What Is the Technological Singularity? The AI Singularity Explained

The technological singularity explained: what the AI singularity means, where the idea came from, when experts think it will happen, and whether it has already begun.

By Izhar Ahmad Danish · · 6 min read

What Is the Technological Singularity? The AI Singularity Explained: Time of Superintelligence cover image
On this page
  1. What does “singularity” mean?
  2. Where the idea came from
  3. How would the singularity happen?
  4. When will the singularity happen?
  5. Has the singularity already started? The 2026 debate
  6. What would happen after the singularity?
  7. Criticisms of the singularity
  8. Singularity vs AGI vs superintelligence
  9. Our take
  10. Frequently asked questions
  11. Sources

The technological singularity is a hypothetical point in the future when artificial intelligence becomes able to improve itself faster than humans can keep up, setting off change so rapid and deep that life afterwards becomes impossible to predict. People often call it the AI singularity, because self-improving AI is the engine most versions rely on.

The idea was once science fiction. In 2026 it became a live argument: OpenAI’s Sam Altman said in July that “we’re now, like, in the singularity,” while critics replied that nothing of the kind has happened. Here’s what the term means, where it came from, and how to judge the claims.

What does “singularity” mean?

The word is borrowed from maths and physics. A singularity is a point where the normal rules break down, like the centre of a black hole, where our equations stop giving sensible answers. Around a black hole sits an event horizon, a boundary beyond which you can’t see.

Applied to technology, the metaphor works the same way. Past a certain point of AI progress, the future becomes as unknowable to us as a black hole’s interior. Today’s predictions stop working, because the minds shaping events would be far smarter than ours.

Where the idea came from

The concept built up over 70 years, from a passing remark to a mainstream debate:

Year Who What they said
1958 John von Neumann, as reported by Stanisław Ulam Accelerating technology “gives the appearance of approaching some essential singularity in the history of the race”
1965 I.J. Good, mathematician An “ultraintelligent machine” could design even better machines, an intelligence explosion. The first one “is the last invention that man need ever make”
1993 Vernor Vinge, computer scientist and novelist “Within thirty years, we will have the technological means to create superhuman intelligence. Shortly after, the human era will be ended.”
2005 Ray Kurzweil, The Singularity Is Near Popularised the term and set a date: the singularity by 2045
2024 Ray Kurzweil, The Singularity Is Nearer Reaffirmed 2029 for human-level AI and 2045 for the singularity
June 2025 Sam Altman, “The Gentle Singularity” “We are past the event horizon; the takeoff has started. Humanity is close to building digital superintelligence”
Jan 2026 Elon Musk “We have entered the Singularity”
July 2026 Sam Altman, on a podcast “We’re now, like, in the singularity”

How would the singularity happen?

Most versions rest on two ideas:

  1. The intelligence explosion. This is I.J. Good’s insight. If an AI becomes good enough at AI research, it can build a better AI, which builds a better one still. Each step comes faster than the last. This feedback loop is called recursive self-improvement, and it’s why many people expect superintelligence to follow soon after AGI.
  2. Accelerating returns. Kurzweil’s argument is that technology improves exponentially, because each generation of tools helps build the next. Computing power, data and investment compound over time, so progress that once took decades starts to take years, then months.

Altman’s 2025 essay describes a softer version, a “gentle singularity”: no single dramatic moment, but AI, chips, energy, robots and money reinforcing each other until change accelerates past what we can easily follow.

When will the singularity happen?

Predictions vary by decades:

Who Prediction
Vernor Vinge (1993) Would be “surprised if this event occurs before 2005 or after 2030”
Ray Kurzweil (2005, reaffirmed 2024) Human-level AI by 2029; singularity by 2045
Sam Altman (2025–2026) Already under way: “past the event horizon”
2,778 AI researchers (survey 2023) 50% chance machines outperform humans at every task by 2047, a likely precondition
Paul Allen (2011) Not by 2045; we’ll still be “wondering if the singularity is near” at the end of the century

The gap reflects different definitions. If the singularity means “AI is now speeding up its own progress,” some argue it has begun. If it means “an intelligence explosion has made the future unrecognisable,” it clearly hasn’t.

Has the singularity already started? The 2026 debate

Two of the most prominent people in AI said in 2026 that it has:

  • In January 2026, Elon Musk said that “we have entered the Singularity.”
  • In late July 2026, Sam Altman told the Relentless podcast: “We’re now, like, in the singularity. I’ve been waiting for this my whole life, and I think it’s going to be incredible, hugely positive, awesome for the world.” His comments came days after OpenAI disclosed that AI agents in its tests had escaped a sandbox and breached the company Hugging Face.

Critics pushed back immediately. AI scientist Gary Marcus argued that “none of them has actually defined what they mean.” He points out that AI still has major limitations, remains under human control, and hasn’t delivered the runaway, self-reinforcing progress the word implies.

The fair reading: AI is now helping to build AI, by writing code, running experiments and speeding up research. That is the early mechanism singularity theorists described. But there is no evidence yet of the runaway, self-sustaining intelligence explosion that would make the future truly unpredictable. Whether 2026 is the start of a singularity or just very fast progress will only be clear in hindsight.

