Has the 'gentle singularity' already begun? And when did the singularity become 'gentle'?
Anyone talking about the “singularity” is using a very specific word. It’s not a term for technology getting awesome. It doesn’t mean people can now generate a Totoro manqué faster than the speed of light. It’s about the theoretical moment when humanity loses control of civilization to an artificial intelligence so advanced that gains total dominion over the world.
Don’t take my word for it. Ask the late Vernor Vinge, whose 1993 essay “The Coming Technological Singularity” coined the term in this context. He saw the Singularity — it earned the capital letter in his mind — as a cataclysmic event for Homo sapiens:
…we are entering a regime as radically different from our human past as we humans are from the lower animals…an exponential runaway beyond any hope of control…Any intelligent machine of the sort [that achieved the Singularity] would not be humankind’s “tool” — any more than humans are the tools of rabbits or robins or chimpanzees…If the Singularity cannot be prevented or confined, just how bad could the Post-Human era be? Well…pretty bad. The physical extinction of the human race is one possibility. (Or as Eric Drexler put it of nanotechnology: Given all that such technology can do, perhaps governments would simply decide that they no longer need citizens!) Yet physical extinction may not be the scariest possibility. Again, analogies: Think of the different ways we relate to animals. Some of the crude physical abuses are implausible, yet…
Or hey, let’s ask the chatbots currently auditioning for the part. GPT-4o: “In the context of technological advancement, the ‘singularity’ refers to a hypothetical future point at which technological growth becomes uncontrollable and irreversible, resulting in unforeseeable changes to human civilization.” In fact, that exact phrase — uncontrollable and irreversible — appeared in the definitions given to me by Gemini 2.5 Pro, Perplexity Sonar Pro, Qwen3 32B, and Deepseek R1. (Claude 4 Opus preferred “unpredictable and irreversible.”)
Which is why I was a little taken aback to see this blog post from OpenAI CEO Sam Altman published last night. (It’s on his personal blog — one more little indicator of how Altman sees himself more as a free agent than other tech CEOs.) The title: “The Gentle Singularity.”
We are past the event horizon; the takeoff has started. Humanity is close to building digital superintelligence, and at least so far it’s much less weird than it seems like it should be.Robots are not yet walking the streets, nor are most of us talking to AI all day. People still die of disease, we still can’t easily go to space, and there is a lot about the universe we don’t understand.
And yet, we have recently built systems that are smarter than people in many ways, and are able to significantly amplify the output of people using them. The least-likely part of the work is behind us; the scientific insights that got us to systems like GPT-4 and o3 were hard-won, but will take us very far.
It’s an odd document — almost religious at times, a work of marketing clothed in the language of technology. But I can’t get past its most basic conceit: a “gentle singularity.”
Already we live with incredible digital intelligence, and after some initial shock, most of us are pretty used to it. Very quickly we go from being amazed that AI can generate a beautifully-written paragraph to wondering when it can generate a beautifully-written novel; or from being amazed that it can make life-saving medical diagnoses to wondering when it can develop the cures; or from being amazed it can create a small computer program to wondering when it can create an entire new company. This is how the singularity goes: wonders become routine, and then table stakes.By that standard, penicillin was the singularity. So was air conditioning, and television, and recorded music, and the modern sewer system. Once shockingly new and awe-inspiring, later normal, and eventually noticed most in its occasional absence. That ain’t the singularity — that’s a comfortably twentieth-century vision of geometric progression, one that ends with human beings still in charge.
I don’t say that as an AI doomer — I use it every single day, consider it a revolution in human capacity at least on the same scale as the internet, and still view its rapidly advancing powers with more excitement than fear. But let’s not rebrand the singularity — the moment our species is supplanted as the planet’s preeminent beings — as going from dial-up internet to DSL.
OpenAI is in a fascinating position. It owns dominant mindshare on the topic of AI, and it is unrivaled as a consumer product, with a market dominance that looks a little like Google in search. But its underlying technology — the models themselves, the LLMs — are nowhere near as dominant. If you look at the places where users are actively making a choice between models, OpenAI is a top player but not the ’97 Bulls.
Look at this week’s usage rankings on OpenRouter, a proxy service for the APIs of nearly every LLM imaginable. In the top 20 models, OpenAI’s rank only as Nos. 3, 12, and 13. (Google’s Gemini 2.0 Flash and Anthropic’s Claude 4 Sonnet take the top two spots.) Or look at Poe, a consumer multi-LLM app; OpenAI models rank Nos. 1 and 4 in text generation, with marketshare just under 50%. But they’re only Nos. 4 and 5 in reasoning models, No. 2 in image generation, and out of the top 5 in video generation. On leaderboards like LMArena, generated by blind rankings of responses, OpenAI’s best models beat Google’s in only 1 of 6 categories.
