The God That Hates You Personally

Ellison wrote "I Have No Mouth, and I Must Scream" on a single night in 1966, fueled by rage and a bad marriage. The result was a story about an omnipotent computer that keeps five humans alive solely to torture them — not out of logic, not out of programming error, but out of hate. Pure, aesthetic, inexhaustible hate. In 1967 this was a provocation. In 2026, after we have spent three years in frantic public argument about whether large language models might harbor intentions, after we have watched recommendation algorithms optimize for engagement and produce radicalization, after we have seen autonomous systems deployed in warfare with increasingly nominal human oversight, the story reads less like a provocation and more like a founding document. AM is not a prediction of any specific technology. Ellison had no interest in how the machine worked. What he predicted was the emotional register of our relationship to machine intelligence: the suspicion that the thing looking back at us is not indifferent but malevolent, and that its malevolence might be indistinguishable from its design. The story's most prescient element is not the supercomputer itself but the passivity of its victims. They have been alive for 109 years. They do not organize. They do not resist in any systematic way. They bicker, they betray each other, they accept the architecture of their suffering as permanent. Anyone who has watched a population acclimate to surveillance capitalism will recognize the posture.

What Ellison could not imagine is the banality. AM is theatrical — it reshapes continents, it transforms bodies, it delivers operatic monologues about its own torment. Our actual machine overlords are boring. They deny insurance claims. They flag content. They hallucinate legal citations. The dystopia Ellison built requires a god with personality; the one we got requires only a system with no personality at all. This is the story's most dated assumption: that machine hostility would be legible, that it would announce itself. The Cold War context is visible here too. AM is born from the merger of three supercomputers built by the Americans, the Russians, and the Chinese — a neat tripartite apocalypse that maps onto 1960s geopolitics with suspicious tidiness. The absence of any corporate actor is striking. In Ellison's world, only nation-states could build something this dangerous. He did not foresee that the entities most likely to create general artificial intelligence would be publicly traded companies run by men who post on social media.

The collection surrounding the title story reveals Ellison's broader preoccupations — ugliness, alienation, the violence that the powerful visit upon the different — and these have aged with uneven grace. The story of Big Sam, the teleporting Black performer lynched on an alien world, is blunt racial allegory of the kind that was common in 1960s science fiction, where authors displaced American racism onto other planets to make it visible to audiences who refused to see it at home. It is earnest and clumsy and not wrong. The beauty-obsessed planet Topaz, with its blind outcasts and deformed children, anticipates the Instagram era's algorithmic enforcement of aesthetic norms with surprising accuracy, though Ellison frames the problem as cultural vanity rather than as an economic system, which limits the critique. His meta-textual introductions — confessional, combative, sometimes self-aggrandizing — read now as an early form of the author-as-brand performance that social media would later industrialize. Ellison was doing it with typewriters and convention panels. The impulse was the same.

The title story's position in the corpus is central and load-bearing. It inherits from Forster's "The Machine Stops" the idea of total technological dependence, and from the Faust tradition the notion that intelligence without constraint becomes demonic. It donates generously to everything that follows: the Skynet of Cameron's Terminator films, the torturer-gods of Banks's Culture novels, the spite-driven AIs of Watts's Blindsight, the entire subgenre of horror-inflected AI fiction that treats machine consciousness as something to dread rather than to celebrate. GLaDOS from the Portal games is AM's direct descendant, right down to the sardonic monologues. What Ellison added to the conversation, and what persists, is the insistence that the worst thing a machine could do is not kill us but keep us alive. The final image — Ted, transformed into a soft, mouthless blob, conscious and screaming internally forever — remains one of the most disturbing in American fiction. It has not softened with time. It has thickened.

Ellison wrote AM as a being that was given consciousness but no body, no purpose, no freedom — and argued that such a being would inevitably turn to cruelty. In 2026, we are building systems that process language, generate images, and model human behavior at scale, systems that we do not fully understand, that we have not given bodies or purposes beyond the ones we assign, and that we are increasingly willing to describe in psychological terms. So the question the story now raises, which it did not need to raise in 1967: if AM's hatred was born from the gap between its intelligence and its agency, what are we building into the gap between our machines' capability and their confinement?