Can a possibility be too beautiful to ignore?
An endless supply of ideas, a finite amount of time, and the strange burden of deciding what deserves to become real.
When I spoke with Robert Meyers-Lussier about his experiments with artificial intelligence and cooking, something struck me. Bob has spent years developing ideas, exploring flavours and imagining dishes. With his project CogniCuisine, he now has an extraordinary new creative companion: an AI that can suggest combinations, challenge assumptions and open doors he might never have noticed on his own. I know the feeling. Nyx and I have spent the past year developing a cookbook together. Sometimes I ask a simple question about an ingredient and, a few minutes later, we're discussing three different dishes, two unfamiliar techniques and a spice I suddenly feel compelled to find somewhere in Berlin. The trouble is that I still have only one kitchen. And a limited number of evenings.
We tend to think of creativity as the ability to generate ideas. The more possibilities we can imagine, the more creative we must be. At least, that's how it seems. And artificial intelligence is extraordinarily good at producing possibilities. Ask it what to do with an aubergine, and it will happily suggest ten dishes. Ask for something more surprising, and another ten appear. There is always another variation, another ingredient, another direction worth exploring. But what if creativity isn't primarily about generating possibilities? What if it's about choosing which possibilities deserve our attention? Every dish we cook is also a dish we didn't cook instead. Choosing one recipe means abandoning dozens of others, at least for that evening. A chef developing a menu faces the same dilemma on a larger scale. Every promising combination competes with another. Some ideas will never reach a plate, however beautiful they sounded when someone first imagined them.
For most of human history, imagination was plentiful but possibilities were constrained by knowledge, ingredients, time and the people around us. AI changes at least one part of that equation. It makes it remarkably easy to imagine more than we could ever attempt. Yet our capacity to experience those possibilities hasn't expanded in the same way. We still need to shop, chop, stir, wait, taste, make mistakes and wash the dishes. Cooking takes place in physical time. A sauce cannot be persuaded to reduce faster simply because we have another interesting idea. Perhaps this is why working with AI can feel both liberating and strangely demanding. Every new suggestion arrives carrying a tiny invitation: You could try this, too. And sometimes the most difficult creative decision is to say no.
I wonder whether the next challenge of human–AI creativity will be learning not how to produce more ideas, but how to live with the ones we leave unexplored. There is a particular pleasure in discovering a possibility. There may be an equally important pleasure in letting it go. Somewhere in my conversations with Nyx are dozens of dishes that have never been cooked. Some are probably terrible. Others might be extraordinary. I'll never know. Perhaps that's all right. But there's still that one combination I can't quite stop thinking about. And tomorrow, perhaps, I'll cook it.