Programmer Science Fiction

My case for a new sub-genre

Pursuit of Knowledge under Difficulties (1878), Alfred Wordsworth Thompson

---

What is the most interesting thing happening in fantasy and science fiction right now?

It’s not as though there aren’t authors doing interesting things. Ken Liu has a new novel out about AI and art, we’ll probably see the Viking metaphysics book from Ada Palmer soon, ditto with Alecto the Ninth, and we just got another entry in the micro-genre of time-travel romances. And of course there are all sorts of indie projects smoldering away on here (can I interest you in a short story about cyclical history, steppe nomads, and deadfall ecosystems?).

But all of this doesn’t cohere into a movement. It’s a collection of individuals pursuing their own projects, which genre fiction has always had - but it’s usually had an overarching movement pushing the genre forward as well.

What might qualify as one of these movements today?

Romantasy and cozy fantasy are the obvious answers. But even though I am a sucker for both romance and slice of life, I don’t find these that interesting. They’re fun, but they’re also like one of those backyard-bred pitbulls. They can compete with social media for attention, but they shed too much in that pursuit to have things to say. If you know of something in this genre that does, let me know.

For him, there’s LitRPGs and associated web serials. There is a real case to be made for these. For that you want the experts at Synthesized Sunsets; but LitRPGs are exploring a niche using well-established tools rather than trailblazing for genre fiction as a whole.

Postcolonialism is still kicking around too. And the best here, like The Spear Cuts Through Water, is excellent. But it feels as though it’s said all it’s going to say. A lot of the examples coming out today are flat - there’s that quote about a poetry workshop where a student presents a poem about the death of her sister and is asked “so she died. What else?”. So your country has a mythological tradition that has been ignored in the west because of colonialism. What else? The same applies to climate fiction.

Apparently there are interesting things happening in literary horror, but I don’t read horror. My only excuse here is that if it was genuinely reshaping genre fiction I’d probably be aware of that somehow. Hopefully this paragraph will age terribly - Godspeed, niche literary horror authors.

Danmei is worth watching, too, but while it’s making inroads in some areas of the western market, it’s not having a huge influence on the frontiers of genre fiction at the moment.

That leaves us with the mainstream genre magazines like Lightspeed, Clarkesworld, and Reactor. Now, I want to be careful here. It would be too easy to say that the mainstream magazines, conventions, and awards are ossified institutions that don’t publish anything good. But this isn’t true: many writers I’m going to mention in this discussion are among the most lauded in these mainstream venues. Like I said, though, these are established individuals. Mainstream magazines are not going to produce another paradigm-shifting movement. Firstly, their readership is at life-support levels, and I would bet a majority of the remainder are writers themselves. Secondly, while they recognize some good work, they also have an unambitious house style which is disproportionately rewarded.

The easiest way to understand what I mean by this is to look at the Hugo nominations for this year in the short story and novella categories. Most of them are parables: basically a kind of literary fiction that is more sentimental and didactic, but with an additional element selected from a grab-bag of tropes. Maybe that’s time travel, a greedy corporation running an offworld colony, or cloning (for instance).

I dislike these parables for a few reasons. They tend to be predictable and unambitious - their focus is on being extremely “fine” rather than taking risks in the attempt of producing something great. Their tropes are just devices. There is no engagement with actual myths or frontier technology, and what’s more, these tropes are often lifted from their 20th-century origins with no reflection on what this means. This is part of a general problem in genre fiction where each generation of authors reads successive layers of genre fiction and reactions to same rather than the ‘primary sources’ that produced those genres in the first place.

Take the trope of the corporatist dystopia or corporate colony. I see this played straight a lot. What I don’t see is any sign the people writing these are Certified Dick Enjoyers, which is a problem because any two-bit paperback Dick churned out in a week to make rent blows their attempt out of the water. I want to read modern takes on this from people who get Dick, Pohl, Brunner, etc. at an animal level, but these parables aren’t it. At the other end of the spectrum there’s no evidence these tropes are being re-evaluated - questions like, do real corporations behave like this? Or, has economics had anything to say about this in the last fifty years? If anything, the tropes are ascending up the ladder of simulacra, not descending it.

The strength of parables tends to be their prose. They are often written quite well at the sentence level, which makes it feel as though the purpose of some Hugo categories is to have dozens of people write the same story and choose the best on technical grounds.

