Monday Musings (Questions about Education)

Hi friends,

Greetings from Austin!

One my biggest surprises of 2022 was how much time I devoted to learning science. The launch of Write of Passage Liftoff (our program for high schoolers) has me thinking about our approach to writing education. Our goal is to eventually have thousands of kids in the program. To do that, we need to both persuade parents that Liftoff is a good use of their kids’ time — and design a program that ensures it is.

After a year of intense study, here are my burning questions about education:

1. How can we better frame tests to help students learn faster?

Students loathe tests. Education reformers want to eliminate them and colleges are dropping standardized tests from admissions criteria. Their complaints have some merit. Yes, tests are stressful and nerve-wracking. Yes, teachers “teach to the test” because they’re evaluated on their students’ test scores. And yes, the way we currently test kids leads them to only care about what they’ll be tested on rather than learning for its own sake. But tests are a remarkably effective learning tool. Long-term recall improves when students are asked to remember information. The problem is tests are overused at the end of the learning process, which makes them an intimidating form of judgment instead of a valuable learning tool. Intermittent tests are useful because they reveal skill gaps, which helps students focus their learning efforts.

— —

2. Can AI tools solve Bloom’s Two-Sigma Problem?

In 1984, Benjamin Bloom published a paper showing how Mastery Learning combined with one-on-one tutoring led to a two-sigma improvement in learning outcomes. We know this faster learning curve is possible. The problem is that we can’t possibly give every student a personal tutor. It’s too expensive. But digital tutors might offer the answer. In 2009 The United States Navy showed how effective digital tutors can be for technical training. Using their tool, new recruits took 16 weeks to reach the level of knowledge of somebody with five years of experience. That was more than a decade ago. Modern technology can help us do even better and ultimately solve Bloom’s two-sigma problem while meaningfully lowering the cost of education.

— —

3. How can we identify profitable business models to attract top talent?

People get a sour taste in their mouths when they hear about for-profit education. But if we want a revolution in education, the industry needs to attract the best people. Everybody wants teachers to be paid more, but changing the industry's fundamental economics is the only way to do that. Few of the smartest people I know work in education because it's historically been grueling, low-paying work. Instead of helping people learn, ambitious types pursued lucrative (but hollow) careers optimizing ads in the tech industry or making markets marginally more efficient on Wall Street. There’s very little entrepreneurship in education either. Until education becomes profitable (and solutions aren’t limited by student to teacher ratios), we will be stuck with rising costs, falling outcomes, and government-employed teachers who select for sinecure. We can do better. Education should have as many options as the rest of the free market. There are fifty cereal brands in a supermarket, but when it comes to educational approaches, you only get a few options. Most schools are the same: curriculums are fragmented by subject, and classrooms are lecture-based and strongly teacher dependent. Profitable business models will attract high-quality talent and incentivize entrepreneurs to innovate.

— —

4. How can we make learning apps fun?

Learning apps are a crucial education tool because they lower costs and scale well. The problem is they’re boring. They look and feel like school. My friend calls them “chocolate coated broccoli.” The gap between what people do for fun and what they do in school doesn’t need to be as large as it is today. Through games, kids can learn to read, solve problems, and receive feedback when they mess up. Video games like Factorio captivate kids while teaching them principles of construction and management. The same kids who are told they can’t focus... often pull all-nighters gaming. We need more games like that for education.

— —

5. If homeschooling rises and kids learn on their computers, how will they be socialized?

Parents resist homeschooling and online learning tools out of fear their kids won’t be socialized. We need people to dissect the problem more. One definition of socialization is intellectual conformity. Many people see “common knowledge” as the bedrock of a cooperative society. Another social definition is the ability to get along with others. But kids may not need schools to achieve this. I know plenty of formerly homeschooled students who don’t struggle with social skills. They learned with other kids in micro-schools, volunteered at their local churches, and played on afterschool sports teams. Since most parents don’t like taking risks with their childrens’ education we need better data on which activities socialize kids the best.

— —

6. What could a universal learning profile look like, and how can it help students learn faster?

Learning apps will make it easier to build datasets for learning. Every kid could have their own learning profile, similar to the comprehensive ones for patients at hospitals. Students could use this data to measure their progress and identify gaps in their own learning. For teachers, student learning challenges would be easier to diagnose. With all this data, the apps could personalize the learning experience for each student.

— —

7. How does age segmentation impact child development?

I have a friend who insists on speaking to his kids like adults. Naturally, they developed an advanced vocabulary and matured quickly. Then, they started kindergarten. At school, they were surrounded by kids with limited vocabularies and “baby-talking” teachers. Their speech degraded. So did their interests. By winter break, he yanked them out of the classroom and started homeschooling. Sorting kids by age is a relatively new phenomenon. The parents I know who support age segmentation believe it decreases bullying and improves social skills. But kids were historically surrounded by both younger and older peers. My hunch is age segmentation removes mentorship and role models, which inhibits learning. It prevents adoption of mastery learning since smart kids cannot skip multiple grades.

