the bottleneck between the mind and the work
A few months ago, I started writing a blog post arguing that our growing dependence on AI would eventually lead to a kind of cognitive atrophy—like outsourcing too much of our thinking would make the mind lazy, slow, and less imaginative. It felt like a neat conclusion at the time, almost too neat. I framed it as a warning: if we let machines think for us, we’d forget how to think for ourselves.
But the more I’ve worked with AI since then, especially building tools and integrating LLMs into real products, the more I’ve realized how broad and premature that conclusion was. Not wrong in spirit, but wrong in its sweeping assumption. What I called “dependence” might actually be something closer to an extension of thought, not its replacement.
And that shift in perspective brought me to a more interesting idea:
One of the biggest constraints on human creativity isn’t a lack of ideas, it’s the bottleneck between the mind and the medium.
Tyler Hobbs's Code Goes In,Art Comes Out
This is evident in most creative fields:
Music arrives in the head faster than fingers can capture it (think about that bathroom epiphany where you were suddenly beyonce).
Poems appear whole, but fall apart when you sit down to write and you find your words falling under the weight you put them.
Paintings exist as perfect images in the mind but become compromises on canvas as the strokes progress.
Even code—the logic, the structure, the architecture often comes as a clean flash before you start typing and debugging.
There’s always a gap.
And in that gap… things leak, fade, or lose their initial power.
If then we could transcribe ideas as they emerge, exactly as they first appear, before the world, the hand, or the keyboard slows them down; our creative output would be unrecognizably rich.
This is partly why the criticism of “vibe coding” always amuses me. People miss the point. It’s not sloppiness; it’s speed. It’s the closest thing we have to moving from thought directly into code. It’s an attempt to reduce that creative bottleneck—to let the idea breathe before rules, structure, and syntax dilute it.
And maybe that’s the real frontier:
Not AI replacing thinking, but AI helping to remove the friction between inspiration and expression.