What's modet's origin story?
Last August I wanted to understand how to apply AI to Product Management, so I started Andrew Ng’s Coursera Machine Learning course.
Everything starts great. I watch the videos. Do the quizzes. Week 1 check. Same for Week 2, but then I get to the programming assignment. And suddenly I think, "what's going on?" I don't even know how to start.
Okay, so I go back, rewatch the videos, make detailed notes. And you should know, I'm pretty good at writing notes these days. But when I get to the assignment. It’s the same! 10 hours later. Nothing. I still can’t start! "What's wrong with me?"
Okay, let's do this one more time. This time I apply two insights:
What if I structure my notes the way Stripe does their dev docs?
- The left side is simple, simply what you do
- The right side gives you a working example, the relationship between what you do and how to implement it, and a visual of the UI
This one needs a little backstory. You see, I sucked at telling stories up to this point. So I've been devouring books and videos on these the last three months:
- How writers like Walter Mosley, Malcolm Gladwell, Anne Lamott create and tells stories, give you a puzzle to solve
- How filmmakers like Pixar and Disney engage you, make you relate, be an active participant and wonder what's coming next
- How William Zinsser and Larry McInerney teach you all the wrong things you’ve been taught about writing and make you unlearn it
- How illustrators sketch simple characters, use storyboards, and visually construct a story using shapes and the camera for a scene
I write everything they say. My insights, my aha’s. To the point where I could teach it. And now I know how a writer, a filmmaker, an illustrator thinks.
I rewatch the videos, second by second, and rewrite the lesson as writers would ー I change the jargon to simple words, strip the filler words, break them down into steps so an 8th grader can follow, explain through analogies and metaphors, show examples, show the process Andrew’s mind follows but doesn't write out for you.
Back to the course, a third time. This time, I can start. I'm making progress! But as I follow Andrew's instructions, it's not working. What's wrong? I could have looked up the simple answer by googling people's solutions on GitHub, but then I'm learning to the test. Ten hours later, I find it. But, wait, this one formula on slide 17 is wrong? He means a double summation but wrote just one! Of course he would say, it’s not wrong; you should know to infer it’s two. Argh!!!! Twenty hours for all of that.
Coincidentally, I was also reading Malcolm Gladwell's David and Goliath book. He tells stories of the brightest young adults’ struggles at elite colleges. How 30% of them drop or transfer out of STEM. That's when I knew I wasn't alone, and if I could solve it for myself, I could prevent the same struggles for every future student.
Next, I go pick up books and talks on the science of how our minds learn. How our machine learning techniques are inspired by the science and insights of our mind. I’m fascinated! I’m in!