Last year, Jamie Herre and Kate Blumberg were confronted by a dilemma. Their young son, Benno, had reached kindergarten age, and it was time to pick a school for him. Yet like many other members of San Francisco’s affluent class of technologists and entrepreneurs, Jamie and Kate could not purchase for Benno the one thing they wanted more than anything else: a good public education.
In most parts of the country, this transaction takes place through the real estate market: You find a good school and buy a home nearby. The better the school, the more expensive the home. But the San Francisco public school system, which includes high-performing and struggling schools, uses a complicated lottery system for admissions that doesn’t guarantee spots in neighborhood schools. Chance, not your address, determines where your children go.
Jamie and Kate wanted to raise Benno in the city. Aware that the lottery might not go their way, they started to explore alternatives to public school. Like any city, San Francisco has a complement of traditional private schools that cater to the local elite. But they, too, can be hard to get into. So Jamie and Kate began exploring the host of education-related start-ups and experiments in the Bay Area that are based on the idea of transforming schools through the use of “disruptive” new technologies. Among them is a company called AltSchool, founded in 2013 by a former Google executive named Max Ventilla. Kate stumbled across a link to the company one day and sent it to Jamie, who was intrigued.
Backed by $33 million in venture capital from such heavyweight investors as John Doerr, Andreessen Horowitz, and Founder’s Fund, AltSchool is a for-profit corporation that describes itself on its website as “a world-class team of educators, entrepreneurs, and technologists … working together to build a network of micro-schools that offer personalized, child-centered learning experiences.” (In early May, the corporation received an additional $100 million from funders, this time including Mark Zuckerberg.) AltSchool, Jamie learned, was in its first stage of expansion, and was opening its micro-schools in a number of neighborhoods around San Francisco. The ads and brochures described in detail how AltSchool uses technology, including iPads and Chromebooks for every child—helping to prepare children, as the website puts it, “for 2030 and beyond.”
Jamie is the chief technology officer of a business-analytics company, his fourth start-up. During his career, he has accumulated expertise in a range of computer programming languages. He is about as steeped in the theory and practice of IT as one can imagine. When he saw pictures of AltSchool students using all those iPads and Chromebooks, and thought of little Benno using them, his reaction was visceral.
He hated it.
Jamie knew that tech start-ups can be dynamic and innovative, but he also knew how vulnerable they are to utopianism and fads. And deep down, he wanted Benno to be educated by a person, not an iPad.
But competition for school spots in San Francisco is fierce, and Jamie wanted to keep all options on the table. So he went to an AltSchool open house and met Ventilla, who quickly allayed his fears. AltSchool, Ventilla convinced him, was developing a sophisticated way of delivering more personalized education to children, not less.
Using technology to give people exactly what they want and need is the great obsession of modern high-tech companies in Silicon Valley. Google customizes your search results based on who you are, where you live, and how you’ve searched before. Netflix mines a vast database of movie-watchers to anticipate exactly the film or TV show you’ll most enjoy next. Amazon does the same for books. Increasingly, Facebook algorithms determine which news stories people read. Uber brings a car to wherever you are.
Ventilla was primed to see the potential of personalization for education. He had founded a social-network search company called Aardvark in 2007, sold it to Google in early 2010, and then worked on Google’s forays into social networking, spending almost two years as the company’s head of personalization. His job was to personalize the Google experience by marrying the company’s vast search database with information from the Google Plus social network and publicly available databases such as airline schedules, traffic, and weather. Based on real-time traffic databases, a Gmail receipt for airline tickets, and a GPS-determined physical location, for example, Google could alert somebody if they were in danger of missing an upcoming flight.
While Ventilla was working at Google, he was also busy raising a young family. He and his wife knew they would soon be confronted with the same dilemma as Jamie Herre: They wanted to stay in the city but, despite their prosperity, couldn’t guarantee their kids a good education in its public schools. So Ventilla did something that’s actually kind of normal in Silicon Valley’s bubble of money, hubris, and desire to change the world: He went looking for venture funding. If he couldn’t find a school to give his kids the personally tailored education they needed for the 21st century, he would create one himself.
There are about 50 million children enrolled in American public schools. We can’t hire 50 million adults to teach them, so instead we’ve designed our schools to facilitate a relationship of one-to-many—on average, 20 to 23 students per classroom, although the number can be much higher in poor neighborhoods.
