The Singularity Is Near and Nearer: Kurzweil's Books and Track Record

The Singularity Is Near (2005) and Nearer (2024) summarized: Ray Kurzweil's key ideas, his 2029 and 2045 dates, and his prediction track record.

By Izhar Ahmad Danish · · 6 min read

The Singularity Is Near and Nearer: Kurzweil's Books and Track Record: Time of Superintelligence cover image
On this page
  1. The two books at a glance
  2. What is The Singularity Is Near about?
  3. What is The Singularity Is Nearer about?
  4. Kurzweil’s prediction track record
  5. The 2029 bet with Mitch Kapor
  6. What critics say
  7. How Kurzweil’s view differs from the “intelligence explosion”
  8. Our take
  9. Frequently asked questions
  10. Sources

The Singularity Is Near (2005) and its sequel The Singularity Is Nearer (2024) are Ray Kurzweil’s two books arguing that technology improves exponentially and that humans will merge with artificial intelligence. They are the source of the two dates you see in almost every article about the technological singularity: human-level AI by 2029, the Singularity by 2045. This page summarises both books and checks how Kurzweil’s predictions have held up.

Short answer: The Singularity Is Near (Viking, 2005) argues that a “law of accelerating returns” will bring human-level AI by 2029 and a merger of human and machine intelligence by 2045. The Singularity Is Nearer (Penguin, June 2024) updates the case for the age of ChatGPT and keeps both dates. Kurzweil’s record is mixed but better on AI than on most topics.

The two books at a glance

The Singularity Is Near The Singularity Is Nearer
Subtitle When Humans Transcend Biology When We Merge with AI
Published September 2005, Viking 25 June 2024, Penguin (delayed from 2021)
Length Long, data-heavy 419 pages
Core claim Exponential growth in computing, genetics, nanotech, robotics Same, with LLMs as evidence that the curve is on schedule
Key dates Turing test passed by 2029; Singularity 2045 Human-level AI 2029; merger with AI 2045
Tone Visionary, technical More about life getting better and the LLM era
Spin-offs Films Transcendent Man and The Singularity Is Near (2012) —

Sources: Wikipedia entries for both books.

What is The Singularity Is Near about?

Short answer: exponential technological growth, and what happens when it reaches intelligence itself.

Kurzweil’s argument rests on a few ideas:

  • The law of accelerating returns. Information technologies improve exponentially because each generation of tools is used to build the next. Progress feels slow, then sudden.
  • Paradigm shifts as S-curves. Any single technology levels off, but a new one takes over, so the overall trend stays exponential (vacuum tubes, transistors, integrated circuits, and whatever comes next).
  • GNR. Three overlapping revolutions: Genetics, Nanotechnology and Robotics, by which he means strong AI.
  • Reverse-engineering the brain. Brain scanning will show how intelligence works, which will feed into AI design.

From these he derives specific forecasts: about $1,000 of computing matching a human brain around 2020, machines passing the Turing test by 2029, hardware for brain uploading around 2040, and in 2045 a Singularity in which $1,000 of computing buys about a billion times the power of all human brains combined today.

Kurzweil’s Singularity is optimistic. He does not imagine machines replacing people; he imagines people enhancing themselves until the line between human and machine intelligence disappears.

What is The Singularity Is Nearer about?

Short answer: the same thesis, updated for large language models, with the same dates.

Kurzweil started the sequel years before ChatGPT, and the release slipped from 2021 to 2022 and then to June 2024. By the time it appeared, LLMs gave him his strongest evidence yet that his 2029 date was realistic. The book:

  • repeats that AI will reach human-level intelligence by 2029;
  • predicts that by 2045 people will merge with AI, expanding human consciousness rather than being replaced;
  • argues at length that human life has improved dramatically over decades and centuries, and will keep improving;
  • presents the Singularity as a positive development overall, with the perils of AI treated as problems to manage.

Reviews were less warm than for the first book. Becca Rothfeld in The Washington Post called it hard to summarise, unfocused and oversimplified, and said some predictions read like religious texts.

Kurzweil’s prediction track record

Short answer: some striking hits, several clear misses, and a self-reported 86% success rate that critics dispute.

Here are the predictions most often cited, with what actually happened.

Prediction Made in Target Outcome
A computer will beat the world chess champion The Age of Intelligent Machines (1990) “by the year 2000” ✅ Early: Deep Blue beat Kasparov in May 1997
Rapid growth of worldwide internet use 1990 book 1990s ✅ Happened
A US company worth over $1 trillion 1999 forecasts for 2009 2009 ❌ Not by 2009 (cited by Newsweek’s Daniel Lyons)
Most text created by continuous speech recognition 1999 forecasts for 2009 2009 ❌ Did not happen
Self-driving cars using sensors in highways 1999 forecasts for 2009 2009 ❌ Not by 2009
A 20-petaflop supercomputer 1999 forecasts for 2009 2009 ⚠️ Disputed: Kurzweil said Google’s network counted
Overall 2009 forecasts 1999 2009 ⚠️ Forbes: 7 wrong, 4 partly right, 1 right. Kurzweil (2010): 115 of 147 “entirely correct”, 86% correct or “essentially correct”
$1,000 of computing matches a human brain The Singularity Is Near (2005) ~2020 ⚠️ Depends on how brain compute is estimated; no agreed verdict
A machine passes the Turing test The Age of Spiritual Machines (1999) 2029 ⏳ Pending formally (Long Bet runs to 2029); GPT-4.5 was judged human 73% of the time in a 2025 three-party test
Solar power meets all energy needs 2008 NAE panel ~2028 ⏳ Not on track as of 2026
Medicine adds a year of life expectancy per year 2013 interview ~2028 ⏳ Pending
Human-level AI 2005 and 2024 books 2029 ⏳ Pending; now within mainstream lab forecasts
The Singularity: humans merge with AI 2005 and 2024 books 2045 ⏳ Pending

Sources: Predictions made by Ray Kurzweil (Wikipedia), Kurzweil’s 2010 self-assessment, Long Bets, Jones & Bergen (2025). ✅ came true, ❌ missed, ⚠️ disputed, ⏳ not yet due.

