I’ve been thinking about this question for a while: when AI can do the engineering, what does that leave for the human?
The honest answer starts with a confession. I am nearly always drawn to the engineering side of things. Given a choice between fixing the system and sitting down to write, I’ll fix the system first. Every time. There’s something satisfying about it. A problem has a shape. You can hold it, turn it, find the angle that gives. You know when you’ve cracked it.
Writing doesn’t work like that. It doesn’t tell you when you’re finished. The end point is softer and harder to trust. So I’ve often wondered whether my pull toward engineering is something I should push against, a default I reach for when the more uncomfortable work is waiting. I don’t think it is. I think it’s temperament. And maybe something more than that.
The instinct to fix isn’t only technical. It shows up when a conversation is stuck, when a process isn’t working, when a team is going around in circles. I notice what’s tangled and I want to help it find a better shape. Engineering is one expression of that. The wider instinct is something like: I help things move forward.
But here’s where AI makes the question more interesting, and more uncomfortable. Earlier this week I had a meeting about the church website I help support. The hosting contract is up in July. The platform is expensive and proprietary, and I’d been thinking for a while that we should move it to WordPress. I estimated it would take a month. Custom theme, CSS, probably some PHP. The usual.
Then I thought: I have an AI team. That is, quite literally, one of the things they’re for. I set up a subdomain, installed WordPress, pointed Arbor at the existing site and said: recreate this. Within a day, the core rebuild was done. Not without iteration, not without me saying, “I don’t like how that hover effect works, can you make it darker, give it a drop shadow?” But those were niggles, not problems.
What would have been a month of properly in-the-weeds work… done in a day. Phenomenal. That really is the only word for it.
But I’ve been sitting with what that means. If the functional barriers come down, what happens to the people who made their living clearing them? I don’t have a settled answer. I hope it goes the way manufacturing automation went, at least partly: not fewer people, just people doing better work with better tools, and more of the population getting access to things that were previously out of reach. I hope. I’m not certain.
Maybe part of the answer depends on what kind of work we are talking about. For some work, the output carries most of the value. A website needs to work. A report needs to be clear. A dashboard needs to help someone make a better decision. In those places, AI doing more of the heavy lifting feels fair game. But personal work is different. Writing, faith, art, reflection, anything that is meant to reveal something of the person… that kind of work cannot be reduced to output. The process matters because the process is where some of the honesty is found.
What I’m more certain about is something the AI experience has revealed rather than resolved. When I asked someone close to me whether they could do what I did with the Bridge site, they probably could. Maybe in two or three days rather than one. Which raises the real question: if the technical barrier to creativity is gone for almost everyone, why don’t more people create?
I think the real barrier was never the tools. It was exposure.
Creativity reveals you. It shows the gap between what you imagined and what you managed to make. It shows your taste, your limits, what you care about, what you avoid. Before AI, it was easier to have somewhere to hide. I’d write more if I had a better system. I’d publish if I had the right tools. AI removes a lot of those excuses. But once the excuses are gone, the deeper question surfaces: now that you can create, what happens if what comes out isn’t very good? What if you’re not quite what you hoped you were?
That’s the mirror creativity holds up. AI can remove the burden of execution. It cannot remove the awkwardness of being seen by your own work. In fact, by taking away the practical hiding places, it might make that risk more visible.
There’s a useful frame for this from the Old Testament. Bezalel, in Exodus 31, was filled with the Spirit of God for skilled craft work. Not preaching. Not leading. Technical, practical, creative making. The Spirit and the craft were not in tension. The craft was one of the ways the Spirit was at work.
AI can help with the doing. I’m not sure it can help with the heart of it. Not in the same way. The direction still has to come from somewhere. And the question of what is worth making, what purpose it serves, who it serves… those questions don’t get any easier just because the making has become more accessible.
Maybe that means the line is not “AI or no AI”. It is whether the tool is rightly ordered. The idea has to begin with me, and the final responsibility has to come back to me. AI can help in the middle. It can research, challenge, smooth and suggest. But if it starts deciding what matters, what I mean, or where the work is going, then collaboration has become surrender. If anything, the questions get harder.
When everyone can produce polished things, polish stops being impressive. The new skill isn’t fluency. It’s honesty. It’s being willing to make something alive rather than something merely finished.
Discipline does not disappear when AI removes friction. It moves higher up. It becomes less about grinding through every manual step and more about discernment, restraint and deciding what the freed time is for. Not every pause needs to become productive. Not every thought needs to be captured, improved, polished and published.
So speed probably isn’t the test. Not really. The better test is whether the thing is true, loving, useful, fruitful and Godward. Whether it serves. Whether it helps. Whether it carries something real rather than adding to the noise.
AI can open the door. We still have to walk through it.