The overwhelming challenge for any teacher faced with a room full of 23 students is how different they are from one another. Children come to school from different home lives and cultures, and in the classroom demonstrate different levels of socialization and emotional development. Their interests and motivations aren’t the same; a book or film that resonates with one might utterly fail to connect with another. Their skills and natural abilities can vary widely, too: A room full of third graders might include some children reading at the first-grade level and others at the sixth. Their respective fluency in mathematics might be the reverse.
The best education meets students where they are. If a child is reading at the eighth-grade level, she needs to be guided and prodded toward the ninth, regardless of how old she is. If she struggles to stay on task and manage frustration, she needs a teacher who understands that and reacts accordingly. If she has certain passions—robotics, the Paleozoic Era, basketball—she’ll benefit if those interests are used as vehicles for learning to read, write, and think.
Nobody seriously debates any of this. Parents with money overwhelmingly choose to buy personalized, supportive learning experiences for their children. But it’s very difficult to deliver customized learning to many children simultaneously. Historically, the only way for a school to pull that off has been to spend a lot of money hiring unusually good teachers to lead unusually small classes. Most schools can’t afford to do that. Instead, they fall back on process and procedure, dividing a child’s educational life into yearlong chunks, so that eight-year-olds are all taught “third-grade reading,” for example, which is defined as being somewhere near the midpoint of where eight-year-olds tend to be. The instruction itself is often mechanical. At best, teachers modestly differentiate lesson plans and curricula that are essentially the same; at worst, they broadcast a uniform set of facts and instructions from the front of the room.
The educational shortcomings of this model accumulate over time: Older students suffer its consequences even more than younger ones do. Most teachers and school leaders understand this, but they haven’t had available to them a cost-effective means of solving the one-to-many dilemma. That may at last be changing, thanks to recent developments in information technology.
There’s nothing new about using information technology in the classroom, of course. Written words are a kind of technology; so are printed books. Two generations ago, teachers used ditto machines and A/V equipment to improve education. By now, most public school classrooms in this country are wired for the Internet, and many feature some combination of computers, tablets, and expensive interactive whiteboards that can project video or educational software onto a screen. Pretty much all of these technologies have been deployed for the same purpose: to bring children into contact with knowledge and wisdom that doesn’t reside in their teachers’ minds.
But while all this information technology has vastly expanded access to information, a major challenge remains: How do you figure out what kind of information each child needs, when, and in what way?
Not surprisingly, Silicon Valley is home to an unusual number of schools devoted to solving this problem. Rocketship Education, for example, is a chain of charter schools based in San Jose that has been successful using a blend of computer-and teacher-based education to serve low-income students. The Silicon Schools Fund, for its part, provides “venture philanthropy” funds to schools that use “innovative educational models and technology to personalize learning.” Recipients tend to be public or non-profit charter schools that manage costs by combining teacher-led classes with what might be called algorithmic personalization: computers gauging the strengths and weaknesses of individual students based on their success in completing computer-based work, and customizing future schoolwork accordingly.
AltSchool differs from these organizations in several ways. First, it has grand ambitions to grow into a nationwide for-profit educational network. Second, despite a multimillion-dollar investment in research and development, and a faith in large-scale data analysis straight out of Google, AltSchool believes that educational personalization in the future will ultimately stay where it has always been: in the hands of teachers.
AltSchool’s campus near Fort Mason, in San Francisco, consists of a half-dozen classrooms. They look ordinary. The younger children spend their days moving about on floors furnished with soft blue-and-white circular floor mats, surrounded by tastefully Nordic shelves full of books and educational toys. They’re watched over by adults speaking in the professionally soothing tones used by any good elementary school teacher. Granted, there’s a little more computer hardware than you might expect: A fleet of iPads lives on a wall rack next to a row of cubbies and pinned-up drawings of houses and trees, and the teachers pull out their smartphones a lot, tapping and typing for a few seconds and sometimes flipping the screen around for students to see. But it’s only when you look beyond the classroom walls that the scope of Max Ventilla’s vision starts to become clear.
That’s because on the other side of the walls there is none of the apparatus we typically associate with school: no cafeterias or gymnasiums or assistant principals’ offices. The whole place sits inside a two-story commercial space on Buchanan Street, flanked by a Starbucks and a small office building, steps away from Marina Boulevard and the San Francisco Bay. Before AltSchool took the property over, it was a gym.
All AltSchool campuses are small, with only six to eight classrooms and 60 to 100 students, who range in age from five to 13. Food and janitorial services are outsourced. All of the central administration takes place at the company headquarters. In a way, an AltSchool micro-school functions like the search box that appears on Google’s homepage: Both are minimal and offer users only what’s necessary for them to access a vast and invisible background network of information that can be tailored to their own needs.
