There’s an AI singer charting right now, and everyone knows it’s fake. The voice was synthesized, the face was generated, the “artistic journey” was a prompt. And people are still streaming it. Not ironically. Not as a novelty. They just… like the song. I’ve been thinking about what that means, and I think it’s the most honest thing that’s happened to music in years — because it strips away the last pretense that we were ever really buying the humanity. We were buying the feeling. The machine just got better at delivering it.

This is the part where someone usually argues that we need to fight back. Protect the human. Resist the tide. I used to think that too. I don’t anymore.

Every Time We’ve Tried to Outwork a Machine, We’ve Lost

When spreadsheets arrived, accountants worked longer hours to prove their value. When ATMs rolled out, bank tellers unionized to hold the line. When search engines indexed the world, encyclopedias printed more volumes. The story always ends the same way, and it’s not because the humans weren’t trying hard enough. It’s because you can’t out-effort a force multiplier. That’s what a force multiplier is.

The people who came out ahead in each of those transitions weren’t the ones who worked harder at the thing the machine was taking over. They were the ones who moved fast enough to catch what fell off the back of the truck — the work the machine created by existing, the problems it surfaced by solving the previous ones, the new human layer that every wave of automation seems to need around its edges.

When ATMs spread across every corner, the number of bank tellers actually went up for a while. Not because banks needed more tellers to do teller things. Because branches got cheaper to run, so banks opened more of them, and someone had to manage the customer relationship the machine couldn’t fake. The machine didn’t replace the human. It changed which humans were valuable and why.

Getting Out of the Way Is a Skill

Here’s what I’ve noticed, building software for the better part of two decades: the engineers who struggle most with AI tooling are the ones trying to supervise it like a junior developer. They’re reviewing every line, correcting the style, nudging it toward how they would have written it. They’re treating the tool like a fast intern instead of a different kind of intelligence with different strengths and a very different failure mode.

The ones who’ve adapted — and I mean actually adapted, not just added Copilot to their existing workflow — have learned to operate at a different altitude. They’re thinking about architecture, about what to build and why, about the interfaces between systems that require judgment the model doesn’t have yet. They’re not doing less. They’re doing different. And they got there by getting out of the machine’s way fast enough that they could see what it exposed.

That’s a skill. It doesn’t feel like one because it looks like stepping back. It looks like letting go. In my experience, that’s the hardest thing to get engineers to do — and honestly, the hardest thing to get myself to do.

The AI Singer Already Won

Back to the singer. What’s actually happening there isn’t that AI has gotten good enough to fool us. We know it’s fake. What’s happening is that we’ve quietly admitted that the thing we wanted was never the struggle behind the art — it was the resonance the art created. The song that makes you feel something on a Tuesday afternoon. The voice that lands exactly right. We dressed that up in the mythology of suffering and authenticity for a long time because we needed the story to justify the price. The machine just made the story optional.

Some people will find that depressing. I find it clarifying.

The musicians who will matter in ten years won’t be the ones who made the most technically perfect thing, or even the most “human” thing in some defensive, provable sense. They’ll be the ones who figured out what the machine can’t do — not as a clever workaround, but as a genuine creative position. The live experience it can’t replicate. The community it can’t sustain. The specificity of a person, in a place, at a time, saying something that couldn’t have come from a prompt.

That’s not a consolation prize. That’s always been the thing that mattered. We just had to lose the parts that didn’t to see it clearly.

The Question I Keep Sitting With

If the machine is going to do the thing anyway — and it is — the only real decision left is whether you move early enough to be part of what comes next, or late enough that you’re defending what’s already gone.

I don’t think that’s a question with a clean answer. I think it’s a question worth sitting with for a while before you decide what you’re protecting and why.