AGI: Are We There Yet?

For years, we’ve been told that Artificial General Intelligence is just around the corner.

2025,2026,2027

The date keeps moving.

But here we are.

Are we there yet?

The answer depends on what you mean by AGI.

If AGI means an AI that can write code, analyse documents, create images, reason through complex problems, use tools and perform tasks that previously required highly educated humans…

Then honestly, we’re getting uncomfortably close.

But if AGI means a machine that can generally understand the world, learn new things independently, adapt to unfamiliar situations and perform across almost any intellectual task at human level or above, then the answer is still:

No.

At least, not convincingly.

The problem with the AGI debate

Everyone seems to have their own definition.

One company releases a new model and calls it a major step toward AGI.

Another says we’re already basically there.

Someone else says today’s models are nothing more than sophisticated autocomplete.

And somewhere in the middle, the actual technology keeps improving.

The interesting thing is that you don’t necessarily need AGI for AI to completely transform society.

That’s the part people may be underestimating.

You don’t need a machine that can do everything.

You just need machines that can do enough.

If AI can perform 30% of the work done by knowledge workers, that’s already a massive economic change.

If agents can operate software, communicate with customers, write and maintain code, analyse financial information and execute business processes, entire industries start changing before anyone agrees that AGI has arrived.

Maybe we’re asking the wrong question

Instead of asking:

“Have we reached AGI?”

Maybe we should ask:

“How much human work can AI already replace, augment or automate?”

Because that’s the metric businesses actually care about.

A company doesn’t care whether an AI system officially qualifies as AGI.

It cares whether that system can do the job.

And this is where AI agents become particularly interesting.

The next stage isn’t necessarily a smarter chatbot.

It’s an AI that can act.

It can receive a goal, break it into tasks, use software, access information, make decisions and keep working until the objective is completed.

That’s much closer to having a digital worker than having a chatbot.

And then comes the uncomfortable part

What happens when AI becomes good enough to perform most computer-based work?

What happens to junior developers?

Accountants?

Customer support?

Researchers?

Designers?

Analysts?

Marketing teams?

What happens when one person with a collection of AI agents can do the work that previously required an entire department?

That’s when the AGI conversation stops being a futuristic debate.

It becomes an economic question.

A political question.

A question about ownership.

Because if AI dramatically increases productivity but the infrastructure is controlled by a handful of companies, the benefits aren’t automatically going to be distributed equally.

We could end up with incredible abundance.

Or incredible concentration of wealth and power.

Possibly both.

So, are we there yet?

No.

But maybe that’s not the most important part.

The more important reality is that we are already entering the period that everyone thought AGI would create.

AI is increasingly capable of reasoning, coding, creating, researching and acting.

The boundary between “AI assistant” and “AI worker” is becoming thinner.

And eventually, the argument over whether we have technically reached AGI may become almost irrelevant.

Because by then, we’ll already be living with the consequences.

The question won’t be:

“Is this AGI?”

It will be:

“What happens now that machines can do this?”