Scrape Your Knees: Why Falling is Essential to AI Mastery

This weekend, I found myself in a familiar parental scenario: teaching my daughter to ride a bike. Despite having older sisters who ride with ease, she was struggling, held back by a paralyzing fear. As I watched her hesitate, I realized: if she doesn't fall, she can't learn to ride.

She has to fall if she wants to learn.

And seemingly worse — I’ve got to let her fall if I want to teach her.

Drawing from a distant memory of a study that suggests pain creates new neural pathways (in fact, children who are given baby walkers take considerably longer to learn to walk — “might the same be true for training wheels??” I wondered…) I decided to shift our approach.

"You're going to have to fall at least 10 times before it'll work!" I told her. "So let's start counting!"

The transformation was immediate and profound. Suddenly, falling wasn't a dreaded outcome but a necessary step toward success. With each tumble, we'd all cheer — daddy and big sisters included — "That's one more! Five falls left to go!"

The fear dissipated, replaced by determination and, surprisingly, joy.

The Learning Paradox: Why We Need to Fall

This experience crystallized a paradox at the heart of learning: to succeed, we must first embrace failure. It's not just about bikes. From a baby's first steps to a scientist's breakthrough discovery, progress is paved with stumbles, missteps, and falls.

As I reflected on this parental epiphany, I realized its profound relevance to another frontier I've been exploring with a handful of spectacular partners in learning (shout out to the Portland Trail Blazers!): the world of AI. In a recent interview on Beyond the Prompt, Shir Yehoshua, Notion's Head of AI, shared a strikingly similar philosophy:

"I actually have a bunch of dance training, and I've learned a lot (about business) from that world. And if you're trying to learn a new trick, you can't ever get it until you fall a million times. You have to learn how to fall first, before you can actually hit the turn that you want, the leap that you want, the whatever move that you want. It's the same (with learning AI):

You have to be able to safely fail in order to succeed."

Rewiring Our Brains for AI

This approach to learning — recognizing that a fundamental rewiring has to occur — resonates deeply with me. Just the other day, I had the privilege to deliver an AI keynote at Stanford’s annual Campbell Trophy Summit. One of the award recipients in the audience followed up after the session: “my big takeaway was that a combination of relatable use cases (shooting stars) and daily experimentation can be the key building blocks to rewiring the brain.” (Brief explanation: I was emphasizing the importance of new inputs to stimulate insights, much like "shooting stars" were to Johannes Kepler, gazing into the night sky, which was believed to be a fixed substance, “the firmament.”)

I think his takeaway is exactly right, with one caveat: in organizational life, it’s not enough to simply rewire our own brains. We need to rewire our teams’ collective understanding and expectations, too.

The Joy of Falling: Celebrating AI Mishaps As a Team

There was something really special about cheering my daughter on with her sisters by my side, enthusiastically celebrating each fall as a step in the right direction. Shir's team at Notion has found innovative ways to celebrate the necessary stumbles on the road to building one of the most celebrated AI-rich feature deployments in recent memory:

"Anytime somebody has learned something or failed at something or had an idea that failed... they succeeded because they learned something. Some of the demos that get the most amount of applause are — for examples, a couple of months ago, somebody demoed their PR (a “pull request”). They just showed all of the code that they deleted, and there were roaring cheers from the whole team."

Creating Your AI Learning Playground

So, how do we apply a "falling forward" mentality to AI learning? Here are a few suggestions:

1. Seek fresh input. Recognize that the first order priority is to spark your imagination, and find sources of inspiration that help you imagine possibilities you haven’t considered.

2. Block time on your calendar. What’s true of innovation broadly is true of acquiring the new language of innovation, fluency with Generative AI: it requires time to explore.

3. Institute a "bad output" quota. Aim for at least 10 "falls," or suboptimal AI outputs, before expecting success. Each fall is a necessary step toward mastery.

4. Reframe your expectations. Remember, every "failure" is actually a success in learning what doesn't work. So celebrate them, just like innovators of past eras famously have.

(Ready to start your AI learning journey in a structured environment? Check out my new '5-Day AI Learning Sequence' email course. It's designed to guide you through your first 'falls' and celebrate your progress along the way. Sign up now!)

From Scraped Knees to AI Mastery

As we navigate the exciting, sometimes daunting new frontier of collaborating with AI, it’s helpful to remember the lessons from the bike path. Every scraped knee, every deleted line of code, every unexpected AI output is a step toward mastery. By embracing these "falls," celebrating our mishaps, and persistently getting back up, we're not just learning to use AI – we're rewiring our brains for a future where human creativity and artificial intelligence dance in harmony.

So, are you ready to scrape your knees on the path to AI mastery? Block your calendar, set your falling quota, celebrate your stumbles, and remember – the joy isn't just in the destination, but in knowing that tumbles are the only way to mastery.

If you aren’t falling, you aren’t learning!

And if you don’t let your team fall, they’ll never learn, either.

Related: Unlock AI’s Potential
Related: Beyond the Prompt: Shir Yehoshua, Notion’s Head of AI
Related: 5-Day AI Learning Email Course
Related: Jeremy Utley: AI Ignites Human Creativity (Campbell Trophy Summit)
Related: Make Time for Exploration
Related: Celebrate!

Join over 24,147 creators & leaders who read Paint & Pipette each week