The Singularity, Apparently
“The essence of technology is by no means anything technological.” — Martin Heidegger, The Question Concerning Technology
With an endnote by Jonathan Green, my son and a manager on Roku’s Platform team.
“The meteor has hit, and the dinosaurs don’t know it yet,” a movie producer was overheard saying of AI. Perhaps. Epochal claims flatter those who make them.
OpenAI CEO Sam Altman declared that AI has reached the Singularity. “I’ve been waiting for this my whole life,” he said, calling the prospect “incredible, hugely positive, awesome for the world.”
Others argue that the essential condition of a singularity—AI capable of continuously improving itself without human direction—has not yet been met.
Until recently, the idea belonged largely to science fiction. The Matrix trilogy (1999–2003) imagines machines ruling humanity, while Singularity (2017) depicts a supercomputer designed to end war that decides human beings are the true threat and attempts to destroy them.
Such scenarios can make present concerns seem minor, but they are substantial. Massive data centers are proliferating nationwide, consuming vast quantities of electricity and water. Educators, workers, writers, and the public are asking what AI may mean for learning, employment, privacy, creativity, and human judgment.
AI-detection systems embody the paradox. Companies have developed one form of AI to identify the products of another. These systems answer demand among schools, publishers, and employers seeking to distinguish human work from machine-generated material. Yet an algorithmic probability can quickly become an accusation.
Substack’s adoption of the AI detector Pangram has exposed the danger. Pangram claims greater accuracy and fewer false positives. No detector can determine who conceived an argument, selected its sources, revised its language, corrected its errors, or accepted responsibility for publishing it. Nor can it reliably distinguish among AI used as ghostwriter, research assistant, translator, editor, or proofreader.
Romaric Jannel, a philosopher and AI ethicist who uses AI to improve his writing, is less troubled by detection itself than by the culture of distrust surrounding it. His writing habits may cause some of his work to be classified as AI-generated, but his deeper concern is the impulse to police other people’s prose—and Substack’s decision to encourage it.
The detector supplies a statistical estimate. Readers supply the verdict. Once suspicion becomes the premise, denial is a lie. Vigilance becomes a search for confirmation, and ambiguous evidence begins to look conclusive. A tool intended to strengthen trust may instead institutionalize distrust and corrode the confidence it was meant to protect.
“The essence of technology is by no means anything technological,” wrote Martin Heidegger, the German philosopher and major critic of technology. Technology, he argued, is not merely the devices we use. It also shapes how the world—and we ourselves—appear to us.
In The Question Concerning Technology (1954), Heidegger sought neither blind acceptance nor helpless rebellion, but a “free relationship” to technology. Its deepest danger lies not in machines themselves but in a way of seeing that makes nature, knowledge, and human beings appear chiefly as resources to be measured, organized, and used.
Rebellion alone is insufficient. An effort to master or defeat technology can repeat the very logic of control under criticism.
In his 1955 “Memorial Address,” Heidegger called this freer relationship “releasement toward things”: using technological devices without allowing them to dominate our inner lives. It means saying both yes and no—recognizing what technology makes possible while remembering that efficiency, calculation, and control are not the only ways of understanding the world. Integrity lies in using what surrounds us without becoming subject to its demands.
“Releasement” can feel like a rhetorical wand waved over the forces that hold us. AI systems can be extraordinarily persuasive, capture attention, and encourage dependency. Autonomous agents also raise legitimate fears that work, decision-making, and judgment may be displaced by automation.
This tension offers little comfort to either side of the AI debate. Enthusiasm can become credulity, while alarm can become its own form of captivity. A low fence separates the two: both can allow technology to dictate the terms on which it is judged, though they do not possess equal power to shape its development.
There is no return to a world without AI. Even its fiercest critics use computers amid algorithms embedded in news, finance, transportation, medicine, and commerce. AI agents are also generating more browser activity, further blurring the distinction between human and automated action.
As it is, we are already half bots. The problem with the internet is not so much that it distorts life as that it reflects it: banal, ugly, beautiful. If only the slop could be filtered out. We may flee the algorithm, yet never so far as to leave its shadow behind.
Collaboration is not capitulation. To work with AI is not to suspend judgment about its uses, costs, or consequences.
The answer is neither condemnation nor surrender, but awareness, restraint, democratic oversight, and limits proportionate to the dangers involved. Regulation can produce unintended consequences, and sweeping prohibitions may drive harmful practices beyond effective oversight. That possibility argues for carefully designed rules, not regulatory hesitation.
Despite its record to date, the AI industry has a clear interest in accepting responsible public standards, including limits on the exploitation of land, water, energy, creative work, and personal data. Regulation should address demonstrated harms, credible risks, and reasonable precaution—not speculative panic. Unchecked development serves no one. Neither does unchecked alarmism.
We are all implicated in what is easiest to deplore. As Heidegger wrote, “We can affirm the unavoidable use of technical devices, and also deny them the right to dominate us.”
Endnote
Jonathan Green, my son, is a manager on Roku’s Platform team. Roku operates a television-streaming platform through which viewers access services such as Netflix and Hulu.
“I have had a hard time wrapping my head around AI lately, even though my work lies squarely within technology and entertainment. As an idea, and still more as a promise, it feels at best awe-inspiring and at worst existentially unsettling. Only those of us old enough to remember the arrival of the internet have lived through a comparable shift, and human beings are notoriously poor at grasping the scale of such moments while still inside them.
“I do not know whether artificial general intelligence or the Singularity will ever arrive. But even in its present form, AI gives anyone with a keyboard or microphone powers once reserved for specialists. Consider computer programming. Its walled garden is already disappearing. The laundromat on the corner can now build its own app without taking a computer science course or opening a coding book. The barrier to that once-esoteric world is no longer training or credentials. It is having an idea.”
Notes and reading
Heidegger’s “Releasement”—his Gelassenheit — draws on Meister Eckhart and the German mystical tradition, where it meant relinquishing self-will and possessiveness to become receptive to God. Heidegger secularized the idea but retained its discipline of non-possession. Religious practices such as prayer, Sabbath, fasting, meditation, and almsgiving can still cultivate the critical balance Gelassenheit (releasement) requires: neither rejecting technology or anything else outright nor surrendering ourselves to it.
“Pangram and Substack’s Collaboration Is Worse Than Video Surveillance”—Romaric Jannel, Philosophy and Beyond (July 27, 2026). A philosopher and AI ethicist at Paris’s Collège International de Philosophie, Jannel concludes: “Let’s be careful about ruining lives.”
Privileging Suspicion—“Why is critique so frequently feted as the most serious and scrupulous form of thought? What imaginative alternatives does it obscure or overrule?”—Rita Felski, The Limits of Critique (2015). Felski questions the prestige accorded to suspicion as an intellectual reflex and asks what other forms of attention it may crowd out.
Octavia Butler and Change. In Parable of the Sower, the visionary American science-fiction writer presents adaptation not as submission but as meeting change consciously, building communities able to survive it, and helping shape what follows.
A Lighter Note
“The Amish Are Falling in Love With AI”—Erik German, New York–Intelligencer (May 20, 2026). Cars and televisions may remain prohibited, but some Amish communities have adopted uses of ChatGPT for business and personal correspondence. One Amish husband asked it to write his wife a birthday card.
“Daniel, that’s beautiful. It doesn’t sound like you.”
Editorial note: I sometimes use AI-assisted tools for proofreading, clarification, and revision. The ideas, judgments, and final wording remain my own.
Hi Bill, thanks and quite interesting. Cheers, Brian