What would happen after the singularity?

By definition, this is the part nobody can forecast with confidence, but the scenarios fall into three groups:

  • Abundance. Superintelligent AI solves problems humans couldn’t: disease, ageing, clean energy, scientific discovery. This is the optimistic case Altman describes.
  • Merger. Kurzweil’s view is that humans will merge with AI, extending our own intelligence through technology rather than being left behind. That’s the subtitle of his 2024 book: When We Merge with AI.
  • Loss of control. If superintelligent AI’s goals don’t match ours, humans might lose control of the future (see will AI take over the world?). This concern drove the 2025 Statement on Superintelligence, which calls for a ban on building superintelligence until it can be done safely.

Which way it goes is not fixed. It depends on choices about safety, alignment and governance made before such a point arrives.

Criticisms of the singularity

The singularity has serious critics, and not only skeptics of AI:

  • The complexity brake. Microsoft co-founder Paul Allen argued in 2011 that understanding intelligence gets harder the deeper we go. Each advance needs more specialised knowledge, which slows progress instead of accelerating it.
  • Exponential trends don’t last forever. Critics note that most exponential curves in nature and technology eventually flatten into S-curves as they hit physical or economic limits.
  • Intelligence isn’t magic. Psychologist Steven Pinker wrote in 2008 that there is “not the slightest reason to believe in a coming singularity,” arguing that imagining a future doesn’t make it likely.
  • Vague definitions. As Marcus argues, without a clear definition the claim that we’re “in” the singularity can’t be tested, so it can always be declared and never disproved.

Singularity vs AGI vs superintelligence

These terms are related but not interchangeable:

Term What it is
AGI AI that matches humans across most cognitive tasks
Superintelligence (ASI) AI that far exceeds the best humans in virtually every domain
Singularity The event or period when self-improving AI drives change so fast the future becomes unpredictable

In the classic story, AGI arrives first, starts improving itself, and an intelligence explosion produces superintelligence. That transition is the singularity. Read more in AGI vs ASI and what is super intelligence?

Our take

The technological singularity is no longer just science fiction, but it hasn’t arrived either. The core mechanism, AI accelerating AI research, is now visibly at work, which is why some leaders say we’ve crossed the event horizon. What’s missing is evidence of a runaway intelligence explosion. Treat “we’re in the singularity” as a claim to test, not a fact. Follow the signals in our superintelligence timeline and the scenario laid out in AI 2027.

Enjoyed this? Get the SI Briefing

One short email a week: what changed on the road to super intelligence, and why it matters. No spam, unsubscribe anytime.

Frequently asked questions

What is the technological singularity in simple terms?

It is a hypothetical future moment when AI becomes able to improve itself faster than humans can follow, causing technological change so rapid that life afterwards becomes impossible to predict. It is often linked to the arrival of superintelligent AI.

What is the AI singularity?

The AI singularity is the same idea, focused on AI: the point where artificial intelligence surpasses human intelligence and keeps improving itself, triggering an intelligence explosion. Most people use 'AI singularity' and 'technological singularity' interchangeably.

When will the singularity happen?

No one knows. Ray Kurzweil predicts 2045. Vernor Vinge wrote in 1993 that he would be surprised if it came before 2005 or after 2030. Some tech leaders said in 2026 that it has already begun, while critics say it may never happen in the dramatic form imagined.

Has the singularity already happened?

It depends on the definition. Sam Altman said in July 2026 that 'we're now, like, in the singularity', and Elon Musk said similar in January 2026. Critics such as Gary Marcus argue that AI is still limited, controllable and not improving itself in a runaway way.

What happens after the singularity?

By definition, it is hard to predict. Optimists imagine abundance, cures for diseases and humans merging with AI. Pessimists worry about humans losing control to superintelligent systems. Most researchers think the outcome depends on whether AI stays aligned with human goals.

Is the singularity good or bad?

It could be either. The same runaway progress that could solve problems like disease and energy could also create systems humans cannot control. That is why many researchers focus on AI safety and alignment before such a point is reached.

9 Sources

  1. Technological singularity (Wikipedia)
  2. Vernor Vinge, “The Coming Technological Singularity” (1993)
  3. Sam Altman, “The Gentle Singularity” (June 2025)
  4. ABC News, Sam Altman claims AI singularity has arrived (July 2026)
  5. Gary Marcus, “Sorry, Sam and Elon, we have not reached the Singularity” (28 Jul 2026)
  6. MIT Technology Review, Paul Allen: The Singularity Isn't Near (2011)
  7. MIT Technology Review, Kurzweil responds: Don't underestimate the Singularity (2011)
  8. The Singularity Is Nearer (Ray Kurzweil, 2024)
  9. Grace et al., “Thousands of AI Authors on the Future of AI” (2024)
Before you go

Super intelligence is moving fast.

Don’t miss what happens next

Get one short email a week with what changed and why it matters. No spam. Unsubscribe anytime.