What is the technological singularity?
Where the idea of a technological singularity comes from (von Neumann, Good, Vinge, Kurzweil), what it claims, and why it is not the same thing as AGI, superintelligence or an intelligence explosion.
In short: The technological singularity is the idea of a point beyond which technological change, driven by intelligence greater than humanity’s, becomes so fast and so profound that the future can no longer be predicted from the present. The term was made famous by the computer scientist and novelist Vernor Vinge in 1993, building on remarks attributed to John von Neumann and on I. J. Good’s idea of an intelligence explosion. It is not a synonym for AGI or superintelligence: those describe kinds of system, while the singularity is a claim about what might happen to the world after such systems exist. It is a hypothesis, often used loosely, and Vinge’s own dated prediction has not come true.
Where the word comes from
“Singularity” is a mathematician’s word. In mathematics and physics it marks a point where a quantity becomes infinite or a model stops giving meaningful answers, as at the centre of a black hole. Applied to technology, the metaphor is the second sense: a point past which our models of the future break down.
The first recorded use in this sense comes from the mathematician Stanislaw Ulam. In a 1958 tribute to his colleague John von Neumann, Ulam recalled a conversation about “the ever accelerating progress of technology”, which gave the appearance of “approaching some essential singularity in the history of the race” beyond which human affairs “could not continue”. Von Neumann was talking about technology in general, not specifically about artificial intelligence.
The link to machine intelligence came from I. J. Good in 1965. He argued that a machine better than humans at designing machines could design ever better ones, producing an “intelligence explosion”. Good did not use the word singularity, but his argument supplied the mechanism later writers relied on; our intelligence explosion page sets it out.
Vinge, 1993
Vernor Vinge, a mathematician at San Diego State University and a science-fiction writer, joined the two ideas. In a paper given at a NASA symposium in March 1993, “The Coming Technological Singularity”, he wrote: “Within thirty years, we will have the technological means to create superhuman intelligence. Shortly after, the human era will be ended.”
Vinge’s point was partly about forecasting. If entities smarter than humans began driving change, human models of the future would stop working. He listed four possible routes: computers that become “awake” and superhumanly intelligent, large computer networks that “wake up” as a superhuman entity, human-computer interfaces so close that their users could be considered superhumanly intelligent, and biological improvements to human intellect. He gave a range for his own forecast, writing that he would be surprised if the event occurred before 2005 or after 2030.
The thirty years ran out in 2023, and there are now fewer than four years left of his range. Superhuman intelligence, in the sense Vinge meant, has not arrived. Whether he was early or wrong in kind is still argued over.
The popular version: Kurzweil
The inventor and author Ray Kurzweil turned the idea into a bestseller. In The Singularity Is Near (2005) he argued that progress in computing follows predictable exponential trends, and set dates: computers passing the Turing test by 2029, and the singularity, which he described largely in terms of humans merging with machine intelligence, by 2045. His 2024 sequel, The Singularity Is Nearer, kept both dates.
Kurzweil’s version differs from Vinge’s in tone and content. It is optimistic, it puts human enhancement at the centre, and it rests heavily on extrapolating trends. These are forecasts by one author, not findings, and many researchers regard the precision of the dates as unjustified.
Four terms that are not the same
The singularity is often used as if it meant AGI, superintelligence or an intelligence explosion. It does not.
| Term | What kind of thing it is | In one line |
|---|---|---|
| AGI | A kind of system | AI that can do roughly any intellectual task a person can |
| Superintelligence | A kind of system | AI that greatly exceeds the best humans in virtually all domains |
| Intelligence explosion | A hypothesised process | AI improving AI in an accelerating loop that could produce superintelligence quickly |
| Technological singularity | A hypothesised event or period | The point at which change driven by greater-than-human intelligence makes the future unpredictable |
The relationships between them are claims, not definitions. An intelligence explosion is one proposed route to superintelligence, but superintelligence might in principle arrive slowly. A singularity is one proposed consequence of superintelligence, but some researchers who think superintelligence possible doubt there would be a single, sharp break. And AGI could exist without leading to either: whether it would, and how fast, is the subject of AI takeoff.
How the word is used now
Among AI researchers the word is used less than it was. Technical and policy writing in 2025 and 2026 tends to talk about specific, measurable things instead: how much of AI research is automated, how fast capabilities are improving, whether an intelligence explosion is under way. Where “singularity” does appear in research, it is usually in that narrower sense. A July 2026 paper from Epoch AI, for example, argued that limits on how much research can usefully be done in parallel could “delay a technological singularity” even after AI research is automated.
In popular writing and in marketing the word is used more freely, sometimes for any rapid advance in AI. When you meet it, it is worth asking which claim is meant: that AI will become far more capable than people, that this will happen in a runaway way, or that the result will be beyond prediction. Those are three different claims, and each needs its own evidence.
The September 2026 decision by the US administration to rename artificial intelligence as “Super Intelligence” has no bearing on the singularity. It renames existing technology; our dated timeline records what was done.
Sources
- Stanislaw Ulam, “John von Neumann 1903–1957”, Bulletin of the American Mathematical Society, vol. 64, no. 3, 1958, pp. 1–49.
- I. J. Good, “Speculations Concerning the First Ultraintelligent Machine”, Advances in Computers, vol. 6, 1965, pp. 31–88.
- Vernor Vinge, “The Coming Technological Singularity: How to Survive in the Post-Human Era”, VISION-21 Symposium, NASA Lewis Research Center and Ohio Aerospace Institute, 30–31 March 1993.
- Ray Kurzweil, The Singularity Is Near: When Humans Transcend Biology (Viking, 2005).
- Ray Kurzweil, The Singularity Is Nearer: When We Merge with AI (2024).
- Nick Bostrom, Superintelligence: Paths, Dangers, Strategies (Oxford University Press, 2014).
- Phil Trammell, Epoch AI, “Even after R&D is automated, parallelization constraints could delay a technological singularity”, 29 July 2026.
Common questions
- Who coined the technological singularity?
- The idea goes back to remarks by John von Neumann, recalled by Stanislaw Ulam in 1958. The computer scientist and novelist Vernor Vinge made the term famous in a 1993 paper, predicting superhuman intelligence within thirty years.
- Is the singularity the same as superintelligence?
- No. Superintelligence describes a kind of system: AI that greatly exceeds the best humans in virtually all domains. The singularity is a claim about what might happen to the world after such systems exist: change so fast that the future can no longer be predicted.
- When is the singularity predicted?
- Vinge said he would be surprised if it came before 2005 or after 2030. Ray Kurzweil has predicted 2045. These are individual forecasts, not findings, and Vinge's thirty-year forecast has already passed without the event.