Two patterns stand out. Kurzweil is strongest on raw computing trends and weakest on how fast society adopts technology: the 2009 misses are mostly about deployment (speech typing, smart highways), not about what computers could do. And his self-grading is generous; “essentially correct” does a lot of work in the 86% figure.

The 2029 bet with Mitch Kapor

In 2002 Kurzweil and Lotus founder Mitch Kapor made the first bet on Long Bets. Kapor predicted that “By 2029 no computer - or ‘machine intelligence’ - will have passed the Turing Test.” The stakes are $20,000, going to the Electronic Frontier Foundation if Kapor wins or the Kurzweil Foundation if Kurzweil wins. The bet uses detailed rules, including long interviews and expert judges, so a five-minute online study like the 2025 GPT-4.5 experiment does not settle it. Still, few people would now bet heavily on Kapor.

What critics say

Short answer: exponential curves flatten, the brain is harder than he thinks, and the vision is quasi-religious.

  • Curves don’t stay exponential. Critics such as Theodore Modis argue logistic (S-shaped) curves fit real processes better than endless exponentials.
  • The brain is hard. Microsoft co-founder Paul Allen wrote in MIT Technology Review (2011) that “the Singularity isn’t near” because understanding cognition requires a “complexity brake”: the deeper we look, the more complex the brain turns out to be. Neuroscientist David Linden argued that Kurzweil confuses collecting data with understanding biology.
  • Rapture for nerds. Alex Beam called the vision a “Buck Rogers” version of the Christian Rapture; philosopher John Gray tied it to apocalyptic myths.
  • Too rosy. Janet Maslin in The New York Times found much of it “pie in the sky” and focused on upside over risk. A reviewer in Philosophy judged Nick Bostrom’s Superintelligence more realistic, largely because Bostrom takes the dangers seriously.

How Kurzweil’s view differs from the “intelligence explosion”

Kurzweil and the intelligence explosion school both expect AI to surpass humans, but the stories differ. Kurzweil sees a smooth exponential, predictable decades ahead, ending in merger: we become the superintelligence. I.J. Good, Vernor Vinge and Bostrom describe a possible sudden jump driven by AI improving itself, ending with a separate super intelligence that may or may not share our goals. That difference explains why Kurzweil is an optimist and Bostrom a worrier.

Our take

Twenty years on, The Singularity Is Near reads better than most futurism of its era. The 2029 date for human-level AI looked eccentric in 2005; by 2026 it sits inside the range many lab leaders give, as our superintelligence timeline shows. The weak spots are the same ones critics named: Kurzweil underrates how slowly institutions change and how little the exponential story says about risk. Read Near for the argument, Nearer for the update, and keep the track-record table in mind when weighing the 2045 date.

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 Singularity Is Near about?

Ray Kurzweil's 2005 book argues that computing, genetics, nanotechnology and AI improve exponentially, so human-level AI arrives around 2029 and humans merge with machine intelligence around 2045, an event he calls the Singularity.

What is the difference between The Singularity Is Near and The Singularity Is Nearer?

Near (2005) set out the theory and the 2029 and 2045 dates. Nearer (2024) is an update written after the rise of large language models. It keeps both dates, adds evidence that life is improving, and frames the Singularity as humans merging with AI.

What year does Kurzweil predict the singularity?

2045. He predicts AI will reach human level by 2029, and that by 2045 humans will have merged with AI, multiplying our intelligence many times over.

How accurate are Ray Kurzweil's predictions?

Mixed. Kurzweil claimed in 2010 that 86% of 147 predictions were correct or essentially correct. Forbes judged his 2009 predictions mostly wrong. His early chess forecast came true, and LLMs have made his 2029 human-level-AI date look less far-fetched than it did in 2005.

Is The Singularity Is Near worth reading in 2026?

Yes as history and as the clearest statement of the exponential-trend argument behind today's AI optimism. For a current view, read The Singularity Is Nearer, which covers the LLM era.

8 Sources

  1. The Singularity Is Near (Wikipedia)
  2. The Singularity Is Nearer (Wikipedia)
  3. Predictions made by Ray Kurzweil (Wikipedia)
  4. Ray Kurzweil, “How My Predictions Are Faring” (2010, PDF)
  5. Long Bets #1: Kapor vs Kurzweil, Turing test by 2029
  6. Jones & Bergen, “Large Language Models Pass the Turing Test” (arXiv, 2025)
  7. MIT Technology Review, Paul Allen: The Singularity Isn't Near (2011)
  8. Superintelligence: Paths, Dangers, Strategies (Wikipedia), incl. comparison with Kurzweil
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.