Because students of the same age can vary greatly in their academic development, and because its classrooms are small, AltSchool places students of different ages together. Instead of yearlong grade levels, students are grouped into broad ranges—lower elementary, upper elementary, and middle school. There are no enrollment cutoff dates that make children born on October 1, say, have to wait an entire year to start educational programs available to children born just a day earlier, on September 30.
Personalization is the name of the game. At the time they enroll, students are evaluated in a variety of core subject areas and assigned a unique curriculum based on their teachers’ assessments of who they are and how they are most likely to move ahead in different subjects. Both the curriculum and the teachers’ evaluations evolve constantly. The key to managing the process is a technology platform built around a set of “cards,” personalized for each student, that can be displayed on a computer, tablet, or mobile phone, and accessed by students, parents, and teachers at any time. Each card is about the size of a standard tablet screen and displays a particular educational goal, along with a list of assignments the student needs to complete in order to meet the goal. Many of the goals derive from the Common Core Standards—a set of ambitious educational goals, established by a consortium of state education officials, that have been adopted by 43 states and the District of Columbia.
In any given week, students have 15 to 25 personal cards active in the system, and these determine what they study. As students do their assignments and achieve their goals, teachers review and grade their work, electronically attaching teacher feedback and evaluations, photos, and progress reports to the cards. Completed cards are filed away as new ones are created and used, gradually building a database of exactly what each student has been working on and learning. It’s like the mythical Permanent Student Record, powerfully enhanced.
That database consists of more than just students’ cards, however. Bolted onto the interior walls of every AltSchool classroom, halfway between the ceiling and floor, are a series of transparent boxes, each housing a mass of black, blue, and red wires that connect a wireless router to a small HD video camera and a professional-grade microphone. These capture the sights and sounds of the classroom in a way that can be accessed by a team of software engineers at AltSchool’s headquarters. Next year, similar spaces will open up in Brooklyn, Palo Alto, and elsewhere in San Francisco, all collecting terabytes of audio and video footage.
The audio and video data help teachers manage the problem of one-to-many. In one lower-elementary class I visited, the students watched a video of the three little pigs and then broke into four small groups to come up with a plan for building their own wolf-proof house. The teacher moved from group to group, listening and helping. But she was particularly concerned with the progress of one five-year-old, who had recently joined the class and was struggling to collaborate with peers. Normally, a teacher in this situation has to make a choice: spend the whole time observing the new student’s group, neglecting the others, or miss out on important information. This teacher was able to go back later in the day and watch exactly how the new students interacted with other children for the whole class period, both with an adult present and when the students were alone.
At AltSchool headquarters, meanwhile, staff use the data to see how space is being utilized so the designs of future micro-schools can be improved and refined. The system also provides a wealth of information to review how teachers spend time in the classroom. AltSchool also has some more speculative ideas: Perhaps the database could automatically record each time a student’s name is spoken out loud, and how often the context is positive or negative, to see who is receiving more attention and why. Facial-recognition software could allow teachers to track individual student movements, which might be useful for understanding group dynamics. Perhaps the lighting could be set to automatically dim to soothing levels when the volume of chatter gets too high.
Or maybe they’ll learn something else entirely from the audio and video data when it’s all combined with the growing file of student cards filed in the database. AltSchool has plans to gather more information than any human being can process manually, but we’ve arrived at a point in history when computers routinely search through huge data sets for patterns and relationships that people would never have thought to look for. This is the process of machine learning, a pillar of modern artificial intelligence theory.
Technology has changed so quickly, it’s easy to forget that none of this would have been feasible even 10 years ago. The AltSchool personalization model depends on constant high-speed wireless connectivity, infinite cloud-based data storage capacity, and powerful, inexpensive computer hardware running software that normal people can use with intuition and ease. Once those building blocks were in place, it became possible to design an educational organization from the ground up whose teachers are authentic knowledge workers, independent and empowered, instead of what they often end up being today: cogs in a creaky machine.
One day last October, I visited the Fort Mason campus and spent some time in the classroom of a teacher named Paul France. He and a co-teacher are responsible for 12 kids, ranging in age from seven to 10. France is in his mid 20s. He spent four years teaching in the Chicago suburbs before moving to the Bay Area for this job.
