The Janitor of the Cosmos Has Some Notes

Stanislaw Lem spent decades complaining that science fiction was unworthy of the universe it claimed to inhabit, and in *Microworlds* he assembled the receipts. Published in 1984 as a collection of critical essays spanning roughly two decades of thought, the book reads less like literary criticism and more like a series of dispatches from a man who wandered into a cathedral and found it full of people selling keychains. Lem's central argument is ruthlessly simple: science fiction, as a genre, had the rare privilege of engaging with the actual strangeness of the cosmos and the genuine destabilizations of technology, and it squandered this privilege in favor of recycled adventure plots, market-driven mediocrity, and a "humanized" universe scrubbed clean of anything truly alien. Forty-two years later, this diagnosis has not aged. It has metastasized. The genre Lem critiqued has since become the dominant aesthetic vocabulary of global entertainment — Marvel films, streaming dystopias, AI thrillers — and the ratio of keychains to cathedrals has only worsened. His complaint that science fiction domesticates the cosmos now applies not just to paperback novels but to the entire cultural imagination. When he wrote that SF's reliance on "convenient tropes like time travel and faster-than-light travel" had led to "a loss of the cosmos' true complexity and mystery," he was describing a tendency. Now he is describing an industry.

What Lem got right with unnerving precision was the structural problem: that market forces would continue to select against genuine intellectual ambition in speculative fiction, and that the absence of serious theoretical criticism would leave the genre without immune defenses against its own worst impulses. He predicted, in effect, the content farm — not in its specific digital form, but in its logic. His observation that science fiction existed in "continual oscillation between the Upper and the Lower Realms of literature" perfectly anticipates the present condition in which prestige television adapts Philip K. Dick while algorithmic recommendation engines flood readers with self-published LitRPG. His insistence that the authority of a statement is not proportional to the circulation of its medium now reads as a prophecy about the attention economy, where virality and significance have almost fully decoupled. He could not have foreseen the specific mechanisms — social media, large language models generating text at industrial scale, the collapse of traditional publishing gatekeepers — but he understood the underlying dynamic with the clarity of someone who had watched Soviet and American cultural systems produce different flavors of the same intellectual poverty.

His blind spots are instructive. Lem's framework is almost entirely Western and male, with the literary conversation running from Wells through Stapledon to Dick and Borges. The absence of any sustained engagement with writers outside this axis — no Octavia Butler, no Ursula K. Le Guin beyond passing mention, no consideration of non-European speculative traditions — reflects both the limitations of his access (a Polish writer behind the Iron Curtain had genuine logistical barriers) and the limitations of his imagination about who might be doing the serious work he demanded. His conviction that "truly scientific fantasy" must adhere to rigorous internal coherence and scientific methodology is a powerful standard, but it also excludes vast territories of speculative fiction that have since proven their worth precisely by abandoning that standard in favor of other forms of rigor — sociological, phenomenological, mythic. Lem wanted science fiction to be a laboratory. Some of the genre's most important work since 1984 has functioned more as a clinic, or a séance. He also could not anticipate that the cosmological strangeness he cherished — dark energy confirmed in 1998, gravitational waves finally detected in 2015, the Event Horizon Telescope imaging a black hole in 2019 — would arrive and be greeted by the culture with approximately forty-eight hours of awe before the scroll continued. The problem was never just that science fiction refused to engage with real cosmology. The problem, it turns out, is that the culture itself has developed antibodies against wonder.

The essays on Borges and Todorov remain the sharpest pieces in the collection, and they hit differently now because structuralism's corpse has been so thoroughly picked over that Lem's critique of it feels less polemical than archaeological. His reading of Borges as a writer who "transforms established cultural concepts using their own logical frameworks" is a description that now applies, with eerie aptitude, to the large language model — a system that recombines the entire library using the library's own statistical regularities, producing text that is "more related to the crafts than to creative art." Lem's distinction between craftsmen who reproduce and revolutionaries who invent is the distinction that generative AI has made existential. His insistence that "only a few can amaze, educate, and move" is no longer a critical standard. It is a survival criterion. Meanwhile, his analysis of Philip K. Dick as the writer who mastered the "transformation of trash" — using the genre's own debased materials to produce genuine ambiguity — has been vindicated so thoroughly by Dick's posthumous cultural dominance that it barely needs restating. Lem saw it first, and said so, and was right.

*Microworlds* sits in the corpus as the work of a man who wanted science fiction to earn the universe it had been given, and who documented, with exasperated precision, its failure to do so. He took from Wells and Stapledon the conviction that speculative fiction could be a cognitive instrument, and he gave to successors — Gwyneth Jones, Kim Stanley Robinson, Ted Chiang — the permission to demand more. The book's deepest contribution is not any single argument but its insistence that criticism of speculative fiction should itself be speculative: rigorous, imaginative, willing to follow implications past the point of comfort. Now that machines can generate plausible science fiction at the speed of electricity, and the market dynamics Lem described have accelerated beyond anything he could have modeled, the book raises a question it could not have raised in 1984: If the genre's failure was always a failure of seriousness, what happens when seriousness itself becomes computationally simulable — when the keychains can be made to look, sentence by sentence, exactly like cathedrals?