“The way we typically teach is backward. In schools, we start with the theory, then slowly build up examples — but when we learn sports or music, we watch a full game or play a song first, and only later break it down into techniques. That’s how we should teach coding and machine learning too.”

Why this hit me deeply:

I’ve always learned best by seeing the full picture first — getting inspired, getting curious, and then slowly unpacking the pieces. Jeremy Howard confirmed for me that this isn’t just a preference — it’s a better way to learn. He brings this into how he teaches AI: by showing something powerful right away, and then guiding students through the concepts that make it work.

What I admire:

• He’s an AI researcher, yet chooses to focus heavily on how to teach well.
• He shares real, working code upfront, to motivate and excite learners.
• He isn’t afraid to say the education system is flawed — but he acts by building something better.
• He shares his credentials not to brag, but to earn trust — because he knows that’s how some people decide who to listen to.

My teaching reminder for next semester:

Start with something real. Show students what’s possible first. Let them get curious, excited, and motivated. Then teach them the tools. Teach like coaching a sport — not like a math textbook.

🧠 "Let them play the game first." 🎓