By Hadi Brenjekjy

There’s something wild about watching a human grow.

When my son was born, he was all instinct..tiny fists, milk-drunk sighs, and total dependence. I was his entire interface to the world. He didn’t know how to sit up, or that night and day were different things, or that spoons weren’t chew toys. Every need was urgent. Every cry, a system crash.

But here’s what I noticed over time:

He didn’t need perfect care. He needed predictable, “good enough” care.

When he was a newborn, I responded instantly. But later? I’d wait a few seconds before picking him up from the crib. Let him try. Sometimes he would self-soothe. Other times, I’d still step in, but later than before.

This wasn’t neglect. It was design.

Tiny doses of frustration. Tiny experiments in autonomy.

That’s how he learned:
1. That he can survive without instant help.
2. That the world doesn’t end when things don’t go his way.
3. That not every signal gets a response, and that’s okay.

The gaps between his need and my response were the training ground for independence.

Now here’s the twist:
I’m watching the same thing happen again, but this time, not with my son.
With my LLM and AI tools 

So what is Artificial General Intelligence (AGI)?

Let’s zoom out for this. Think of today’s AI like a really smart intern who only works when you ask. You tell it what to do, it does the thing, then waits for the next instruction.

AGI – Artificial General Intelligence, is the moment that intern becomes a teammate.
It doesn’t just wait for orders. It learns, adapts, makes decisions, asks smart questions, remembers past projects, and gets better with time. It thinks and acts like a generalist human, not just a task-doer, but a goal-seeker.

AGI is NOT about creating a robot with feelings. It’s about building a system smart and capable enough to handle anything you would reasonably hand to a colleague. From writing a memo to planning a strategy, without needing to be babysat.

Back to the story.. 

When I first started using AI tools, it was like managing a baby. I’d summon them when needed, feed them prompts, and close the session once I got what I wanted. Total micromanagement. Like handing my toddler a toy, watching every move, and saying “Wow!” when he didn’t eat it.

But lately? Things are shifting.

My AI agents are starting to do things without me hovering. They remember. They propose. They check in. Sometimes I leave them running and come back to progress. They are starting to live in the in-between moments, those small gaps where I’m not watching, but they are still thinking.

Sound familiar?

It’s the same arc. From full dependency to “leave it for a bit and see what happens.”

Which brings me to how I now define AGI:

AGI is not when an AI becomes perfectly human. It’s when it makes economic and cognitive sense to leave it on.

Like a human teammate. Like a child who’s grown past the stage of needing constant supervision.

We will know we have hit AGI when we choose not to turn the agent off. When resetting it becomes inefficient. When keeping it running becomes the norm, not the exception.

That’s the milestone. Not fooling us with clever language. Not beating us at chess. Not passing a moving benchmark of “economically valuable work.”

It’s when the agent earns its right to stay awake.

And to do that, it will need everything my toddler’s learning too:

• Memory: What just happened, what worked, what didn’t.
• Adaptability: If the toy breaks, find a new one. If the plan fails, try again.
• Goal generation: Wanting to try. Wanting to explore.
• Initiative: Not waiting to be told.
• Trust: The deep sense that I can step back, and it won’t all go sideways.

None of these are binary. They are messy, overlapping, and context-dependent. Just like raising a human.

We are not there yet.

But we are crawling.
Then we will toddle.
Then we will run.

And soon, the day will come when turning your AI off will feel as counterproductive as unplugging your Wi-Fi between emails.

Raising humans is hard.
Raising agents might be harder.

But when done right?
They don’t just do what you say.
They grow.

Just like my son.
Just like the LLM I’m training.