Have you ever wanted to grow your career at work? What did you do? Most likely you took on a tough new project that would force you to learn new skills to get the work done. The project put healthy pressure on you to gain those skills, all because your manager was counting on you to get it done. A good project builds on existing knowledge, but also forces you to grow in areas that make you better. It’s that sweet spot of flow. Not too easy, but also not too hard.
If that’s how adults learn in the real world, why don’t kids do the same thing?
How you learn in school is completely different than how you learn on the job.
School Based Learning
- Read books
- Classroom lectures in groups of 20+ people
- Take tests to prove knowledge
- Knowledge is siloed, you learn math then science then social studies separately
- Everything is paced at a specific speed
- No control over what you want to learn, if you hate your classes you just need to suck it up and power through them
Work Based Learning
- Learn 1n1 from managers, mentors, and teammates
- Finish projects to prove knowledge
- Projects combine many skills into one problem to work on
- No pacing, advance your career as fast or as slow as you desire
- Choose your own adventure for what you want to learn, if you hate your job you get a new job
Kids are on these artificial train tracks for 22 years of their life, then all of a sudden they are ripped off and thrown into the wild west once they start their careers. What if they started work based learning sooner? Maybe even from day one?
The hardest part of work based, aka project based, learning is scaling the 1n1 aspect. This is where AI comes in. AI is quickly becoming the smartest thing in the world, which means it can be the best teacher in the world at anything. Kids in rural India will have the same access to AI as trust fund kids in New York City. It will be the great equalizer in education.
What if kids were allowed to learn what they wanted, when they wanted? Project based learning can truly be a choose your own adventure while still covering most of the basics of what every kid needs to know. A kid could probably learn calculus before the age of 12 if they truly loved math. Why slow them down with traditional pacing of going from first grade, to second grade, to third grade, etc. This becomes supercharged with AI.
Large language models like GPT-4o from OpenAI can tailor education to each student based on what they like doing the most. If a kid loves F1 racing then imagine learning about the chemistry of racing fuel, the physics of aerodynamic cars, even to how racers build their personal brands to reach millions. That kid will learn 100x faster doing that than reading out of a standard textbook and taking a standard test.
We learn best by doing, not by studying.
In a new AI world, it doesn’t make sense to study flashcards when the smartest thing in the world are AI models that everyone will eventually access for free. Being the best at remembering stuff is not a competitive advantage in a new AI first world. Instead AI should be teaching kids through projects that they’re the most interested in. People impact the world by doing things they love, so why not have kids build that sense of agency around doing interesting things while learning along the way as early as possible?