The question isn't whether AI replaces human creativity. It's whether we can recognise craft when it changes form.
There's a particular joy in managing a music library. Not a playlist — a proper library, with curated tags, album art, and metadata. A niche pursuit in the age of streaming. The app that makes it possible is Meta, built by Benjamin Jaeger at Nightbirdsevolve Software. A small, focused Mac app. One developer, one purpose, beautifully executed.
A recent email exchange with Ben wandered from a licensing issue into something more interesting — what AI means for the craft of building software. His position was clear: he builds nothing using AI. He values thinking creatively, figuring things out through trial and error. The fun of life, he said, is working things out on your own. We try, we fail, we learn.
It's a position worth sitting with rather than dismissing.
But here's what made the exchange valuable: it wasn't an argument. Two people who build things for a living, looking at the same technological moment, drawing different conclusions. Both honest. Both valid.
Ben sees AI as useful within boundaries — debugging, problem solving, understanding systems — but a risk once it removes the satisfaction of making something. The other view is that human value is shifting. The thinking about what to build and why is becoming the craft. The implementation is increasingly shareable with machines. A terminal with a large language model powering execution has transformed what's possible for a single person or a small team.
The current discourse tends to collapse this into a binary. Embrace AI fully or be a luddite. Craft is dead or AI is a bubble.
The reality is more textured. Craft relocates. It always has.
The industrial revolution didn't eliminate craftsmanship — it moved it. From hand-production to design, engineering, and systems thinking. The resistance was understandable then too. And yet, new forms of mastery emerged that the previous generation couldn't have imagined.
The craft is moving from syntax and implementation toward architecture, constraint design, and orchestration. The SaaS market senses this already — stock prices across the sector have taken hits precisely because custom software, built quickly with AI tools, can now replicate what took teams and months to ship. The economics of "good enough, built fast, tailored to you" are starting to win.
None of which settles the question. Ben's perspective — that the joy is in the figuring out — isn't wrong. It's a reminder that not everything worth doing needs to be optimised. Sometimes a beautifully crafted, focused app that manages music metadata is its own answer.
Every technological revolution produces this tension. The people who mourn the old form of craft miss the emergence of the new one. The people who dismiss the old form lose something worth preserving.
The interesting place is in between — recognising that craft doesn't disappear. It relocates. And the real skill is noticing where it's gone.