Before AI, Learn to Think

When I was a kid, judo also educated me. I was a big fan of Toshihiko Koga, whose poster hung above my bed.

Koga was a role model to me because he competed in my weight class (-71 kg) and mastered the same signature throw I loved: ippon seoi-nage. But he did it differently. Instead of the classical parallel foot placement, Koga developed a “split-hip” entry, placing his right leg deep between his opponent’s legs while gripping the armpit rather than the lapel.

With this adaptation, he could throw opponents far heavier than himself. In 1990, weighing just 71 kg, he fought his way to the final of the All-Japan Championships, an open-weight tournament, facing heavyweight world champion Naoya Ogawa. He lost, but the performance became legendary: the smallest fighter to reach that final in decades.

Here’s what strikes me now: when Koga taught ippon seoi-nage, he always started with the classical form. He believed mastering the fundamentals was essential before developing variations. The unspoken truth? Most judokas never truly master the basics. When I received my black belt, I was taught: “Now you can finally start learning.”

Why am I telling you this on the Education Futures blog?

Because when we discuss education and AI, I don’t think we’re being reactionary to suggest that:

• Before learning to run, you must learn to walk. Before using a calculator, you must learn to count. Before using AI, you must learn to think. AI doesn’t replace thinking, it reveals whether thinking was ever there.

• Role models in education matter more than we often acknowledge.

• Education is not the monopoly of schools.

Koga taught me that innovation requires foundation. His revolutionary technique only worked because he had internalized every principle of the classical form. The same logic applies to AI in education: these tools amplify capability, but only when built on genuine understanding.

Laurent