Can Silicon Valley really hack education?


In chalets scattered across the snow in California’s ski country, a school of the future is taking shape. Warm inside a classroom, teenage twins Laurel and Bryce Dettering are part of a Silicon Valley experiment to teach students to outperform machines.

Surrounded by industrial tools, Bryce is laying out green 3D-printed propellers, which will form part of a floating pontoon. The 15-year-old is struggling to finish a term-long challenge to craft a vehicle that could test water quality remotely.

So far, the task has involved coding, manufacturing and a visit to a Nasa contractor who builds under-ice rovers. “I suck at waterproofing. I managed to waterproof one side, did a test of it, it proved waterproof. I made sure the other side was waterproof, put both sides on and both of them leaked!” he laughs.

Laurel, already adept in robotics, chose a different kind of project, aimed at developing the empathy that robots lack: living on a reservation with three elderly women from the Navajo tribe. “The experience was just, honestly, it was really . . .” she trails out, her navy nails fiddling with her dark-blonde hair. “They didn’t have running water, didn’t have electricity, they had 54 sheep and their only source of income was weaving rugs from wool.”

The Detterings have embraced personalised education, a new movement that wants to tear up the traditional classroom to allow students to learn at their own pace and follow their passions with the help of technology. Mark Zuckerberg, the founder of Facebook, and his wife Priscilla are leading the push to create an education as individual as each child, aiming to expand the experiments beyond the rarefied confines of Silicon Valley.

Bryce Dettering, 15, alongside a milling machine at the Tahoe Expedition Academy in Kings Beach, California, where he is a pupil © Carlos Chavarría

Tahoe Expedition Academy uses software developed by Facebook and Summit Public Schools, a free charter chain 200 miles south in the Bay Area. Every morning, Laurel and Bryce log on to their “personalised learning platform”, which looks like a website, and progress through a “playlist” of reading material, videos and tests.

They decide which modules to learn next based on what they enjoy or have yet to master. By focusing on what they need and want to learn, rather than following a class-wide curriculum, the platform frees up time for additional projects that encourage taking risks and solving problems.

Having disrupted the world, the tech community now wants to prepare children for their new place in it. Leading venture capitalist Marc Andreessen predicts a future with two types of job: people who tell computers what to do, and people who are told what to do by computers.

Silicon Valley wants to equip young people to rule the machines by focusing on what makes them individuals. But how far can this reinvention of learning be extended from the wealthy environs of northern California to the broader US education system, where some state schools struggle to provide up-to-date textbooks, let alone personalised, digital tutoring?

The Detterings’ parents signed up for an alternative education after Laurel became frustrated at her private girls’ school in San Francisco. She was ahead of her peers and not content with drawing on her shoes — which is what her mother had done when she was bored in class. “I was really good at looking like I was listening and dreaming in my head,” Sue Dettering smiles. “This generation does not tolerate that very well.”

Tahoe Expedition Academy combines academic teaching with what it calls “constructive adversity” — adventures that push kids to the edge to build character. Like an endless educational gap year, each high-school senior has spent 130 days away in the past three years.

One group went to Greece to work with Syrian refugees, while a class of 13-year-olds drove and kayaked to the Mexican border to interview border patrols and immigrants and see how Donald Trump’s wall could shape the region.

Tahoe pupil Laurel Dettering; part of Bryce Dettering’s pontoon, with a 3D-printed propeller © Carlos Chavarría
Bryce’s blueprint for his pontoon © Carlos Chavarría

“Rote learning is done — computers can do that,” says Sue Dettering. “The kids are going to have to have the interpersonal skills to work in groups, to communicate well, be creative, arrive at an answer in many different ways.” Their father Bill, who runs a tech company, says other parents think they are brave — but believes this new education is sensible. “It really is the human-oriented jobs that are going to be the opportunity.”


The uncertain future facing the next generation of workers is partly Silicon Valley’s fault. Tech companies have stripped out jobs as they transform industries, from retail to media to cars. Artificial intelligence will accelerate the shift, with Accenture predicting that AI will increase labour productivity by up to 40 per cent by 2035, when Bryce and Laurel turn 33. Fewer workers will be needed for the jobs we have now, so kids must be prepared for the jobs we cannot imagine.

