Are We Really Approaching the Singularity?

The word singularity has crept back into the public conversation, not from science fiction writers, but from the people building today’s most advanced artificial intelligence systems. When Sam Altman recently suggested that we may already be “in the singularity,” it sparked a fresh wave of excitement, anxiety, and confusion. So what’s actually going on? And are we really standing at the edge of something world‑changing?

What is the singularity

The singularity is the hypothetical moment when artificial intelligence becomes smarter than humans and begins improving itself at a pace we can’t match or manage. It’s the point where technological progress becomes unpredictable and potentially unstoppable.

Why people are talking about it again

AI systems have made huge leaps in the last few years. They can write code, reason across long tasks, operate as agents, and interact with tools. Some early experiments even show models taking initiative in ways that look autonomous. Combine that with massive investment in data centres and specialised hardware, and it’s no surprise that some leaders feel like something big is happening.

But we could argue that capability growth, even rapid growth, isn’t the same as a runaway intelligence explosion.

Where AI stands today

Let's briefly consider the following.

• AI is advancing quickly, especially in planning, reasoning, and multimodal understanding.
• Agentic AI is emerging, allowing systems to act, not just respond.
• Investment is enormous, pushing hardware and data pipelines to new scales.
• Some bottlenecks remain - energy, data quality, reliability, and alignment are still hard limits.
• And AI is jagged - brilliant in some areas, but still a bit clumsy in others.

These are signs of acceleration.

Why experts disagree

Predictions for artificial general intelligence (AGI) and singularity range from “a few years” to “several decades.” Some experts see exponential curves and assume they’ll continue without diminishing returns. Others point out that intelligence isn’t just more compute. It is stability, reasoning, autonomy, and self‑direction, none of which current systems truly possess.

The disagreement itself is a sign of how uncertain this territory is.

The real frontier: alignment and control

If AI ever does cross into self‑improving territory, the biggest challenge won’t be intelligence. It will be alignment. We need systems that are reliable, predictable, and safe, before they become more capable than us.

So, are we in the singularity?

I would argue, not yet. I believe what we’re experiencing is pre‑singularity acceleration, a period where AI feels like it’s moving faster than society can fully absorb. The singularity remains a possibility, not a present reality. The real question isn’t when it arrives. It’s whether we can build AI that’s aligned, safe, and beneficial before capabilities outrun our ability to guide them.