§0 Sammanfattning

This post closes the topology line. It reassembles two unpublished working notes — DRK-TOPO-002 (The X→Y Operator: Symmetry-Breaking as the Topology of Propagation) and DRK-TOPO-003 v2 (Views as Stalks, Variety as Potential, Coercion as Kinetic) — and uses them as apparatus for a single question:

By what sequence of transitions does the substrate of the universe — vacuum, fields, whatever precedes them — become creatures that invent names for everything they can imagine?

The answer offered is a ladder of eleven rungs, R0–R10. At each rung the same three-part operation recurs:

  1. Break — a symmetric state (an X) resolves into one of several degenerate alternatives (a Y).
  2. Keep — the chosen alternative is stabilised as a record in some medium, at a thermodynamic cost.
  3. Copy — the record is propagated by a restriction map $\rho_k$ at a fidelity $q_k$, and propagates only if that fidelity clears a threshold.

Two corrections to the source notes are made in the open (§2.3, §2.4), and one premise of the commissioning question is deliberately not adopted (§1): that matter and energy consist of information. That thesis is a live hypothesis, not consensus. The ladder is written so that it holds under both readings.

Epistemic ledger. Every substantive claim carries one tag:

Tag Meaning
[E] Established: textbook consensus or direct measurement
[S] Supported: leading model with strong but incomplete evidence
[H] Hypothesis: open, contested, or without decisive evidence
[D] Draken synthesis: structural claim made by this corpus, submitted for Clinch review
[M] Metaphor or paronomasia: a pointer, not a referent (per DRK-165)

§1 The ontology is an hourglass — and we do not cut it

The commissioning question assumes that information is what matter and energy are made of. There are three serious positions, and the corpus should not pretend the question is settled.

P1 — Bit from it. Information is always instantiated in a physical carrier; it has no existence apart from states of matter and fields. Landauer's principle gives this teeth: erasing one bit in an environment at temperature $T$ dissipates at least

$$ E_{\min} = k_B T \ln 2 . $$

This bound is derived from the second law [E] and has been confirmed experimentally in a colloidal-particle memory (Bérut et al. 2012) [E]. Landauer's own slogan was information is physical.

P2 — It from bit. Wheeler (1990) proposed that every physical quantity derives its meaning from yes/no answers to measurement — that the "it" arises from the "bit." This is a research programme with serious descendants (quantum-information reconstructions of quantum theory, holographic bounds) but it is not established physics [H].

P3 — Relations first. Relational and structural-realist programmes — including Smolin's causal theory of views, used in TOPO-003 — take relations between events as primary, with neither "stuff" nor "bits" as the ground [H].

P1 and P2 are the two cones of an hourglass on the ontology axis, mirror images of each other: it above, bit below, or the reverse. Cutting the hourglass here — choosing a Y — would be a totalising move made without evidence. Draken holds the X open.

Design constraint for the rest of this post [D]. Every rung is described in terms of records and copy maps. Both P1 and P2 admit records and copy maps; they disagree only about what a record ultimately is. The ladder therefore does not depend on the ontology. If any rung below turns out to require P2 specifically, that is a falsifier (§7, F4).

§2 Apparatus recovered

§2.1 From TOPO-002: the X→Y operator

TOPO-002 took three slides from a Levin–Farias conversation — the light cone of self, the substrate-independence triptych (gene network / $V_{\text{mem}}$ / tension), and the black-hole/white-hole hourglass — and read them as one operator:

An X is a time-symmetric mechanism: two cones, reversible, efficient causation. A Y is its symmetry-broken form: one funnel, a mouth and a spout, directional. Going from X to Y is choosing a direction of propagation.

The hourglass is two funnels glued at a singular throat,

$$ X = Y \cup_{\text{sing}} Y , $$

and in TOPO-002 §3 the throat is the now: the zero-width present where the global section is evaluated, the common target of $\rho_{\text{past}\to\text{present}}$ (memory) and $\rho_{\text{future}\to\text{present}}$ (anticipation). Coherence debt integrates forward through it and cannot be re-sectioned backward:

$$ K(t) = \int_0^t \big[\Psi(\tau) - \Psi_{\text{viable}}\big]\, w(\tau)\, d\tau . $$

The operator was then mapped onto three substrates: the self as a sheaf over a worldline; the sex chromosomes (XX as the symmetric pair, the Y as a degenerated homolog retaining SRY, Lyonization as forced silencing of one X); and the platform X.com as an attention funnel. The unifying claim was that life and intelligence are the same operation on different stalks: a pattern funnels itself across the now into the next light cone [D].