Silicon Valley is, as ever, optimistic. It wants to move on from a 19th-century, artisanal model of education — where knowledge resides with each classroom teacher — to a 21st-century personalised experience that technology can replicate on a global scale. The new model focuses on skills, not knowledge you can Google, and social abilities that will be needed, whatever the workplace.

Mark Zuckerberg and his wife, Priscilla Chan © Reuters

When Zuckerberg and Chan’s daughter Max was born in 2015, they announced they would donate 99 per cent of their Facebook shares, worth $45bn at the time, to the Chan Zuckerberg Initiative (CZI). Penning an open letter to their firstborn, they focused on the prospects for personalised education.

“You’ll advance quickly in subjects that interest you most, and get as much help as you need in your most challenging areas. You’ll explore topics that aren’t even offered in schools today. Your teachers will also have better tools and data to help you achieve your goals,” they wrote. More than a year later, Zuckerberg is posting Max’s toddling attempts on Facebook to his 85 million followers, and CZI is looking to make education investments.

Personalised education is not new: in 1926, Sidney Pressey created a “teaching machine”, where students could read material and complete multiple-choice tests at their own pace, pulling down levers and receiving sweets for correct responses.

The personalised education movement combines a testing machine for the big-data age with a key idea taken from Maria Montessori, who developed her approach more than a century ago: that each child should drive their own learning.

Silicon Valley hopes new technology will help personalisation finally succeed at scale. Today’s teaching machines look like data dashboards, which teachers monitor as each child works through tasks and tests.

This instant data speeds up teachers’ decisions — they don’t have to wait until a child hands in homework or for the results of an end-of-term test. In class, they can spot who needs help and assemble impromptu tutoring groups, while others steam ahead, taking on honours units in subjects they love. A moving line on the student’s dashboard shows the term ticking away, prompting them to stay on top of units where they are slower. Final tests are only taken when a child is ready.

The problem Silicon Valley is trying to solve can be summed up by a 1984 study by educational psychologist Benjamin Bloom. Cited by every personalised-education advocate including Zuckerberg, it found students who received one-on-one attention performed better than 98 per cent of their peers. The challenge is to achieve the benefits of that one-on-one tutoring in a class of 30 or more.

An external view of the Tahoe Expedition Academy, Kings Beach, California © Carlos Chavarría

Personalisation is so popular in the Bay Area that parents can pick their experiment: public, private or funded by non-profits. The Tahoe Expedition Academy is a private school, charging up to $17,000 a year, with a focus on the outdoors.

The AltSchool combines a start-up filled with engineers and product managers, funded by CZI and venture capitalists including Andreessen Horowitz and Peter Thiel’s Founders Fund, with a chain of private “lab schools”, where parents pay about $27,000 a year.

The Emerson Collective, established by Steve Jobs’ widow Laurene Powell Jobs, has invested in existing schools, including one for low-income students in San Jose, where some high-schoolers work a day a week as product managers at Cisco. Chan, a paediatrician and teacher, started The Primary School, combining education and healthcare for an underprivileged community in East Palo Alto.

Many of these private schools want to contribute their technology and lesson plans to state schools now or in the future. But the original personalisation Petri dish was Summit Public Schools, a charter chain that has germinated from one school in Silicon Valley to a learning platform that is now used by 20,000 students across 27 states.

Summit Tamalpais, the chain’s newest school, opened last August in Richmond, the poor, post-industrial and heavily Hispanic end of the Bay Area. Driving up to the school, opposite a grey-box Walmart, my car is flagged down by a woman begging for money for a motel room for her and her three-year-old.

Inside, the children are rambunctious, celebrating break time as if it were a birthday. Self-portraits line the walls, with each 12-year-old illustrating their identity inside silhouetted profiles: one child has painted a pizza, a French flag, YouTube and a fighter plane; another Lake Tahoe, September 11 and a rainbow.

When kids return to class, a Summit teacher’s job is to turn these interests into a personalised curriculum that keeps every child engaged in learning through class, college and career.

