There’s an interesting contradiction in how we talk about AI.

There’s panic on one side. AI is coming for our jobs. Developers will be obsolete. The robots are taking over.

On the other side: people casually trying to get AI to generate a thousand startup ideas, register the domains, build landing pages, and deploy them. All in one command. As if building something valuable is just a matter of stringing together the right prompts.

We’re clearly not at “AI takes over” yet. What we are at is a weird in-between where the loudest voices are either predicting doom or treating AI like a startup slot machine.

Meanwhile, those of us actually working with AI know the reality is far less dramatic. It’s a tool. A powerful one. But “find startup ideas and just build them” skips the part where you need to understand problems, validate assumptions, and decide what’s actually worth making.

That part? Still very much a human job.