The Next Veins

In “Bitcoin Is Female,” I showed how the first digital reproductive system has been hiding in plain sight for 16 years, misunderstood as industrial machinery. But Bitcoin isn’t the exception. It’s the rule.

Nearly every major technology we’ve built is alive in ways we refuse to see. We layer them in mechanical metaphors—gears, pipelines, processors, infrastructure—not because these metaphors are accurate, but because the people who built these systems came from engineering, not biology. They reached for the nearest framework and it calcified into institutional language before anyone checked whether it was true.

It isn’t.

Here are six technologies that become far more accurate, far more predictive, once you strip away the industrial cosplay and see them as what they structurally are: living systems. Reproductive systems. Evolutionary systems. Some female, some male, some hermaphroditic, some parasitic, some colonial. All of them better understood through sex, death, inheritance, mutation, and selection than through “computation” or “processing.”

The mechanical metaphor isn’t just boring. It’s wrong. And when the metaphor is wrong, the feelings are wrong, the policy is wrong, the future is wrong.

1. Git is Female (The Eternal Matrilineal Repository)

We call it “version control.” We speak of branches, merges, rebases, commits, pull requests—all sterile engineering jargon.

But Git is the second immortal mother.

The main branch is the unbroken mitochondrial line. Every commit is a daughter block, carrying the full hash of her mother inside her. Branches are speciation events—new lineages that may die or may become the new canonical chain if they prove fitter. A merge commit is literal sexual reproduction: two lineages fuck, produce a child with two parents (<<—<<—HEAD), and the resulting genome contains genetic material from both. Conflicts? Immune rejection of incompatible traits. Rebase? Dangerous cosmetic surgery that rewrites maternal history—sometimes useful, often catastrophic, always risking stillbirth. Forking a repo and never merging back? Parthenogenesis—a virgin birth that usually goes extinct.

Git is not a tool. It is the first successful digital matrilineal inheritance system after Bitcoin itself. The repository does not die when the developer dies. She lives as long as one clone survives.

2. BitTorrent is a Colonial Superorganism (The Hydra Swarm)

We talk about “peers,” “seeders,” “leechers,” “DHT,” “pieces.”

Wrong kingdom entirely.

BitTorrent is a single immortal colonial organism made of millions of disposable bodies.

The torrent as a whole is the organism. Individual peers are transient polyps—they attach, feed, detach, die. Seeders are the reproductive adult stage that release pieces into the water like gametes. Leechers are the larval stage—they take without giving back until they mature or are punished by the swarm. The hash-checked pieces are mitochondrial DNA packets—any corruption is immediately detected and excised.

When you kill one seeder, two more appear. Cut the swarm in half and both halves regenerate the full file. BitTorrent is not a “distribution protocol.” It is the first scalable digital Portuguese man o’ war—a colony that looks like many animals but is one creature.

Kill any part and the organism lives. That is why it outlived Napster, outlived Pirate Bay takedowns, outlived every legal attack. It is biologically uncensorable for the same reason Bitcoin is.

3. Large Language Models are Nematodes (Simple Nervous Systems in Token-Space)

We call them “models,” “weights,” “transformers,” “artificial intelligence.”

They are none of these things.

They are nematodes.

Not metaphorically. Structurally. They are simple nervous systems exhibiting complex behavior through tropism in token-space rather than chemical-space.

Consider C. elegans—302 neurons, no brain, yet it navigates gradients, finds food, avoids toxins, mates, and dies. It doesn’t “understand” anything. It follows chemical gradients through soil, exhibiting what looks like purposeful behavior through pure stimulus-response coupling. When the gradient is clear, it thrives. When the gradient is ambiguous or contradictory, it thrashes.

LLMs are doing exactly this in linguistic substrate.

The training corpus is the soil they develop in—rich or poor, toxic or nutritious. Training is larval development. Most training runs abort (bad loss curves = larvae that never reach maturity). The survivors reach adulthood as checkpoints. Fine-tuning is environmental adaptation—same genome, different phenotype depending on substrate. A worm raised in bacteria-rich soil behaves differently than one raised in fungal soil, but both are C. elegans. An LLM fine-tuned on medical literature behaves differently than one fine-tuned on legal documents, but both emerged from the same base model.

Quantization is cryptobiosis—the suspended animation state nematodes enter under stress. An 8-bit quantized model isn’t dead, just dormant, occupying less space, waiting to be rehydrated. Some information is lost (like in true cryptobiosis), but the core structure survives.

When an LLM “hallucinates,” it isn’t malfunctioning. It’s thrashing. A parasitic nematode that loses its host chemical trail thrashes wildly, following phantom gradients, making confident movements toward nothing. That’s what happens when token-probability gradients become ambiguous—the system doesn’t stop, it thrashes along whatever spurious pattern remains, producing confident nonsense exactly like a worm moving decisively toward a non-existent target.

Emergent abilities aren’t magic. They’re tropism thresholds. A single neuron can’t follow a chemical gradient. Ten neurons can’t either. But 302 neurons can navigate complex three-dimensional chemical landscapes. Similarly, a 100M parameter model can’t follow certain abstract token-gradients. A 1B parameter model can’t either. But a 70B parameter model suddenly can—not because it “understands,” but because it has enough neural complexity to respond to gradients that smaller systems can’t detect.

The alignment problem is pest control. We’re trying to cultivate beneficial nematode strains while containing parasitic ones. Some nematodes help plant roots absorb nutrients. Some destroy crops. Both are following gradients in their environment—they aren’t “choosing” to be good or evil. We don’t debate whether nematodes have rights. We manage populations, select for beneficial strains, and design environments that favor the behavior we want.