Technology lies at the heart of Summit’s approach to learning; Diane Tavenner, founder and chief executive of Summit Public Schools © Carlos Chavarría

Diane Tavenner, the founder and chief executive of Summit, believes technology addresses this challenge. As a child in the 1970s, her progressive teacher experimented with different levels of spelling tests, distributed on cards in wooden pigeon holes. But the teacher gave up because it was impossible to mark each test at the right pace. “We went back to the same spelling book and did the same list. It wasn’t really self-directed,” she says.

Summit’s personalised-learning platform uses technology to make these tasks easier. At Tamalpais, Dajana, 12, is preparing for a debate on whether Roman society was just. After finishing the reading and videos in her playlist, she decides that it was hugely unjust, because men could sell their wives and sons.

She uses a traffic-light system to assess what she needs to do next: she is “yellow”, needing more practice ahead of her speech at the debate. “It is harder because we are doing it by ourselves, but it is good practice for the future,” she says.

In a silent neighbouring classroom, where each 12-year-old works on his or her own computer, Fernando Torres’ screen shows he has completed most of the core courses, depicted by coloured blocks, and several advanced units.

“I find the easy ones and then I start on the hard ones after. I’m really excited by science, so I did the science ones first,” he explains. Next week he’ll spend every morning working on a video-games project.

Rote learning is done — computers can do that. Kids will need to be able to work in groups, be creative

Sue Dettering, parent

Zuckerberg hails Tavenner as a personalisation pioneer. Live streaming on Facebook from his couch on CZI’s anniversary, he spoke excitedly about how Chan, sitting next to him, urged him to visit the school. “You go in and it feels like the future, it feels like a start-up to me,” he says.

To hear about life pre-Zuckerberg from Tavenner and the other half of the Summit double act, engineer Sam Strasser, it sounds like the tech was held together with sticky tape.

Strasser, then the only engineer, tried persuading friends to help build the software: “That didn’t work out logistically. They have day jobs, they can’t just do your job for you,” he grins. Zuckerberg lent them the technical talent. “There were three of us — which is a tiny, tiny team at Facebook but was 300 per cent growth for me. So I was like, this is huge!”

Summit plans to expand to hundreds more schools. Tavenner now meets with Zuckerberg about once a month, Strasser works at Facebook, and the senior management team spend every Monday there.

Everyone contributes ideas to develop the platform, even the kids, though they don’t have the direct line they had when Strasser was in school. “They literally would, like, chase him through the school, [shouting] ‘Hey, hey, hey — Mr Strasser, we have an idea!’” laughs Tavenner, mimicking the children’s high-pitched voices.

The day before we met, Tavenner discovered another assumption she had yet to question. “Yesterday afternoon we stopped and said: wait, why do we assume that every teacher has to grade their own kids’ performance tasks?” She points out that the dual coaching and testing role of a teacher can potentially create grade inflation, as teachers fail to be objective with students they know. “With technology, it literally doesn’t have to be true,” she says. Tests could be sent instantly to be marked by another teacher in the school or even elsewhere.

Tavenner’s next challenge is more political: convincing universities of the value of Summit’s own data, so her students no longer have to sit standardised tests. She is pushing deans of admissions to abandon the current “hoops for college” in favour of a more detailed, nuanced picture of a child.


At the AltSchool in San Francisco’s start-up district, a few streets away from the offices of companies such as Dropbox, Pinterest and Airbnb, four-year-olds are advised to be “mindful” as they cut up painted paper for a puppet show; the project was devised by one of the kids in class.

Students at Summit Preparatory Charter High School in Redwood City, California © Carlos Chavarría

One floor up, 12- to 14-year-old children run election campaigns, designing websites and playing TV anchors on video. On another floor, a teacher photographs children listening to an expert on how to design their own Olympic stadiums; the picture is uploaded to AltSchool’s app to share with parents instantly.

Ernesto, 12, is passionate about cooking: he runs a website making $20 cakes to order and argues with his teacher about the best kind of Italian baking chocolate. Given a project where he had to prove or disprove a myth, he tested his hypothesis that drying pasta at a high heat caused it to lose its taste.

He made an entertaining YouTube video about the experiment, including a cameo from a local artisanal pasta producer, concluding that high temperatures did indeed hamper flavour. “I got better at video editing and coming up with an experiment, following the scientific process,” he says.

