> For the complete documentation index, see [llms.txt](https://dyns.ntd.asia/power-dynamics/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://dyns.ntd.asia/power-dynamics/artificial-intelligence/what-is-ai-really.md).

# What Is AI, Really?

## What Is AI, Really? Thinking About the Goal of "Human-Like"

I keep coming back to a simple question: what is Artificial Intelligence actually trying to become?

Not in the marketing sense — not "smarter search" or "your new productivity assistant." I mean the deeper goal underneath all of it. Strip away the branding and I think the answer is this: AI wants to become human-like.

That's a strange goal when you sit with it. So I've been trying to work out what "human-like" actually requires, piece by piece.

### Start with what makes a human...

A person isn't just a body walking around. A human has characteristics — personality, quirks, ways of seeing the world. A human has knowledge, built up over years of reading, listening, paying attention. And a human has experience, the messy kind you only get by living through something, not reading about it.

None of that shows up overnight. A newborn has none of it. Everything a person becomes, they build.

And here's the part I think matters most: humans don't just passively accumulate these things. They need them, actively, in order to function. To learn. To work. To build things — a career, a family, a house, a business, a life.

### Needs turn into skills, and...

Wanting to learn isn't the same as being able to. So humans develop a learning skill — the ability to absorb something new and make it stick.

Wanting to work isn't the same as being employable. So humans develop a working skill — showing up, collaborating, delivering.

Wanting to build isn't the same as being able to construct something that lasts. So humans develop a building skill — turning an idea into something real, something that survives contact with the world.

Skill is what sits between wanting something and having it.

### Now put AI next to that

If AI's ultimate goal is to become human-like, then it needs to walk the same path. It needs knowledge — and this part it's already good at, arguably better than any single human, because it can absorb more text than a person could read in a hundred lifetimes.

But knowledge by itself doesn't make you human-like. It makes you a library. A very fast, very well-organized library — still just a library.

The harder part is what comes next. Experience. Skill. The ability to not just know something, but to have done it, gotten it wrong, adjusted, and done it again.

This is why I think the current shift toward "agentic" AI — systems that take actions instead of just answering questions — is the real story of this moment. It's not a feature update. It's AI attempting the second half of the human formula: turning knowledge into something it has actually practiced.

### Where I land, for now and then...

I don't think AI is human-like yet. Not close. But I think the framework is useful precisely because it's simple: knowledge, then need, then skill, built through doing.

If that's the real yardstick, then the interesting question for the next few years isn't "how much does AI know?" We already know the answer to that — a lot, and growing fast.

The interesting question is: can it earn skill the way we do? Through repetition. Through failure. Through actually building something and watching whether it holds up.

That's the part I'll be watching.

**\[NTD]yns.Asia**