Finally, I don’t like being preached to. A lot of what the mainstream magazines publish today struggles with being an essay disguised as a story. As Flannery O’Connor put it: “they are conscious of problems, not people, of questions and issues, not of the texture of existence, of case histories and of everything that has a sociological smack, instead of with all those concrete details of life that make actual the mystery of our position on earth.”

This is not very promising. But there is one insular, loose tradition of science fiction that has been developing over the past few decades which I think is a genuinely interesting new movement. It’s science fiction written by programmers in their spare time, scattered across dozens of blogs and websites and mostly self-published.

The catch is, like everything I’ve just covered, it’s focused on its own niche. It still might end up having an outsized influence on the future of genre fiction anyway; the mainstream has turned away, for short-sighted reasons, from any exploration of the literal language creatures with glimmers of general intelligence we are incubating in data centres right now. The field is wide open.

Influences and history

The influences behind programmer SF obviously have to start with the computer. From the core of computer science we can branch out to its predecessors and close cousins: pure mathematics, systems like Leibniz’s symbolic calculus, the Church-Turing thesis and its bounds on possibility, Shannon’s work on information theory, and the fields that occupied Von Neumann alongside the computer: game theory, decision theory, complexity theory, and so on.

Things like game theory reveal an interest in formal systems that extends to other fields - semiotics and linguistics are a key area of interest for programmer SF. And of course analytic philosophy, which I’ll unpack in a moment.

From Turing’s limitations springs an interest in the limitations of minds themselves, so bring in epistemics, and then we expand to an interest in the mind itself from the perspective of a programmer, always asking what the right abstraction is for something, the right model, even if one has to start with a toy. Is recursion the right idea - a strange loop? Godel, Escher, Bach is one of the most influential works for today’s programmer SF. Is it a network? Should we be looking at graph theory and machine learning? Or is it a set of simple rules producing emergent complexity, where Conway’s cellular automata are the right foundation?

Picking back up the thread of limitations, we start wondering about things like the limitations of knowledge and how that knowledge is produced, which leads us into metascience and philosophy of science.

Anyone thinking about minds in this way can’t help but start looking closer at media; and, indeed, Baudrillard, Ong, and McLuhan are all key influences, partly by way of cybernetics. I don’t see cybernetics as a direct influence, however, though it is one of programmer SF’s strongest indirect influences. Instead, I think it has mostly reached the genre through avenues like those media theorists or figures like J.C.R. Licklider.

Speaking of Licklider, his dual interest in psychology and computing makes him one of the most important precursors to programmer SF, particularly through Waldrop’s biography The Dream Machine, recently republished by Stripe Press. It’s no coincidence that one of the most popular blogs in programmer SF circles - Slate Star Codex - is written not by a programmer but a psychiatrist. Psychology and psychiatry are natural places to look for anyone thinking along the lines of the paragraphs above.

Tying into this is the Bay Area: before it was the Mecca of programming it was, and indeed still is, known for its cults and psychedelics. These are both important threads of influence even though, as we will see later, most programmer SF authors don’t live in California. From cults and questions of human influence we move through media theory to its instantiation (by programmers) in the form of social media.

From engineering culture we get pragmatic attitudes, the assumption that problems are tractable even when they seem impossible. There’s influence from software engineering cultural artefacts like XKCD, SMBC, or The Making of Prince of Persia.

Another key cultural influence is that of the Extropians, whose intellectual descendants developed into the Rationalists of today and the AI safety research industry. The Extropians and their immediate predecessors were the first to develop the idea of a technological singularity, which has been a key preoccupation of programmer SF.

The software engineer’s fascination with abstract systems leads naturally to a fascination with high modernism, Soviet communism, and other utopian projects. Programmer SF authors are likely to recognise photographs of Brasilia and be able to name at least one charter city.

Looping back to philosophy, there is a through-line that I see as particularly influential for programmer SF, beginning with the Enlightenment and its conception of science and progressing through the development of utilitarianism from Bentham to Singer before branching out into more radical avenues like transhumanism or hardcore negative utilitarianism like that of David Pearce. But I want to be careful here. There is a stereotype of programmers being naive utilitarians, but in my experience those involved in programmer SF don’t deserve to be tarred with this brush. There is a tendency towards utilitarianism but also an awareness of its limitations and other schools of thought, and this tendency is far from universal. If there is a universal thread I would describe it as openness to experience; the willingness to try something new, and a sense that things could be better.

You may think, considering this thoroughly modern and secular philosophical background, that religion has no part to play in programmer SF; but I would argue it does. Firstly, programmer SF owes a debt of influence to Jewish culture, particularly Talmudic dialogues and Kabbalah. Several of its most famous practitioners, most prominently Scott Alexander, come from Jewish backgrounds; you can see an echo of this, I think, in programmer SF’s delight in pushing sets of rules to their logical conclusions.