Throughout the day, France guided his students back and forth between individual and group work. Sometimes they all came together for a group discussion or to listen to Paul talk. At other times, they sat on stools and worked on white workbench-style tables, each with room for a couple of laptops or a pile of books and colored paper. The day’s schedule was written in marker on the wall, next to bookshelves, paintings, a Lego model, and an old-fashioned globe. In the next room, a group of 12- and 13-year-olds slouched on beanbag chairs and explained, in the manner of every adolescent ever, that nobody understands them.
At the time of my visit, the students in Paul France’s class were working on math concepts such as place value, computation, and number sense, at different levels of mastery. At a broader level, they were learning about form and function: How they can relate to patterns in mathematics, how the shape of tools found in an archeological dig can help determine what they were used for, how a text can be structured to explain something with cause and effect. Later, the teachers and students all went outside together to walk around the Marina District and use their iPads to take pictures of natural and man-made forms.
France has a Millennial’s practiced ease with mobile information technology. He can’t imagine doing his job without it. Every 10 or 15 minutes, I saw him pull out his phone, work on it for a few seconds, and then put it back in the pocket of his cardigan sweater. Whenever he needs to during the day, he explained, he can take a picture of a student’s work, tag it with their name and the card it relates to, and store it in the AltSchool database—all in about 20 seconds. “If they saw me using my phone this much in Chicago,” he told me, “they’d assume I was goofing off.”
Last year, France went through the highly respected National Board for Professional Teaching Standards certification process, which emphasizes an iterative, diagnostic approach to teaching—know the child, set an objective, teach, assess, set the next objective. Managing the complexity of that program for many students at once, as France now does in his classroom, would be impossible without the team of software engineers back at AltSchool HQ, who are in constant dialogue with the teachers, making the experience of creating and managing cards, capturing and assessing evidence of learning, and communicating with students and parents as swift and seamless as possible.
Carolyn Wilson, AltSchool’s director of education, told me that the company has more software developers on staff than does the entire Los Angeles Unified Public School District, which spends over $7 billion a year to educate more than half a million students. When Ventilla opened up his first one-room school, in 2013, as a kind of prototype, it took his teachers three hours per student per week, outside of class, to manage the card sets. Now, after practice and a series of software improvements, they’ve got it down to 30 minutes. This means that France can spend almost all of his classroom time giving students his personal attention: Supervising a few kids as they build a model drone, helping a girl who loves dogs to design an independent reading project, talking to a boy who likes to roll dice and see how often different number combinations come up. And he can do all of that knowing that he can log into the database later to catch up on what other students did during the day—and still have time to sleep at night.
This is not school as we all grew up knowing it. And that’s the point. “The purpose of school,” Ventilla told me, “is to prepare kids for the world that they’re actually going to inhabit as an adult. And that world is continuing to change at an accelerating rate. … You need something that doesn’t just create an updated most-common-denominator approach, but that really varies the experience for all the kids and all the families and all the teachers, not just within different locations, but even within the same classroom. When you think about it from a research standpoint, it’s something much more akin to personalized medicine than traditional blockbuster trials.”
Not surprisingly, most of AltSchool’s investment to date has gone toward establishing its central infrastructure: software engineers, data analysts, and administration. Right now, the organization is top-heavy, but the plan is to change that soon by opening lots of new micro-schools. This is basic franchise economics, but it’s never really been applied to place-bound for-profit K-12 education. Edison Schools was launched with much fanfare in the 1990s and reached a billion-dollar stock-market valuation before collapsing in 2001, but its model required entering into contracts to run public schools with public money, which embroiled the company in urban politics and contentious labor disputes. Another company, K12 Inc., serves tens of thousands of students, but its schools exist primarily online.
Ventilla believes the AltSchool network can grow to encompass hundreds of thousands of students, or more, some enrolled directly in the company’s micro-schools and others in schools connected to the AltSchool network. AltSchool’s early-stage expansion into Brooklyn and Palo Alto, already well under way, will be followed by a move into other “creative class” cities, full of well-educated parents with progressive values who can afford annual tuition that runs from $20,000 to $28,000, depending on the location. After that will come greater expansion, all across the country—not just to other cities but also to rural areas, where assembling enough students to fill a standard grade school can mean long car trips to and from school every day.
Ventilla believes he can create a school network that gets better as it gets bigger. Each new school, teacher, and student, the theory goes, will add to the collective intelligence and value of the organization. The audio and video, the completed card sets, the evidence of learning carefully tagged: All of it will cohere into a mass of information that can be mined for organizational insight—how to better hire and support teachers, how to design curricula and classrooms, how to build momentum in student learning over time. This knowledge and wisdom will be valuable for both AltSchool and other school systems and educational organizations, which, Ventilla believes, will pay for access to it. It’s the educational equivalent, in other words, of what Ventilla helped create as the head of personalization at Google.
