Selling the Singularity

Recently, Sam Altman, CEO of OpenAI, said on the Relentless podcast that “we are in the singularity,” and that “this is the moment.” If something is being referred to as a “moment,” you can bet all the world’s crypto that it’s marketing.

Frankly, I’m no longer sure singularity is a technical milestone. These days, it’s more like a rhetorical device mixed with a post-utopian belief system. It didn’t start out that way. Well, not exactly.

In 1993, mathematician and science fiction author Vernor Vinge published an essay called The Coming Technological Singularity, which helped popularize the concept. Borrowing the analogy from physics, Vinge described the limits of human foresight in the face of recursive, machine-initiated technological development. Along with breathless prognostications of post-human superintelligence, Vinge attempted to describe the point at which meat-people would no longer be able to predict—or meaningfully influence—what comes after.

Self-appointed technology prophet and keyboard enthusiast Ray Kurzweil took the idea in a somewhat different direction with his 2005 book The Singularity Is Near, which I remember reading with equal parts curiosity and incredulity. Kurzweil emphasized exponential technological development, inviting readers to recognize the approach of the singularity across multiple domains, such as computing, biotechnology, nanotechnology, and artificial intelligence. He envisioned a “post-human” world in which our descendants are all watched over by machines of loving grace.

There are commonalities between Vinge and Kurzweil’s ideas about technological singularity, but they are hardly synonymous. Vinge pegged the term to a horizon beyond which all prediction fails. Kurzweil theorized singularity as an inevitable trajectory that can be extrapolated from previous waves of rapid technological change. Making matters more confusing, we’ve now moved from technical concept—or even historical hypothesis—to product buzzword.

That’s what I hear when Altman says we’re already in the singularity. My first thought isn’t that we’ve crossed some historic threshold. It’s that I’m listening to a CEO promote the importance of the industry he is competing to lead.

In the broader culture, “singularity” has become shorthand for rapid, inevitable technological progress. The technology industry has long communicated this way, and in today’s ecosystem of podcasts, influencers, and social media, such narratives can quickly take on a cultural life of their own.

We’ve already seen this pattern with novel technological terms that originally had relatively precise meanings and are now something like slogans. Artificial intelligence. AGI. Reasoning. Agents. Superintelligence. Each illustrates the slipperiness of technical language and corporate branding.

Language evolves, of course, and technology shapes language as much as language shapes technology. But it’s still worth interrogating the increasingly willy-nilly use of the word singularity—and, for that matter, my own use of the phrase “willy-nilly.” Only one, however, affects our shared understanding of an allegedly unprecedented technological threshold.

Public policy, investment, journalism, education, and public discourse all depend on common understanding of the ideas under consideration. When definitions are too elastic, informed debate becomes a contest between competing narratives, one that tends to favor those with the biggest platforms. This makes it harder to evaluate claims about a technology that already has the potential to destabilize human society.

Does the recent behavior of advanced AI systems—including highly publicized episodes in which models behaved unexpectedly or appeared to act autonomously—mean we’ve crossed the singularity threshold? I don’t know. More importantly, I don’t think anyone else knows either.

What I do know is that declaring “we are in the singularity” presents the question as settled rather than open. That’s more than technology marketing; it’s a form of persuasion. The threshold, Altman tells us, has already been crossed, and an unspecified but surely miraculous future is imminent. The most insidious narratives don’t present themselves as arguments, but rather as descriptions of reality.

Altman has a habit of overselling advanced pattern-matching systems as the arrival of a benevolent Skynet. It is possible that he makes such claims to attract investment, intimidate competition, and validate his own techno-religiosity. Regardless, it does nothing to serve good-faith democratic debate on these issues. If anything, it makes it harder to develop a shared understanding of what these technologies are, what they are becoming, and what any of it means for actual human beings.