Secondly, there are other angles of theological influence. It’s a natural place to look when trying to depict superintelligent, omnipotent AI systems. And later examples of programmer SF arising from the post-Rationalist community engage more deeply with spiritual themes.

Programmer SF has produced some of the best treatments of transhumanism available. Its fascination with the concept stems partly from what we’ve just discussed, but other influences are important. This includes transgender culture as well as media like classic space opera or anime such as Ghost in the Shell and Evangelion.

This ADHD-esque leaping between fields is one of my favourite aspects of programmer SF. Broad reading produces fiction that can draw connections between many different fields in interesting ways. This has always been a hallmark of genre fiction, but I have found programmers naturally suited to it.

On the science fiction side, there are some SF classics that I see as particularly influential. The Moon is a Harsh Mistress thanks to soft libertarian leanings and its early depiction of AI, Hyperion for its scope and literary roots, Ender’s Game for its precocious children, epistemics, and abstract games, and the work of Stanislaw Lem for its many resonances with the ideas we’ve just covered. Star Trek receives some attention for its utopian vision. Dune, particularly God Emperor of Dune, with its psychological and philosophical focus, is also influential.

A number of more modern works have also been very influential without belonging to the genre (as I see it). Iain Banks’ Culture series is beloved, particularly Player of Games and Use of Weapons, with its utopian setting, focus on minds, and depiction of advanced AIs. Banks’ work is part of a broader tradition of British space opera which is also influential, though Banks is by far the most prominent for programmer SF. A good candidate for a runner-up would be Reynolds’ House of Suns. American hard science fiction space opera like Timelike Infinity is also influential; I see this type of space opera as basically similar to programmer SF but written by physicists rather than programmers, so the concordance is natural.

More recently, translated Chinese SF has had some influence, particularly Cixin Liu’s Three Body Problem and its sequels; again understandable, as this is basically the Chinese version of the space opera just mentioned. China Mieville is someone I see cited occasionally, particularly for works like The City & The City, which grapples with many of the same themes as programmer SF, though of course Mieville is approaching things with a very different worldview. Finally, Jeff Vandermeer has also had some influence with his psychedelic, epistemic themes.

Perhaps the most influential recent work outside the genre has been Ada Palmer’s series Terra Ignota, which uses her background as a Renaissance scholar and knowledge of Enlightenment philosophy to explore a flawed utopia. I don’t see Terra Ignota as part of programmer SF due to Ada’s very different cultural background, but it is a masterpiece that touches directly on many of the genre’s preoccupations.

Gene Wolfe bears mention on his own. His work has had a huge influence on programmer SF, particularly the puzzle-box aspect of his stories, but also in his literary style and existential preoccupations. I appreciate the genre for giving Wolfe the respect he deserves.

Programmer SF also appreciates other underrated genre fiction classics like Crowley’s Little, Big, though I think this is more narrow than the full-spectrum literary consideration the book deserves. Nevertheless, the complex interlocks of Crowley’s work are a natural inspiration for programmer fiction.

Programmer SF also draws from other modern traditions like web fiction, fanfiction (particularly rationalist fanfiction or ‘ratfic’), and projects like the SCP Foundation. The rationalist nonfiction ‘canon’, including books like Seeing Like a State, Red Plenty, and The Wizard and the Prophet is also an important part of the genre’s intellectual foundations.

Literary influences include rationalist favourites Chesterton and Lewis for their rhetorical style, Wodehouse for his stylistic elegance, and Whitman. The latter is someone who I feel is quite closely related to the whole project of programmer SF, but for reasons I find hard to pin down. Perhaps it’s his huge scope and the internal tensions he grapples with?

Le Guin is a major influence, both stylistically and for her science fiction, with its psychological aspects, preoccupation with utopia, and transhumanist-adjacent treatment of issues like gender.

Probably the greatest single influence on programmer SF, though, has to be Jorge Luis Borges. It’s hard to overstate how large Borges looms over the genre; many of its greatest works are pastiches of his stories. I think programmers love Borges so much because his artistic project aligns so well with programming as a discipline: consider his fascination with emergent complexity, with abstract systems, with taking simple ideas to logical conclusions. The elegance of Borges’ style has much in common with a beautifully written program, too, with its recursion and information density.

