craft as ai moat
fluency is cheap. divergence is the moat.
there’s a piece of creativity research every product team in 2026 should know. koivisto and grassini1, published in nature scientific reports in 2023. 256 people tested against three ai chatbots on the alternate uses task, the standard test for divergent thinking.
on average, the machines won. the chatbots scored higher than the typical participant, and produced their ideas instantly.
the best human ideas still matched or beat anything the chatbots produced.
the average moved and the ceiling didn’t.
volume is what got cheap. the top of the range, which is where products actually get made, didn’t move.
the breakthrough product isn’t idea #47 in a list of 50 variations on the same concept. it’s the one from a direction nobody else was even looking.
this is why “generate 100 options and pick the best one” hits diminishing returns fast. if all 100 cluster around the same center, the best one is still the best version of the average. which, by definition, is still average.
there’s a second problem, and it gets less attention. experts rate raw ai output at a c-minus. not because it’s wrong. because it’s fluff. reasonable-sounding but hollow. coherent on the surface, thin underneath.
and ai almost never says “no.”
it doesn’t push back on bad directions. it doesn’t say “this idea isn’t great.” it doesn’t say “i’m not sure.” it just keeps producing, confidently, in whatever direction you pointed it. fluency without judgment. talking a lot without saying anything.
a senior designer’s taste isn’t generating more options. it’s knowing which option is actually different, and which is just a rearrangement of the same thing.
the builders winning right now have the taste to spot divergence when it appears, and the craft to make it real.
the moat isn’t building more.
it’s building different.
ai gives you volume. you bring judgment about which direction is worth obsessing over. that split is the whole game.