the million-dollar target in their first year. If they did, he said, they would move on to building something larger. The model treats business creation as a progression of capabilities, not an outcome reserved for a few unusually gifted founders.

Relationships and access to experienced builders are part of that training. Liemandt recalls a 14-year-old Alpha student cold-calling Jensen Huang. Huang returned the call, heard the student’s pitch, and gave him a brief but specific critique: the student had used the call to ask questions he could have answered through research or by talking to people in Huang’s organization. The student had shown initiative, Liemandt says, but had not prepared well. The experience became a memorable lesson in doing the work before asking for someone’s time.

For Liemandt, the value of access is not only practical advice. Meeting accomplished people can make their work seem possible. He remembers encountering prominent visitors at Stanford and realizing they were impressive but also ordinary people. That recognition, he says, can raise a young person’s sense of what they might do. He invokes the idea that a teacher’s most important job may be raising a student’s ambition for themselves.

Alpha’s projects, Founders School, and the university model discussed by Biyani all put students into work that has consequences beyond completing an assignment. Liemandt’s claim is that students can handle more than adults expect when schools give them ambitious goals, tools to learn, and room to build.