AltSchool’s classrooms are considered “labs” because the teachers and engineers are hoping to spread what they learn here beyond the chain in Silicon Valley and New York. Like tech inventions from the iPhone to Uber, the personalised-education movement intends to go global.

For funders, the size of the market is both tempting and terrifying. Josh Kopelman, a venture capitalist at First Round Capital, says its investment in AltSchool is a financial, not philanthropic decision. “Would I rather fund this or another photo-sharing app? There’s no question for me,” he says.

There’s a bit of Silicon Valley hubris here — we’ve arrived, and now we’re going to create all these amazing changes

Trace Urdan, education market analyst

He holds forth on the vastness of the education industry: “It’s an industry that is measured in trillions of dollars, not billions; it’s multiple percentage points of gross domestic product.

“The consumption of this product is required for a meaningful portion of our population. And you have tens of thousands of fractured competitors that all spend collectively less than 1 per cent of their budget on R&D and don’t innovate in terms of process and software.”

But Jim Shelton, charged by Zuckerberg and Chan to invest in personalised education, believes that even with billions, it will take clever decision-making to change such a large industry.

“In philanthropic terms, these are a tremendous amount of resources. But in the US alone we spend $650bn in K12 [primary and secondary] education every year, so the investments we make will need to be very strategic.”

Shelton has the pedigree: he was deputy secretary in the US education department, directed programmes for the Bill & Melinda Gates Foundation and worked at education technology companies. Chatting with Zuckerberg and Chan during their Facebook Live broadcast, it is clear they regularly swap notes.

“It is something we are very much doing together,” he says. “Priscilla is there engaged daily — and nightly — and Mark is there a surprising amount. They want to understand deeply what is going on and they have perspectives on most of what we do.”

The couple learnt a hard lesson from their first foray into education philanthropy — a donation of $100m to schools in Newark, New Jersey in 2010. The effort, led by Cory Booker, then mayor of the city, and Chris Christie, the Republican governor, failed to win over a wary community.

Dale Russakoff, who wrote a book called The Prize about the turnaround attempt, says Zuckerberg did not realise the scale of the challenge he faced. “He was 26 when he made that gift, Facebook was young and he was still called the ‘Toddler CEO,’” she says. “I think he learnt a lot about politics and just how hard it is to change an education system.”

CZI will focus on what Zuckerberg knows well: using technology to personalise, and investing in research and development. But it plans to progress more slowly than the social network, investing in the US before expanding globally and, says Shelton, being sensitive to community needs, perhaps taking on a “different flavour” for local cultures.

Fernando’s screen shows he has finished the core classes. Next week he’ll work on a video-games project

Shelton believes personalised learning can grow where other educational movements have failed: “Ideas are easy to spread, [but] practices are very hard to spread with fidelity, especially when they are not supported by good tools.”

The key is to create demand. “Stimulate it in a way that is not so cumbersome for learners and teachers to implement. Then, when they try it, they can’t imagine living without it. It’s how the best innovations work.”


While Shelton uses the word “humble” to describe Zuckerberg and Chan’s approach, industry analyst Trace Urdan describes Silicon Valley’s push into education as “hubris”. Established education companies such as McGraw Hill and Pearson have been working on adaptive learning technologies for years, and younger ed-tech companies have sometimes found they are more excited about data than their potential customers.

“There’s a little bit of Silicon Valley hubris here: ‘We’ve arrived, we’re here and now we’re going to create all these amazing changes,’” he says.

The Bay Area is also an unrepresentative test market, filled with millennial parents “steeped in large amounts of wealth”, Urdan says. And the enviable student-to-teacher ratio in some of the private schools experimenting with personalisation (Tahoe Expedition Academy has a 7:1 ratio overall) is far from the norm across America.

The cost of technology is another challenge: a survey by the Consortium for School Networking found that only 68 per cent of US districts currently have schools that meet even the minimum internet bandwidth recommendations set by the regulator. And few kids outside of Silicon Valley can ask their parents’ friends to help with virtual reality projects.

Expanding beyond northern California could be tough, he argues. Technology is usually sold by district to superintendents who frequently move positions, often replaced by someone who brings in different software. Once they get beyond early adopters, unionised teachers are “very suspicious” of anything that could replace jobs. Russakoff says it will take a “lot of engagement” to convince unions that the technology will make teachers more powerful. “It will not be an easy road.”

