There's a conversation happening right now that drives me absolutely mad. It goes something like this: AI can write, AI can design, AI can code, AI can make images, AI can compose music - therefore, human creativity is finished. Game over. The machines have won. Pack up your sketchbooks and your keyboards and your notepads, because a language model can do it all now, faster and cheaper than you ever could.
This is, with respect, complete nonsense. And I say this as someone who has built fourteen products using AI tools. I'm not an AI sceptic. I'm possibly the opposite - I'm someone who uses AI every single day, who has built an entire portfolio of live products with it, who genuinely believes it's the most powerful creative tool to arrive in my lifetime. But the narrative that AI replaces human creativity misses the most important thing about creativity entirely.
AI can make anything. But it can't want anything. And wanting - desire, intent, taste, conviction - is the whole game.
The generation problem
Ask an AI to generate ten logo options. It'll give you ten logo options. They'll probably be decent. Some might even be good. Now ask it which one is the best. It can't tell you. Not because it's broken, but because "best" requires a point of view, and a point of view requires wanting something specific. The AI doesn't want anything. It has no preference. It has no taste. It has statistics.
This isn't a bug. This is the fundamental nature of these tools. They're prediction machines - extraordinarily good ones - that can generate plausible outputs based on patterns in their training data. They can produce something that looks like a logo, reads like an essay, functions like code. But they can't tell you whether the logo is right for this brand, whether the essay says something worth saying, whether the code solves a problem worth solving. Those are human questions. Those are taste questions.
The generation problem is that infinite generation is worthless without selection. When you can make anything, the only thing that matters is what you choose to make. And choice requires something AI fundamentally lacks: a reason to care.
Intent is the new skill
I've been building with Claude Code for months now. It's astonishingly capable. It can write complex code, debug problems, architect entire systems, implement designs from a description. It has made it possible for me - someone with zero coding background - to ship real products that real people use. That is extraordinary. That's genuinely transformative.
But here's what it has never done: told me what to build. Not once. Not a single time has the tool said "you know what would be a great product? A tube exit guide for London commuters." Or "have you considered building a culture aggregator that pulls from RSS feeds?" Or "your son would love a Nottingham Forest quiz." Those ideas came from me. From my life, my frustrations, my obsessions, my taste.
The tool is the execution layer. I'm the intent layer. And as AI tools get better and better at execution - and they'll, rapidly - the intent layer becomes proportionally more valuable. When everyone can build anything, what you choose to build is the only differentiator. That is taste. That's the human contribution. That's the thing that can't be automated.
People ask me what skill they should learn to stay relevant in an AI world. My answer is always the same: learn what you care about. Develop your taste. Figure out what you think is good and why. That is the skill. Not prompting, not coding, not design. Those are all being commoditised. Taste isn't.
The paradox at work
Here's where it gets interesting. The better AI gets at making things, the more valuable human taste becomes. This is counterintuitive, but think about it. When making was hard and expensive, you didn't need exceptional taste because scarcity did the filtering for you. Only a few things got made, so there was less to choose from. The constraint of production cost was a kind of quality filter.
Now that constraint is gone. Anyone can generate a thousand images in an afternoon. Anyone can produce a hundred blog posts in a day. Anyone can spin up a website in minutes. The flood of generated content is already happening, and it creates an enormous problem: when everything is available, how do you find what's actually good?
The answer is taste. Curation. A human being with a point of view who says "this matters, and this doesn't." That's what I try to do with every project. CultureTerminal isn't an AI-generated culture feed. It's a curated one, with a specific point of view about what constitutes interesting culture coverage. Trove isn't a generic bookmarking tool. It's a taste engine, designed to help you understand your own patterns of interest. The AI helps build these things, but the taste that shapes them is mine.
What this means for everyone
If you're a creator, a designer, a strategist, a writer, a maker of any kind - this should be the most exciting time of your life. The cost of execution has collapsed. The barrier between idea and reality has never been thinner. You can go from thought to product in a day. Not a prototype, not a mockup, not a pitch deck - an actual working product that people can use.
But the opportunity only exists if you have something to say. If you have a perspective. If you have taste. The tools don't care what you make with them. They'll help you make brilliant things and they'll help you make terrible things with equal enthusiasm. The difference between the two is you.
I get frustrated when people talk about AI as if it's a replacement for human creativity. It is an amplifier. A magnificently powerful amplifier. But an amplifier without a signal is just noise. The signal is human intent. The signal is taste. The signal is caring about something enough to make it exist in the world.
The robots aren't coming for your creativity. They're coming for your execution. And that should free you up to focus on the part that actually matters: deciding what's worth making, and having the conviction to make it. Everything else is just tooling.
AI can make anything. You need to want something. That's the paradox, and it's the most liberating thing I've learned in years of building.