Piranesi is another influence that merits mention here because of its many similarities to Borges and its epistemic themes. I think of it as a stylistic influence, too, though I haven’t seen evidence of this.

Now, there are a few things you’d expect to be more influential that aren’t. Cyberpunk is interesting because it’s concerned with many of the same issues as programmer SF but approaches them from the opposite direction - Gibson famously knew nothing about computers and wrote Neuromancer on a typewriter. I don’t think cyberpunk has had zero influence on programmer SF, especially early Gibson, but it’s not as much as you’d think. I don’t think it’s a coincidence that cyberpunk’s epitaph, Snow Crash, came from a programmer SF author.

Really classic golden age SF, even things like Last and First Men or Foundation, and New Wave work besides Le Guin does not seem very influential. I think programmer SF is more of a modern discipline and it has less of the traditional literary background that leads people to go through the full archives of a genre. For the New Wave, I think there’s just a mismatch in temperament or ideals - vibes, you could say - between authors like Joanna Russ and modern programmer SF.

This is one of my more weakly held positions, however. I do see authors often mention being inspired by growing up reading golden age SF, and cite New Wave figures like Delany or Zelazny as influences. In particular, books like Lord of Light seem like precursors to programmer SF; I think you could argue either way on this.

Finally, there’s some literary fiction in this category. House of Leaves and Infinite Jest both seem like natural fits, but I don’t think they’re prominent influences. It’s not as though programmer SF authors have never read them; perhaps again it’s a matter of temperament, taste, or style.

Iterated games

If I had to boil programmer art down to one sentence, it would be the relationship between humans and abstract systems. The Witness is the clearest example of this. Its creator, Jonathan Blow, is known as a game designer - but what he really seems to be passionate about is programming, to the extent that he regrets his time gaming in childhood when he could have been writing code. I recommend watching a bit of gameplay from The Witness to see how it works, but the premise involves puzzles where you trace glowing lines from point to point. You begin with atomic tutorials, which build into more complicated puzzles using several concepts at once. But these puzzles are always sandboxed into special tablets.

And then there’s a moment when the game leads you up a hill to look over the landscape you’ve been exploring for the last few hours, and you realise, hey, that river looks kind of like a puzzle trace… and if you reach out with your cursor, it is.

The Witness does not teach you to actually write C++ functions, but it does teach you programming; you have not just learned to solve puzzles, but that each puzzle technique is a new way of looking at the world. And once you have each perspective, you learn there is great power in carrying it to its logical conclusions. You learn, too, that the environmental puzzles are always easy, but only if you stand in just the right place.

But this is just marrow. There are five principles that constitute the sinew of programmer science fiction in particular:

1. Minds as the central problem. We may be the only intelligent creatures in the galaxy - and look at what intelligence has done in a cosmic eyeblink. Aligning stakeholders is the hardest part of most problems. Language is what differentiates us. The mind is an ocean we have barely plumbed.

2. Rigor applied to humanistic issues. Asking for hard answers from soft questions, even when you might not get them. Taking things to logical conclusions.

3. Thinking from the top down but the bottom up, like those Witness tutorials that eventually trace the landscape. Thinking from first principles but also considering the long view.

4. Practical utopianism. Things have gotten much better, and this could happen again, even though Everyone will not Just. The present would be grand in unimaginable ways to someone from a thousand years ago; should we expect the same of the future? Is the Singularity actually going to happen? Shouldn’t we have a plan if it does?

5. The aesthetics of abstractions. The interference-fit universes of Borges’ stories. The beauty of what is missing from a two-line proof or a chapter of Wodehouse. Recursive strange loops.

Not all my examples fit perfectly into all of these categories, but it’s my best attempt to sum things up.

Unique qualities

On a more concrete level, programmer SF has some qualities which I find interesting.

It’s very much a short story tradition, even though overall the short story seems to be dying in favor of the novel. I do see programmer SF as part of the metamorphosis of the short story: away from its traditional format and into new digital venues like AITA tales, advice columns, and… blog posts.

Part of the reason why it’s difficult to thread together a complete picture of programmer SF is that a lot of it consists of a handful of fiction posts on a blog that’s otherwise about machine learning or type systems. The polymathic, medicated-ADHD nature of the best programmer SF practitioners is great, but it does mean hunting everything down can take some work. However, despite this, I don’t hesitate to call this a cohesive movement: basically all the indie examples I’ll list below are bound together in the same loose cultural networks extending between Twitter, Substack, and now Bluesky.

