It Will Do Whatever Fool Thing You Ask It

I’m talking to more people lately who are fearful for their profession in an age of AI. I keep coming back to “The Map Is Not The Territory.” I need to refresh that post.

This week’s stunning release of OpenAI’s new universal model capabilities enable us to “vibe-design” (riffing from the latest “vibe-coding” trend) imagery, infographics, game mockups, transforming between styles and the like. That “vibe-[verb]” thing is the way we iterate on something with an AI model, and our side of the iteration is driven by feeling and intuition to shape what the AI model produces so that it’s most impactful for us.

We cannot digitize feelings or train them into a model. The thing that drives human connection, human creativity, relationships. Standing in front of the Bust of Nefertiti is not the same as looking at a photograph of it.

Sure, a good developer could type code well. A good writer could wield spelling and grammar correctly. A good designer could sketch out a layout.

But that is not what differentiates them even among humans, let alone from GenAI. Sure, GenAI can invoke feelings in us like any other artistic form, but that’s how feelings work, we can’t help but have them. But the GenAI didn’t invoke them, the material from which it was trained did.

Developers create novel code every day, can empathize with the frustration of a user to improve usability. Writers and journalists interview people because the data from a questionnaire isn’t enough. That musician is fed by their audience as much as the audience is fed by their music. Designers feel the ebb and flow of how their designs impact an audience, whether their medium is code, watercolour, sound, or words.