— —

8. How can we raise self-directed learners without ignoring the essentials?

Schools train compliance over curiosity. Instead of following their innate interests, kids are told to conform to what everybody else is studying at the pace of the herd, so they can become docile workers. Leaving school was the best thing that ever happened to my curiosity. Without a teacher to constrain me, I became a passionate self-directed learner. In the Internet Age, where information is so easy to access, the drive to self educate should be a core goal. Nevertheless, there are skills every student should learn, like mathematics, reading, and writing. Though they’ll sometimes be a slog to get through, students shouldn’t graduate with the idea that learning is boring. How can we make the fundamentals fun, while giving each student autonomy?

— —

9. What factors make kids love school the most, without hurting learning outcomes?

Kids spend the first two decades of their life at school. They rate their experiences pretty highly until middle school, when engagement scores begin to decline. The majority of students are disengaged by high school. School should be delicious and nutritious. I learned the most from invigorating creative projects, and many of my best teachers were also the funniest. The link between boredom and learning needn’t be as strong as it is today. Bored students may even learn less! We need language, data, and examples to show that fun learning environments can be the most effective.

— —

10. How can we make spaced repetition easier and more enjoyable?

Learning is a battle against the decay of memory. Not all memorization efforts are created equal. Reviewing the same piece of information five times in five days is a much less effective memory technique than reviewing it five times in five years, spaced out over sequentially longer time periods. Principles like spaced repetition have been proven to boost memory. The problem is flashcards are boring to create and unsatisfying to study. If we embed the idea into software tools that are easy and fun to use, we can increase the adoption of spaced repetition.

— —

11. What should a modern writing curriculum look like? ​
​
Math is the cleanest subject to teach. From simple addition to the complexities of calculus, each layer builds on those before it and leads into what comes next. Math is easy to score for teachers because computers can do most of the grading. For students, gaps in knowledge are easy to spot. Since math curriculums are so similar worldwide, students can easily find help on the Internet whenever they're stuck. Writing education has no equivalent. It’s also a more subjective craft. Nevertheless, there are universal principles that underlie all good writing. The rise of Grammarly and ChatGPT changes what needs to be taught. Though writing education will never be as orderly as math, it’d benefit from a more systematic approach where students can progress in accordance with the principles of Mastery Learning.

— —

12. How will GPT change writing education?

ChatGPT has already altered the playing field for school writing assignments. Every student will increasingly be able to open their computer, type in a question, and instantly produce something decent. Generic writing becomes obsolete. Writing teachers should now be asking: “What’s scarce in a world of GPT writing?” For starters, we should move beyond five-paragraph essays. The technicalities of writing are overly emphasized too. One of my least favorite and most useless college classes focused on memorizing the AP Style Guide. Too many writing curriculums associate good writing with good grammar. Meanwhile, many of the most popular Reddit posts are littered with spelling errors, but nail what ultimately matters: captivating readers. The combination of good ideas, flow, and storytelling is far more essential than the typical writing curriculum would have you realize. Technological changes will force standard writing education to evolve.

— —

13. How can we bring the “Learn Like an Athlete” mentality into the classroom?

School and sports teams have nearly opposite cultures. Schools segment kids by age, while sports segment by ability; schools are slow to adapt to change, while sports teams desperately look for every advantage they can get; schools are afraid to let go of poor performers, while sports teams see it as par for the course; schools train kids to learn by listening, while sports teams encourage kids to learn by doing; schools teach poor study habits, while principles like deliberate practice are the essence of sports practice. Resistance to change leads to ossification. The system’s status quo bias has stunted learning outcomes. When you’re stuck with a problem in education, look to athletics. Drawing from the realm of athletics can free us from the kind of thinking that made the education system so sclerotic.

— —

14. How should we teach the Great Books?

I didn’t have the attention span for Shakespeare when I read his work in the 4th grade. Today’s students definitely don’t either. But the great books are worth learning. Some people dismiss them because they think the ideas are old and no longer relevant... what nonsense. The canon persists because the ideas are timeless. They speak to fundamental truths about the human condition. We don’t need to be introduced to the great books in their original format though. We can convert them into kid-friendly ones. Once kids appreciate the ideas, their curiosity will lead them to the original versions and a lifelong relationship with meaningful ideas.

Earlier this year, Jonathan Bi and I released a 100-minute video introduction to Rene Girard’s philosophy.

We promised six more episodes and now, the wait is almost over. We’re set to release the next two episodes on Friday. If you’d like to receive it by email, subscribe here.

If you want to get up to speed, here’s a link to the lecture we published earlier this year.