For the last 30 years, technical execution was the bottleneck. You needed to know how to code, how to render, how to distribute. Taste was abundant. Execution was scarce.
AI flipped it.
Execution is now everywhere. The bottleneck is knowing what’s worth making.
The earliest Art Blocks scripts were on-chain, readable by anyone. You could run the algorithm. What you couldn’t run was the 10,000 hours of looking at art that let you recognize which outputs were transcendent. The moat wasn’t the code. The moat was the eye.
Everyone who collected early generative art started developing a thesis about what matters. You can’t look at thousands of algorithmic outputs and not build a framework for what’s worth keeping. Taste is the only thing that compounds.
The hardest problem in AI right now isn’t capability. It’s curation. What do you surface? Who decides what’s good? How do you build recommendation systems that don’t just reflect back what people already liked?
Those are all taste problems. And they’re getting harder as output volume increases.
The algorithm is available to everyone.
The eye is yours alone.