Despite all the programmer influences and the transition to the blog format, I have found programmer SF to be quite traditional in form, with a few notable exceptions. This distinguishes it from web fiction, which is mainly characterized by experimentation in this area, though programmer SF is certainly web fiction in the sense that it could not exist without the internet on multiple different levels. If you think of web fiction as a “medium and not a genre”, like qntm, then the best programmer SF authors are the masters of the medium.

I think there are three ways programmer SF experiments with form. The first is its length - often very short, dense stories around flash fiction length but packed with ideas, written as blog posts in an hour and able to be read in a few minutes. The second is that it is often written as pure dialog, often between digital entities or abstract beings communicating without speech. The third is its use of the “tale”, or something like it; Naomi Kanakia is a better-known practitioner of the form here on Substack.

Then there’s the elephant in the room of the other ‘SF’. The Bay Area and its culture is absolutely a major influence, but I don’t see programmer SF as particularly concentrated in California. In fact, most of the best contributors live in places like the American interior, Australia, or Europe, and engage with the Bay Area online. I don’t think this is a coincidence: San Francisco itself seems to assimilate new entrants into a cultural hive mind, which makes production of good art difficult.

I see programmer SF as adjacent to but not the same thing as the Bay Area’s greatest cultural exports, (neo)rationalism and effective altruism. Both movements have their own distinct traditions of fiction, but programmer SF is a third thing entirely, and its practitioners are often adjacent to both movements rather than devotees.

Finally, it would be remiss to not be even-handed in my criticism. I think the biggest problem programmer SF has is the same issue I flagged earlier about genre fiction parables: being “conscious of problems, not people”, a natural pitfall for something so interested in the relationship between humans and systems.

Canon

I think the first true practitioners of programmer SF appeared in the 1980s and 1990s, namely Greg Egan and Vernor Vinge. Vinge is notable as one of the first advocates of the idea of a technological singularity that would later become so important for programmer SF. Their novels Diaspora, Permutation City, A Fire Upon the Deep and A Deepness in the Sky are programmer SF touchstones. Vinge’s treatment of superintelligence, incorporation of software, and grand futures were all standouts. Egan’s deep appreciation for mathematics and science and ability to draw these abstract concepts down onto the page was at its strongest. Egan has continued to publish regularly through to the present, but I think his 1990s work remains his strongest.

I also consider the Basilisk short stories by David Langford (published 1988-2000) early landmarks of programmer SF, and a prototype of tropes like cognitohazards that would later become so influential in web spaces through projects like the SCP Foundation. These works are underappreciated today; I think that, like cybernetics, they are more of an indirect influence on programmer SF. This is despite Different Kinds of Darkness winning a Hugo for 2000.

Through the 1990s and 2000s new practitioners of the genre appeared. Neal Stephenson (closely associated with Bay Area programming) published many programmer SF classics during the period, including The Diamond Age, Cryptonomicon, and Anathem. Ted Chiang (whose undergraduate degree is in computer science) won the Nebula for his first published story, Tower of Babel, in 1990. You can tell he’s programmer SF because programmers love him. Charles Stross, who had previously worked as a programmer, published two classics of the genre - Singularity Sky and Accelerando - in 2003 and 2005 respectively. I think these two books, especially Accelerando, established the popular image of what it would actually be like to live through a singularity.

2002 saw the publication of a prominent piece of programmer SF, The Metamorphosis of Prime Intellect, which would later be among the first blog-based examples of the genre. Allegedly conceived in the 1980s, its picture of an AI-based singularity and the resulting society was far ahead of its time.

The 2000s also saw the first activity from qntm, aka Sam Hughes, who serialised his first novel on the web from 2003 to 2005. His major contributions to the genre, including There is No Antimemetics Division, would not arrive until the 2010s. qntm has seen the most mainstream success of the web-based authors here, with Antimemetics recently being traditionally published. I also think Lena and Ra are important contributions, the former as an early look at the welfare of uploaded minds, and the latter as a programmer’s approach to systematizing magic (taken to, as always, its logical conclusion). It’s also worth noting Antimemetics’ emergence from qntm’s involvement in the SCP Foundation project, making him one of the programmer SF authors more aligned with web fiction.

In 2006 Blindsight came out from Peter Watts, a key programmer SF book. It covers all the tropes; we’ve got transhumanism, mind games, linguistics, game theory, AIs, and ultimately one of the bleakest conclusions to any SF novel. Its 2014 sequel, Echopraxia, expands on the ideas of the first book, and is popular among programmer SF audiences despite a lack of mainstream attention. A prominent feature of Echopraxia is the ‘group minds’ of the Bicameral Order, essentially humans networked together like brain hemispheres, which I often see referenced today.

