The Age of the Mediocre Polymath
AI is removing friction at a staggering pace. Activities which once required years of skill-building, for example designing a UI, writing code, composing music or generating copy, can now be done in minutes by anyone with a prompt and a pulse. We’re entering an era where the bottleneck is no longer can you build it, but should you and will it actually be any good?
This shift is both liberating and deeply misleading.
On the surface, it's a golden age for creative generalists. You can mock up an app in the morning, generate a marketing plan by lunch, and ship a working prototype before dinner. The problem is, ease of execution creates the illusion of quality. Just because you can build something doesn’t mean you’ve built something meaningful. Worse, when everything becomes effortless, there’s less incentive to push for depth, craft, or even basic standards.
AI is turning all of us into passable designers, average developers, and decent writers. But not experts. And if we’re not careful, we’ll start mistaking volume for value, and output for outcomes. “I can, therefore I will” is a thought process that destroys companies.
The real challenge now isn’t capability. It’s discernment. Knowing what to build, why it matters, and how to shape it into something that’s not just functional, but genuinely useful and resonant. That takes more than prompts; it takes experience, taste, and the humility to know when something just isn’t good enough yet.
One of the most insidious side effects of AI-powered creation is aesthetic parity. Everything looks professional. The design is clean, the copy is coherent, the code runs. On the surface, your DIY landing page is indistinguishable from one crafted by a seasoned team. But polish is no longer a proxy for quality. And when the visual and functional markers of a “real product” are available to anyone, how do we know what we’re looking at is actually good?
The hard truth is, we don’t…unless we know what to look for.
AI tools are excellent at replication. They’ve been trained on a corpus of what “good” has looked like in the past, and they can convincingly echo it. But they can’t evaluate why something worked. They can’t tell you if your product solves a real problem, if your flow respects cognitive load, or if your content earns trust with the right audience. This is where expertise still matters. Not for creating the veneer, but for interpreting the substance.
When everything looks good on the surface, signal has to come from somewhere else. You need to watch how users behave. Do they come back? Do they convert, engage, refer? Are they doing what you hoped they would, and if not, why not? You need to listen to real, contextual feedback. Not just “do you like this,” but “did this solve the problem you had before you opened the app?” The more specific the context, the more useful the insight. You need to pay attention to the qualitative nuance. A designer sees friction in a flow others miss. A PM picks up on emotional disconnect in messaging. A researcher spots a pattern in silence. These nuances aren’t things AI is capable of catching. They come from lived experience and deep attention.
The DIY approach can be powerful, especially in the early stages, when speed of iteration matters more than perfection. But there’s a threshold where “just ship it” becomes “we’re building on a shaky foundation.” DIY makes sense when you're testing an idea, validating interest, or learning fast. You don’t need a perfect design to prove someone wants what you're offering. But when the stakes are higher, when you're scaling, entering a new market, building trust with a skeptical audience, or when the product’s success hinges on subtleties like usability, accessibility, or emotional resonance, you need expertise. Real, human expertise.
The paradox is that AI gives us a running start, but it can also dull our instincts. It’s easy to skip the hard questions because the output looks convincing. But in a world where everything appears finished, the real work is learning to see beyond the surface.
