AI Can't Be Creative Because It Can't See What's Missing

Creativity in Negative Space

AI Can't Be Creative Because It Can't See What's Missing

Steve Francia argues that AI operates entirely in "positive space"—recombining existing patterns from training data—while human creativity lives in "negative space": noticing gaps nobody has named and filling them. He illustrates with his own experience building Cobra, the CLI framework behind Kubernetes and Docker, which emerged from an intuition that existing tools were wrong, not from any data point. Fluency and creativity are not the same thing, and as AI makes polished output cheap, the ability to spot what doesn't yet exist becomes the only thing that matters.

A gap nobody has named leaves no trace in the data, so nothing in the weights marks it as missing and nothing in training rewards finding it.
  1. senfiaj

    As I see AI just hasn't the same motivations and desires as humans, it's trained to be just compliant and obedient. Creativity is a subjective term. It's something novel that causes positive emotions. AI hasn't these emotions and desires to experience them.

  2. bshepard

    Linking negative space and creativity is a profoundly useful concept, thank you.

  3. slopnt

    Of course an article talking about how AI lacks some type of human creativity is completely LLM written. You can't make this up.

More from this day

2026-11-20