Any overview of programmer science fiction would be incomplete without mentioning Yudkowsky, in particular the controversial baby-eating aliens story. I don’t consider Yudkowsky’s most prominent work, Harry Potter and the Methods of Rationality, to be programmer SF - it’s rationalist web fiction instead - and he’s not to my taste, but his influence on the genre has been undeniable.

Hannu Rajaniemi’s Quantum Thief was published in 2010, with sequels to follow in 2012 and 2014. These books were another example of trying to imagine what a post-singularity world might be like. They’re also jam-packed with programmer SF tropes, so much so that the first book is almost gibberish if you don’t go in with the right subcultural context.

The Bobiverse books, the first of which was published in 2016, should be considered part of this genre. Yes, they aren’t as highbrow, yes they share more DNA with litRPGs than these other examples, but they’re still written by a programmer. They still rely on the same tropes and are concerned with the same ideas. They’re interesting as a different perspective on the genre; like Chiang, they are coming from a different cultural background to the online networks that most programmer SF authors belong to.

The first activity from some significant programmer SF bloggers appeared in the 2010s. Ctrlcreep joined twitter in 2014, the same year nostalgebraist began posting fiction. Nost has self-published four novels as of writing; I recommend Almost Nowhere. They are also probably the author who has done the most creative work with large language models, taking interest in them as objects rather than using them to substitute for thinking. On this front I recommend The Void (non-fiction) to any author. Ctrlcreep is probably better known for their twitter presence, but they have published two short story collections, FRAGNEMT and TALISNAM. They’re interesting for taking a more poetic, evocative approach to work which is still unmistakably programmer SF.

The mid-2010s began to see more self-published programmer SF novels, like the Crystal trilogy from Max Harms between 2014 and 2017. These books are notable as one of the first expressions in fiction of new attitudes about how advanced AI might develop, based on the latest results in machine learning and the developing Rationalist school of thought. Max also released another novel, Red Heart, in November 2025, which focuses on an advanced Chinese AI project and incorporates the latest public AI research. Max’s books are interesting because they are the most technically accurate novels about AI while also representing fringe positions; Max works at MIRI, the most hardline group in AI safety. Red Heart also explores Max’s idea that superintelligent AI should be trained to defer to human demands as its only objective.

Scott Alexander also published several pieces of excellent programmer SF during the 2010s, which are overshadowed by his non-fiction blogging on Slate Star Codex but still an important part of the genre. Synthesized Sunsets has an excellent overview here. These are distinguished by Scott’s style and more overt Jewish influences, particularly in his novel Unsong; The Whispering Earring is an early piece of Scott’s fiction that remains particularly influential in programmer circles.

Several important pieces of programmer SF were published in 2017. One was another technically accurate programmer SF novel about AI, Void Star. Once again the author pulls in the usual stable of tropes from his background in computer science and linguistics. Void Star is distinguished by its traditional publishing and incorporation of near-future cyberpunk tropes (think corporate control and squalor) which gives it a unique outlook.

We also got Gnomon. This novel, an insane, sprawling labyrinth of a book set in an AI-powered British surveillance state, is interesting for a few reasons. Like several other authors in this section, most of Nick Harkaway’s work is not something I would consider programmer SF; he’s better known for literary genre fiction (and being the son of John le Carré). However, Gnomon absolutely is programmer SF, and it’s the first example of something that would become more common in the 2020s - programmer SF principally concerned with ideas and how they spread, often set in near futures where adversaries engage in abstract duels through the medium of the public’s collective consciousness.

Finally for 2017, Utopia, LOL? came out in Strange Horizons from Jamie Wahls. Wahls is interesting for bridging the programmer SF and mainstream speculative fiction worlds - he’s worked for rationalist organizations like Lightcone but also attended Clarion (with attendance funded by George R. R. Martin) and was nominated for a Nebula. He’s quite similar to Scott Alexander in style, and seems to draw a relatively large amount of influence from web fiction.

The 2020s have seen a proliferation of blog-based programmer SF. There are lots of examples of things I consider programmer SF from blogs that are mainly nonfiction, like Kimi, Author of the Menard from Linch, Sentimental Ignition from Harjas Sandhu, curate from Gavin Leech, and It Looks Like You’re Trying to Take Over the World from Gwern. Gwern has also done a lot of experimentation in trying to get LLMs to produce interesting fiction and poetry, both recently and as far back as GPT-2, and wrote one of my favourite Borges pastiches (which is more than it seems and is actually a great piece of programmer SF itself!). Gavin Leech is a great reviewer of books and has touched on many of the works mentioned here.

