Generative AI is a literacy, not a tool.

When people ask whether they should "start using AI," I hear the same framing every time: AI as a tool you pick up when needed and set down when you're done. Like a calculator. Like Excel.

I think that framing is wrong, and I think it's why a lot of smart, capable people are falling behind faster than they realize.

Tools vs. literacies

A tool changes what you can do. A literacy changes what you can think.

The printing press was not just a tool for making more books. It was a literacy — a shift in how knowledge was authored, distributed, challenged, and remembered. For a century after Gutenberg, the people who won were not the ones who owned the best presses. They were the ones who first internalized what the new information environment made possible.

Generative AI is closer to that. Not a drill you borrow from the shed, but a new medium you become fluent in — or fail to.

What fluency looks like

People who are becoming fluent share a few habits:

• They think with the model, not through it. They treat the model as a colleague whose first draft is often wrong but whose sparring produces something neither of them would have reached alone.
• They develop taste. Fluency isn't prompt engineering. It's knowing when an output is almost right and articulating the small thing that's off.
• They leave a lot of the work undone. Good AI users do less. They let the model take the mechanical 80% and spend their attention on the 20% where judgment actually matters.

What illiteracy looks like

The opposite pattern, which I see in training rooms across Indonesia:

• Asking the model to "just do it" and shipping whatever comes out.
• Using the model to imitate the work of people who know what they're doing, without developing that knowing yourself.
• Treating a wrong answer as the model's failure rather than your own unclear question.

The first pattern produces people whose work compounds. The second produces people who are permanently one model-version behind.

The Indonesian opportunity

For the first time in a long time, the starting line is roughly even. Nobody — not Jakarta, not San Francisco — has had more than a few years with these tools. The advantage does not go to whoever had the biggest head start. It goes to whoever builds the fastest feedback loops.

That's a fair fight. We should take it.