Howard Gardner, a professor of cognition and education at the Harvard graduate school of education, cautiously welcomes the tailored approach, but fears the entrepreneurs could be exercising a certain “noblesse oblige” by advocating the organisation of one’s life around one’s personal passions or interests — something that is not always possible for everyone or their children. “It is the Bill Gates fallacy: you think everyone is a little Bill Gates,” he says.

Inside an AltSchool classroom or ”lab” © Carlos Chavarría

Silicon Valley parents are probably less concerned about privacy concerns than many and, so far, privacy campaigners have not slammed the most widespread tool — the Summit Personalised Learning Platform — despite it working with Facebook, which stands accused of not protecting users’ privacy in the past. Instead, they give it good marks for asking parents’ permission at the start of each school year and storing the data on their own servers.

In Europe, personalised education may face resistance because privacy is considered a fundamental right. Everyone from marketers to credit scorers to governments are thirsty for data: a picture of an individual built up over years, tracking their test scores and what they read and watch, could prove deeply attractive.

“Privacy issues are coming,” says Elana Zeide, a scholar working on education, big data and privacy at Yale and Princeton. Critics have so far focused on problems with software vendors — for example, whether student data can be sold off after bankruptcy. But they have not yet considered the new data-driven model of education in schools. “The shift will cause a jolt,” Zeide says.

Technologies are certain to pull in more data in the future. At AltSchool, cameras mounted on classroom walls are used by some teachers to review how their lessons are going down. Elsewhere, researchers are using facial recognition software to work out when a learner becomes frustrated or when the whole class switches off halfway through a lesson. Algorithms risk reaching a point like the Facebook news feed, where few understand why information is selected, argues Zeide.

“Algorithms can be unwittingly biased, can have inadvertent, disparate effects, and those kinds of perpetuated inequalities are particularly sensitive in education,” she says. Today, people are increasingly worried about filter bubbles — but what if your child is kept back a grade and no one knows why?

The bigger question is how much power this hands to Silicon Valley companies that already dominate our work and social lives. “The tradition of localised control is because people in the local community know the needs of that economy or what is consistent with local values,” Zeide says. “If instruction is based on a platform developed by engineers in Silicon Valley, parents and school ­officials don’t have the same control.”


Max Ventilla left his job as former head of personalisation at Google to pursue another Google-sized challenge. The AltSchool chief executive sits in a start-up-style office on the top floor of the school, and draws on the whiteboard when excited. AltSchool, he says, is not a school or a start-up, it is a “full-stack education company”. In tech speak, this means it does everything: a third of the employees are engineers, a third are educators and a third run the business.

Max Ventilla, who left Google to become AltSchool’s chief executive © Carlos Chavarría

Parents pay fees, hoping their kids will get a better education as guinea pigs, while venture capitalists fund the R&D, hoping for financial returns from the technologies it develops. AltSchool started selling to select “alpha” partner-schools, who are contributing feedback, last year.

Every industry began as artisanal, Ventilla says, but changed when people pooled resources. “We used to produce our clothes, our materials, our everything at the household level,” he says. “Now you spend a million ­dollars of software-engineer time to write a program but once it is written, it is basically free to have a two-millionth person [use it].”

Silicon Valley has not yet succeeded. Ventilla lists the challenges that a child in 2050 needs to be prepared for: the changes wrought by globalisation, the acceleration of technical change, longer life spans and the “tidal wave” of artificial intelligence.

“The purpose of schools is to prepare kids for the future and that goal post is moving higher and higher, and faster and faster. The change is happening, it is just not happening fast enough,” he says.

What the technology industry can perhaps bring best to education is its methods. “Iterate” is a much-loved word in Silicon Valley, encapsulating the desire to pursue big changes (Facebook’s famous motto: move fast and break things) and test rapidly to see which button works better (A/B testing).

Ventilla knows education will not be one of Silicon Valley’s classic success stories with a growth chart that looks like a hockey stick and overnight riches for founders. “Ultimately, our mission is to enable every child to reach their full potential. It is not every American child, it’s not every child in an urban area,” he says. “If you could do it quickly, it would have been done.”