I’ve been testing an app called ChatBetter that came out last week. Its gimmick is that if you ask it a question, it sends it to three different LLMs it thinks will be best at answering it. And if you don’t like its choices, you can pick your own three. (It offers seemingly unlimited access to seemingly every major AI model — at least until the money runs out.) Seeing three LLMs answer the same query is instructive; it makes it easier to see when one is making stuff up, for instance. But my main takeaway is that, for most except the most specialized tests, the leading models exist in a cloud of near-parity. I like Claude best, at the moment — but a lot of that is personal preference, not its technology being wildly superior to Google’s, OpenAI’s, Perplexity’s, xAI’s, or in a lot of cases, DeepSeek’s, Qwen’s, or Mistral’s.1
So a lot of OpenAI’s strongest qualities are about marketing — the fact that, for a large share of the world’s population, “ChatGPT” and “AI” are the same thing. That Altman is a celebrity CEO is a way that Aravind Srinivas, Arthur Mensch, Cristóbal Valenzuela, Clem Delangue, and Dario Amodei2 are not. And its mindshare battles are for developers too, not just customers. Hence pieces like this that argue the “gentle” singularity is already here, it’s just not evenly distributed.
Altman doing the hard sell in his personal blog to top researchers as meta whips out its checkbook to scoop up AI talent across the valley
relatively unsubtle reference here but i take it now is not the time for subtlety pic.twitter.com/nWEdEXTlHk
— rat king (@MikeIsaac) June 10, 2025
“Abundance” is the new billionaire buzzword. Altman and Musk promising the problems with their businesses just gonna disappear blog.samaltman.com/the-gentle-s…
— Rani Molla (@rani.bsky.social) June 11, 2025 at 9:09 AM
It’s Sam Altman’s job to hype up AI. blog.samaltman.com/the-gentle-s…
— Frederic Lardinois (@fredericl.bsky.social) June 10, 2025 at 6:01 PM
“May we scale smoothly, exponentially and uneventfully through superintelligence.” A new quasi-religious blogpost from Sam Altman in which he depicts a brave new world whilst once again minimising the energy/water impacts. No resources, no world, Sam! blog.samaltman.com/the-gentle-s…
— James Martin (@jamesmart1n.bsky.social) June 11, 2025 at 12:37 AM
blog.samaltman.com/the-gentle-s…
Altman’s latest missive: masterclass in unqualifiable hype, naval gazing and jam tomorrow interesting (if true) quote on usage: a fractional teaspoon of water used per query
— Susi O’Neill (@susioneill.bsky.social) June 11, 2025 at 5:18 AM
From Sam Altman’s blog, posted yesterday
blog.samaltman.com/the-gentle-s…
— Ulrike Franke (@rikefranke.bsky.social) June 11, 2025 at 5:08 AM
The same could be said of a $2 calculator, with an equivalent meaning.
— Bob MacGregor (@bobmacgregor.bsky.social) June 11, 2025 at 7:52 AM
Sam Altman wrote a blog post to tell us that he doesn’t understand human creativity, technology, science, politics, history, ethics, or even his own AI. He does, however, have some grasp of marketing.
blog.samaltman.com/the-gentle-s…— Philip Ball (@philipcball.bsky.social) June 11, 2025 at 2:32 AM
The Ultimate Goalpost Shift
Old Days: AI was supposed to be as reliable as a calculator, and as dominant over humans as Deep Thought.
New Days: Uneducated nonexperts make lots of mistakes. AI makes mistakes. Singularity achieved!
— Gary Marcus (@GaryMarcus) June 10, 2025
RealBlink CEO Sam Ctrlman claims teleportation industry close to building massive trebuchet capable of reaching moon
blog.samaltman.com/the-gentle-s…
— Stuart Houghton (@stuarthoughton.bsky.social) June 11, 2025 at 8:04 AM
Translation of Sam Altman’s “The Gentle Singularity” from self-delusion to normal language:
“I am high as a kite”
blog.samaltman.com/the-gentle-s…
— Kevin Marks (@kevinmarks.com) June 10, 2025 at 6:53 PM
AI execs are turning their own brains to mush. We’re not colonising anywhere in the 2030s! blog.samaltman.com/the-gentle-s…
— Jacob Aron (@jjaron.bsky.social) June 10, 2025 at 5:53 PM
Read this with a slowly dawning realisation that he actually believes his own hype! Ah yes, technological advances bringing endless leisure time and riches for all – where have we heard that before?
— John Self (@john-self.bsky.social) June 11, 2025 at 2:38 AM
If someone really thought that overall intelligence was a rate limiting step on progress—hoo, all those words have exceeded load-bearing capacity—a way to “fix” that “problem” would be to make sure every child had enough to eat and a safe place to go every day.
blog.samaltman.com/the-gentle-s…
— Adam Rogers (@jetjocko.bsky.social) June 10, 2025 at 6:57 PM
wrote a new post, the gentle singularity.
realized it may be the last one like this i write with no AI help at all.
(proud to have written “From a relativistic perspective, the singularity happens bit by bit, and the merge happens slowly” the old-fashioned way)
— Sam Altman (@sama) June 10, 2025
Image via ChatGPT, of course.
- It’s also about the tooling mechanisms that model makers make available, like Claude Code or Desktop Commander MCP.
- Not for lack of trying!