Every era had its tool. This one is ours.

The staircase nobody stops to look at

There was a moment when building something physical required massive labor, capital, and years. Then industrial machinery arrived and the cost of production collapsed. Human work didn't disappear, but who could do what changed entirely. The same happened with the internet: suddenly, starting a business didn't require a storefront or a big team. A computer and a connection were enough.

AI feels to me like the next step on that staircase. What I find interesting is what changes in the relationship between an idea and its execution.

Friction is where ideas go to die

Before, if you came up with an app to solve something nobody had figured out well, you needed someone who could code, time to explain what you wanted, money to pay for it, and luck that what they built would actually resemble what you had in your head. Now I can describe that idea, iterate out loud, and arrive at something functional without mastering any programming language. Not because AI does it for me, but because I can do it with it.

That's what feels like the real shift to me. Not that AI is smart. But that it reduces the friction between what you think and what you can build. And that friction, historically, is where most ideas have died.

AI brings your imagination closer to reality so you can create. Yours doesn't.

The genie isn't the best model

I prefer the copilot analogy over the genie in the lamp. The genie does exactly what you ask. The copilot also helps you execute, but knows the territory and sometimes says "there's a storm that way, are you sure?" That version feels more honest about what good AI should be, and also about what it's already becoming for those who use it well.

The step is already there

I don't know if we're at the beginning of something enormous or at the peak before a correction. But I do know that the people building things today with tools that didn't exist two years ago make me think the step is already there.

What I notice in the people getting the most out of AI isn't that they're more technical. It's that they're willing to describe clearly what they need, to iterate, and not to assume the first answer is the final one. That doesn't require technical training. It requires clarity about what you want, and that's always been the hard part.

The question is whether you're climbing it or watching from below.