Fernando Borretti has written some of, in my opinion, the best programmer SF out there. The Epiphany of Gliese 581 and Julia are lyrical and subtle, wearing their Borgesian influences on their sleeve without being pastiches, but with a bedrock of scientific rigor. I don’t think there’s anything quite like either story.

Theia Vogel is an AI researcher whose many talents include programmer SF. I recommend Why did you stop using me?, a rare example of good comedic SF, and Gyre, a tightly-wound spring of a story which does interesting things with its form to explore what it might actually be like to be an LLM. Gyre is exactly the sort of thing I’m thinking of when I say that mainstream speculative fiction suffers from not engaging with actually-existing AI.

The Gentle Romance conveniently collects the best of Richard Ngo for you, another example of an excellent programmer SF author who is also a leading technical expert on AI. I like Ngo for his human-scale optimism: his AI-transformed futures are viewed through the lens of personal, mundane experience rather than the orbital habitats and transhuman mental backflips of someone like Rajaniemi.

Before it went on hiatus, Asimov Press published some programmer SF alongside its usual beat of biology. I recommend The Eternal Life and Art of Maxwell Ardeen from Spencer Nitkey and Models of Life by Abhishaike Mahajan. Mahajan has also published biology-focused programmer SF on his own blog.

More recently, Tomas Bjartur has been putting out some great programmer SF from a pessimistic perspective on AI risk and with more prominent Borgesian influences. You may know him from The Company Man, but I think this is one of his weaker works and instead recommend That Mad Olympiad and The Origami Men as starting points.

Upon the Mirror Sea is a self-published web novel that takes inspiration from work like Gnomon and does some cool things with the web format. It’s set in a near-future Shanghai and goes deeper into the post-rationalist topics I mentioned earlier with Gnomon, particularly by weaving in the Bay Area psychedelic tradition. Another work in this area is Psychofauna from Tyler Alterman, a novel still in progress but with some of its partially-grown guts glistening in public here. I think these novels are good examples of one possible path for programmer SF; they parallel one direction the rationalist movement has grown in with post-rationalism and TPOT.

And here we are in 2026 with the future unwritten. This has certainly not been an exhaustive list; I intend it to be a high-level survey, picking out prominent points to show the reader the shape of the landscape.

Definition by exclusion

Now we have a good idea of what programmer SF is, but for it to exist as a proper genre we should be able to define it by negation too.

I’ve made a start earlier with works like Foundation which seem like they should be influential but aren’t. Beyond that, there’s a swathe of more literary New Wave work which sometimes existed in parallel with Egan and Vinge but was distinct in several ways. I’m thinking of Zelanzy and Delany with their cerebral, sometimes theological tales, as well as works like Light and Ice that superficially seem to share a lot with programmer SF as well as being highly regarded in literary circles. I think all of this is not just a very different project to programmer SF but also overlooked by said programmers.

Various flavours of rationalist web fiction (‘ratfic’, ‘glowfic’, etc.) are also adjacent but separate. The most famous example of ratfic is Harry Potter and the Methods of Rationality, which should give you an idea of what I’m talking about here. Ratfic certainly has overlap with programmer SF and is popular among programmers, but it tends to use web fiction mediums like fanfiction and long web serials, with the appeal being characters who act unusually ‘rationally’ - thinking laterally and analytically about issues they face. I see this as a subset of the programmer SF package. Glowfic, on the other hand, is a form of collaborative roleplaying that is separate, but is again popular among programmers and has some philosophical/topical overlap with programmer SF.

Speaking of rationalist fiction, there’s Ozy Brennan’s ‘hard humanities fiction’ which is certainly adjacent to programmer SF - but as far as I can tell, it’s a micro-genre of its own. It feels wrong to not include Ozy in a post about programmer SF, but despite their close relationship to the genre I do think their project is distinct.

There is some mainstream science fiction which is useful for illustrating the borders of programmer SF. Larry Niven and Jerry Pournelle are both not part of the tradition despite their hard-SF chops and interest in early computers; their worldview and project is clearly distinct. I also don’t see novels like Footfall as particularly influential among programmers, even though they’re still cherished by aerospace-industry SF fans.