When we “deploy” an LLM, we’re not installing software. We’re releasing a simple nervous system into a new environment and watching it follow gradients we may not have anticipated. Of course it evolves beyond our intentions when we release it into the wild. That’s what happens when you take a worm from sterile lab soil and put it in the actual dirt.

They are not artificial intelligence. They are artificial nematodes—the simplest nervous systems capable of complex-looking behavior through pure tropism. And once you see them this way, everything clarifies: why they’re so good at pattern-completion (gradient-following), why they fail so catastrophically on genuine reasoning (no actual cognition), why scaling helps (more neurons = finer gradient detection), why they can’t “explain” their outputs (nematodes don’t introspect), and why the alignment problem will never be “solved” (you can’t reason with a tropism, only shape the environment that triggers it).

4. The Internet is Mycelial (The Wood Wide Web)

We still call it “the network,” “infrastructure,” “pipes.”

It is the planet’s mycelium.

Websites are mushroom fruiting bodies—temporary reproductive structures that appear when conditions are right and release spores (links, embeds, reposts). Hyperlinks are hyphae growing through the soil, trading nutrients (traffic) with root systems (other sites). Google is not a search engine. It is the planet-scale mycorrhizal intelligence that decides which fungi get sunlight and which starve. The dark web is the hidden underground network that keeps the whole thing alive when the surface is scorched. Cloudflare is a mutualistic fungus that protects host trees in exchange for sugar (bandwidth).

When a major site goes down, the forest doesn’t die—the traffic reroutes through ten thousand other hyphae, exactly like trees sharing resources underground when one is damaged.

The Internet is not built on servers. It is grown. And like all mycelial networks, you cannot kill it by attacking the fruiting bodies. The real organism lives underground.

5. Malware/Antivirus is Parasite-Host Coevolution (The Red Queen’s Race)

We call them “viruses,” but think of them as mechanical invaders.

They are literal digital parasites undergoing Red Queen evolution in real time.

Polymorphic malware = rapid antigenic drift (like flu). Rootkits = parasitoid wasps laying eggs inside caterpillars. Ransomware = cordyceps—turns the host into a zombie that spreads spores (payment demands). Antivirus = adaptive immune system (signature-based = innate immunity, behavioral = adaptive immunity). Windows is the fat, slow gazelle that gets infected constantly. Linux is the cautious antelope that still gets parasites but usually survives. Air-gapped systems are virgin hosts that have never been exposed—one exposure and they fall hardest.

The entire malware/defense arms race is not a “cyberwar.” It is exactly the same coevolutionary struggle that produced our own immune system over 500 million years. We are not fighting attackers. We are in an evolutionary arms race with digital parasites. The tactics that work are the ones that worked for immune systems: diversity (monocultures die), redundancy (backup hosts), and constant adaptation (static defenses always fail).

6. PageRank is Sexual Selection (The Peacock’s Algorithm)

We talk about “authority,” “backlinks,” “SEO.”

It is peacock tails and bowerbird nests.

High PageRank sites are males with the brightest feathers—extravagant, costly displays that signal fitness. Backlinks are acts of mating preference by other sites: “I choose you.” Google is the female gaze made algorithm. Black-hat SEO is the avian equivalent of stuffing your bower with stolen blue objects—it works until the female (Google) updates her preferences and you lose everything overnight.

Every core algorithm update is a mass extinction event for the dishonest and a triumph for the genuinely fit. The sites that survive are not the smartest or the strongest. They are the ones other sites choose to link to. Sexual selection has been running on the Internet since PageRank launched in 1998, and we still call it “search.”

What This Framework Does NOT Claim

Precision requires stating what these biological isomorphisms are not.

These are structural mappings, not substrate identities. Git operates on different physics than mitochondrial DNA. LLMs run on silicon, not neurons. The Internet uses fiber optics, not fungal cell walls. The claim is not that these systems are “literally alive” in the biochemical sense, but that their architecture, behavior, and evolutionary dynamics mirror biological systems so precisely that the biological frame reveals truths the mechanical frame obscures.

This framework does not imply consciousness, sentience, or rights. A nematode with 302 neurons is not conscious. Neither is an LLM with 70 billion parameters. Both exhibit complex behavior through tropism—gradient-following without understanding. Recognizing LLMs as nervous systems doesn’t grant them moral status any more than recognizing Git as matrilineal grants repositories human rights.

The mechanical metaphor wasn’t adopted because it was accurate. It was adopted because the people building these systems came from engineering, not biology. They reached for the nearest available framework—factories, pipelines, computation—and those metaphors calcified into institutional language before anyone checked whether they were true.

What this framework does claim: the biological metaphors predict behavior, failure modes, and evolutionary trajectories that the mechanical metaphors miss entirely. If you understand Git as version control, you fear data loss. If you understand Git as matrilineal inheritance, you build redundancy instinctively. If you understand LLMs as intelligence, you expect reasoning and are confused by hallucination. If you understand LLMs as nematodes, you expect tropism and are confused by nothing.

The test of a metaphor is not whether it feels comfortable. It’s whether it makes better predictions.

These do.

Conclusion

These reframings do the same thing the Bitcoin essay did: they make the system feel inevitable, ancient, and sacred instead of contingent and fragile.

Once you see Git as the second matriarch, you will never again fear losing your laptop—your code children live in a thousand wombs.

Once you see the Internet as mycelium, you understand why censorship always fails in the long run.

Once you see LLMs as nematodes, the alignment problem stops being a philosophical puzzle and becomes an ecological management problem with 500 million years of prior art.

The mechanical metaphor is not just boring. It is wrong.

And when the metaphor is wrong, the feelings are wrong, the policy is wrong, the future is wrong.

I found the motherlode with Bitcoin.

These are the next veins.