A teacher-support system that gets better as the AltSchool network gets bigger is a critical part of the company’s expansion plans. Some of the most successful public charter-school networks, among them the Knowledge Is Power Program, seem to depend on a steady supply of highly motivated, unmarried teachers in their mid-20s who work 90-hour weeks in low-income neighborhoods until they burn out. Building a million-student school network using that model will never work. But Ventilla believes that as the AltSchool network grows and improves, he’ll be able to hire regular teachers, with lives and families, and will help them achieve the kind of deep educational personalization that has historically only been possible under rare circumstances reserved for the wealthy.
There’s no guarantee that AltSchool will succeed, of course. The upside to venture capital is that you get enough money to hire talent and build the infrastructure for potential greatness. The downside is that the ticking clock of return on investment forces you to grow quickly—perhaps too quickly. The pressures of expansion can buffet and strain the most well-run organizations, including those whose goals are much less ambitious than solving the basic one-to-many dilemma that has plagued schools for centuries. Ventilla is also competing in a market in which the government subsidizes public schools for about the same amount of money per student, roughly $13,000, that he’s eventually hoping to charge. The playing field isn’t exactly level.
That said, AltSchool’s investors are the same people who saw the promise of companies like Google and Facebook early on, and, unlike many tech start-ups, AltSchool provides a service that people genuinely need and are willing to pay for. A great deal of the money that goes toward schooling today is embedded in mortgage payments for expensive homes that parents bought in order to be near good public schools. Ventilla hopes to pry some of those billions loose.
The success or failure of AltSchool, which at this point serves only a small group of mostly affluent Americans, might not seem relevant to the majority of families in this country, who rely on public education. But innovation often develops first in a market for wealthy people and then spreads outward, as processes are refined and economies of scale are realized. You don’t see a lot of people today driving Tesla automobiles, which can cost almost $100,000. Thanks to Tesla’s innovations and appeal, however, it’s a safe bet that 10 years from now electric cars will be better, cheaper, and more widely used. Ventilla believes that AltSchool can bring about a similar sort of change by fundamentally re-defining the relationships between teachers, parents, students, and schools.
Students tend to be unreliable and uncooperative witnesses to their own learning. What did you learn in school today? Nothing. Really? For six hours, nothing? I don’t know, Dad! Stuff!
Schools aren’t much better, partly because it’s logistically perilous to communicate with many sets of parents about something as complicated as learning, but mostly because in all likelihood the school doesn’t know how your kid’s day was, either. How could it, given the limitations of one-to-many? Why would it, since the education was the same for everyone? Here’s the curriculum; that’s all we know.
So we enter into uneasy relationships of blind trust with our children’s schools, searching for peer testimonials, hyper-alert to neighborhood gossip and reputation. We hedge uncertainty by surrounding our kids with peers from the “right” kind of families, wanting and needing to believe that the local schools are just fabulous—because if they’re not, what kind of parents are we?
The hyper-transparency of the AltSchool model, with every lesson and piece of evidence tagged and available for instant perusal on your screen of choice, creates the potential for something different, a way of educating together that requires more work and less faith, but is also less susceptible to the social distortions, market failures, and mutual resentment that low-information systems tend to create. If it works, it could help liberate schools and teachers from the burden of failing to accomplish a fundamentally impossible task, and give children the kind of education that, regardless of who they are and where they come from, they really need.
That’s why, despite having their son, Benno, accepted by a number of perfectly respectable private schools, Jamie Herre and Kate Blumberg ultimately decided to enroll him in one of AltSchool’s San Francisco micro-schools, where he’s now happy and thriving. “Of all the schools we visited,” Jamie told me, “they were the ones who really listened to what Benno had to say.”
Lead photo: The rise of the omniscient classroom. (Illustration: Brian Stauffer)
Submit your response to this story to letters@psmag.com. If you would like us to consider your letter for publication, please include your name, city, and state. Letters may be edited for length and clarity, and may be published in any medium.
For more from Pacific Standard on the science of society, and to support our work, sign up for our email newsletter and subscribe to our bimonthly magazine, where this piece originally appeared. Digital editions are available in the App Store (iPad) and on Zinio (Android, iPad, PC/MAC, iPhone, and Win8), Amazon, and Google Play (Android).