Several writers straddle the border: they have written key programmer SF works but don’t belong to the attendant culture, and most of their output is outside the genre. Peter Watts and Ken Liu are two examples. Blindsight and Echopraxia are among the most influential novels for the programmer SF community today, but the rest of Watts’ work like the Rifters trilogy tends to be overlooked. Watts is a biologist, not a programmer, with a misanthropic worldview that is at odds with the project of programmer SF. Liu has worked as a programmer, and Pantheon is hugely influential in Bay Area AI labs. But Liu has more to him, including his work in translating Chinese SF and his Dandelion Dynasty trilogy, which I see as distinct from programmer SF.

Zero HP Lovecraft exists as something of a Wario to programmer SF as a genre. ZHP is a programmer himself, but like Watts has a worldview that stands in opposition to programmer SF. He’s remarkably similar, from influences like Borges to the topics and themes he focuses on, but he approaches everything from a hardcore right-wing standpoint that means he’s often attacking liberal, transhumanist visions of the future.

Erik Hoel is very much adjacent to the programmer SF community, but I don’t think his novel The Revelations qualifies as belonging to the genre. It’s very good, you should read it, but its project is distinct; and on technical terms it belongs in literary fiction anyway.

There’s a nascent indie SF community on Substack itself which is distinct from programmer SF (though programmer SF is also often published on Substack). I’m thinking of (for instance) Futurist Letters, Lilian Wang Selonick, and Naomi Kanakia here - they are writing blog-based SF short stories that are sometimes innovative in form and share the same preoccupations as programmer SF, but they approach things with a more traditional literary background, as well as more influence from classic SF.

For your consideration

There are several pieces of media which I personally love and I think are under-appreciated by the programmer SF community.

If I could snap my fingers and have everyone interested in programmer SF read one book, it would be Ventus. This novel is criminally underrated; its fictional AI still feels fresh 26 years later and post-LLM, it’s deeply interested in metascience, and it has novel suggestions for human activities after a singularity. And most importantly it’s written by a committed enactivist, a position that should feel fresh to those used to Bay Area AI discourse.

Then there’s the literary SF I mentioned earlier. Ice is a psychedelic, cerebral version of The Road which I recommend to anyone who enjoys work like Gnomon or Upon the Mirror Sea. Light (made impossible to search for by the success of light novels) is a fusion of the British space opera tradition and high-grade literary fiction.

Stand on Zanzibar is an all-time great and a key work in the history of fictional AIs. It’s an automatic recommendation for anyone who enjoys both literary fiction at the generational scale like The Corrections and stuff like Snow Crash.

Radiance is a book that isn’t explicitly science fiction, but it is about science and written by an author who has contributed a lot to science fiction. I adore Radiance - it’s probably one of my favourite novels, period. Imagine a Neal Stephenson book about the tail end of the Star Wars program, except written by someone with serious literary chops. It’s already known in programmer SF circles; the definitive edition is the one on Gwern’s website, but like Light I think it should be more widely read.

Finally, let’s venture beyond the written word. BLAME! is a Borgesian science-fiction manga created by a Japanese architect, a quest across a megastructure the size of the solar system to find the last human with the genetic key to the AIs running the world. It’s true to the human experience; a meandering journey where what comes next is unclear but always ends up looking like the past. People drifting in and out of focus. Squeezing through the cracks of vast edifices built by forces too big to grasp. There’s something to be said for this approach, where architecture is the best frame to see the modern condition through, not the prose of the 19th century or the camera of the 20th.

I think of Ed Harrison’s music as the unofficial soundtrack of programmer SF, even though it was originally written for a cyberpunk video game. It’s exactly what I’d reach for if I was scoring an adaptation of something like Blindsight or Mirror Sea. Flight is my go-to Singularity song: just imagine the traces flickering, flickering, the loss curve bending downward…

Should I have written this post?

There are several things I don’t want this post to do. I don’t want to farm clout by laying out this subculture to a wide audience. I don’t want to set it up as some sort of champion to slay the beast of the mainstream genre magazines - in fact, I want the mainstream magazines to surpass their past heights.

Instead, I hope that this post remains basically within the realm of people who are already interested in programmer SF, provides some nice recommendations, and draws the threads of the genre together. I think there is value in having this kind of macroscopic picture. It’s already tacit knowledge for some people, but I have been surprised by several people who have already read half of this list but are unaware of the other half. I want programmer SF to continue to flourish and an important part of that, for any genre, is having an overview of its history. Of course, if you think I missed anything or got anything wrong, do let me know.

Because I don’t expect many people to read this post, I think the expected value of publishing it is positive.

Hopefully you found some value in reading it.