Teach Me the Song Before You Ask Me to Dance
Freedom doesn’t come from abandoning structure. It comes from trusting it completely.
People often ask what it was like managing an ecosystem of 30,000 products, 120,000 beneficiaries, more than 200 interconnected data elements, and seven enterprise systems serving an organization of 40,000 employees.
Most people expect the answer to be complexity.
It wasn’t.
The real answer was rhythm.
Every System Had a Beat
Every table had a purpose. Every interface entered on cue. Every data element knew where it belonged. If one part of the orchestra drifted out of time, the rest of the music noticed immediately.
Architecture, at its best, isn’t about diagrams or databases. It’s about creating a rhythm that allows thousands of moving parts to work together without collapsing into chaos.
For years, that rhythm was my world. It taught me something I still believe today: complexity isn’t the enemy. Uncertainty isn’t even the enemy. The real enemy is losing the shared understanding that keeps everyone moving together.
I Don’t Want AI to Dance for Me
That’s why I’ve found the AI revolution both exhilarating and, at times, deeply frustrating.
Everyone is talking about making AI more creative, more autonomous, more capable of generating code, documents, designs, and decisions. Those are remarkable achievements. But I find myself wanting something much simpler.
I don’t want AI to dance for me.
I want it to dance with me.
There’s an important difference.
Great Improvisation Only Works Because Everyone Already Knows the Song
Think about jazz. To someone listening from the audience, it sounds wonderfully spontaneous—as if the musicians are making everything up as they go. Anyone who has ever played in an ensemble knows the truth.
They understand the rhythm. They trust the bass line. They know when to step forward for a solo and, just as importantly, when to step back and let someone else carry the melody.
Freedom doesn’t come from abandoning structure. It comes from trusting it completely.
That’s the relationship I hope we eventually build with AI.
Today, too many interactions still feel like teaching someone a dance by shouting individual steps across the room: Forward. Left. Turn. Stop.
Recovery Is Part of the Performance
Now imagine something changes. The music skips. One musician loses the beat.
In a great ensemble, nobody pretends nothing happened. The musicians look at one another. They listen. They adjust. They find the rhythm again before the audience even notices.
That’s where I think today’s AI still has the most room to grow.
Making mistakes has never bothered me. Every engineer makes mistakes. Every architect has designed something they later improved. Every teacher has watched a lesson go sideways. Mistakes are part of learning.
What matters is what happens next.
When experienced people realize they’ve lost the thread, they slow down. They ask questions. They retrace their steps. They revisit assumptions. They say, “Something doesn’t feel right.”
Too often, AI does the opposite. It continues confidently. The words keep flowing. The code keeps growing. The conversation moves forward. But the rhythm has already been lost.
As an architect, that’s the moment I care about most—not the mistake itself, but the recovery. Because recovery is where trust is built.
AI Needs Compass Checks, Too
Imagine an assistant that paused and said, “I think I may have violated one of your architectural principles.”
Or, “This solution doesn’t seem consistent with the rest of your system.”
Or simply, “Before I continue, can we make sure I’m still following the rhythm you intended?”
That’s an AI I would trust—not because it never makes mistakes, but because it knows how to recover from them.
For decades, architects have designed systems that allow thousands—sometimes millions—of independent actions to remain beautifully synchronized. We didn’t achieve that through rigid control. We achieved it by creating shared contracts, dependable rhythms, and enough structure that people could solve new problems without breaking everything around them.
Maybe that’s the lesson AI has been waiting to learn.
Teach Me the Song Before You Ask Me to Dance
Because the future isn’t machines replacing human creativity.
It’s humans and machines learning to improvise together—guided by shared rhythms, grounded in trust, and always willing to pause long enough to find the beat again.
I have a feeling that’s when the real music begins.