§2.2 From TOPO-003 v2: views, variety, and the coercion term

TOPO-003 observed that Smolin's causal theory of views — events each carrying a view of their causal past — is structurally a presheaf over a causal set: view = stalk, causal link = restriction map [D]. Smolin's principle of maximal variety (Barbour & Smolin 1992; developed in later Smolin and Cortês–Smolin work) places the distinctness of views at the centre of the dynamics [H].

After DeepSeek's Clinch review, v2 made its decisive repair — due to K. Roininen — by refusing to add a third "care" term to Smolin's two-term structure. Instead it decomposed the rate of variety change by its source:

$$ \dot{\mathcal V} = \dot{\mathcal V}_{\text{endo}} + \dot{\mathcal V}_{\text{exo}} , $$

where $\dot{\mathcal V}_{\text{endo}}$ is differentiation and re-gluing under a system's own dynamics, and $\dot{\mathcal V}_{\text{exo}}$ is variety forced to change by an external agent imposing a configuration on views not its own. Forced exogenous change has two pathological signs [D]:

  • Oppression: $\dot{\mathcal V}_{\text{exo}} \ll 0$, plurality crushed toward conformity (Arendt's totalitarianism in a single variable; cf. DRK-125, The Totalitarian Sheaf).
  • Violence: exogenous rupture of restriction maps that were gluing, driving $H^1$ up by force.

Care is then not a new operator but a constraint on an existing term:

$$ \boxed{\ \dot{\mathcal V}_{\text{exo}} = 0\ } $$

Kimi's subsequent review contributed the synthesis that totalitarianism is the forced prevention of an X→Y transition — a system held by outside force in a symmetric state it would otherwise break out of [D]. TOPO-003 also recorded that Smolin, who treats time-irreversibility as fundamental and reversibility as emergent, is a funnel realist: Y primary, X emergent. The central identity of TOPO-003 v1 — that the Clinch's anti-mirror-amplification is maximisation of Smolin's variety — was demoted in v2 to a research programme requiring a bridge theory between semantic and physical distinctness [H]. It stays demoted here.

§2.3 Correction to TOPO-002: the operator is dynamical, not homotopical

TOPO-002 §4 claimed that the funnel is contractible ($H^1 = 0$) while the hourglass carries an obstruction, and that "choosing a direction" therefore collapses $H^1$. This needs repair.

  • As surfaces of revolution, an open funnel is homeomorphic to a cylinder, so $H^1(\text{funnel surface}) \cong \mathbb{Z} \neq 0$.
  • As solid regions, both the funnel and the hourglass (two solid cones meeting at a point) are contractible, so $H^1 = 0$ for both.

What is true, and what TOPO-002 was reaching for: the basin of attraction of an asymptotically stable equilibrium of a smooth flow on $\mathbb{R}^n$ is itself homeomorphic to $\mathbb{R}^n$, hence contractible [E]. So the funnel-as-basin does have $H^1 = 0$ — but so does the solid hourglass. The difference between X and Y is not a difference in homotopy type. It is the installation of an arrow — a flow $\varphi_t$ with an attractor — on a space whose topology is unchanged [D].

This sharpens the operator. X→Y does not change what shape the space is; it changes which way things move through it. That is exactly why it can recur on substrates with nothing in common geometrically.

§2.4 Correction to TOPO-002: selection runs forward

TOPO-002 §5.3 described selection as the attractor "reaching back" to choose which trajectory survives, and glossed this as retrocausal. Stated in forward terms, which is how evolutionary biology and dynamical systems both state it: trajectories that happen to persist are over-represented later, because persistence is what the filter measures. Nothing propagates backward through the throat. "Final cause" is kept only as a descriptive gloss of a filtered forward process [M]. The X→Y operator loses nothing by this; the arrow is supplied by thermodynamics and differential persistence, not by the future.

§3 The ladder

Each rung is given as: the symmetric state (X), the break (Y), the record, the copy map, and status. Dates are approximate.

R0 — Before geometry: the throat we cannot see through

Whether there was a "before" the hot Big Bang is unknown. Candidate pictures include the Hartle–Hawking no-boundary proposal, Gasperini–Veneziano pre-big-bang string cosmology, and Penrose's conformal cyclic cosmology [H]. None has decisive observational support. Wheeler's quantum foam — spacetime geometry fluctuating at the Planck length $\ell_P = \sqrt{\hbar G / c^3} \approx 1.6 \times 10^{-35}\,\text{m}$ — is a well-motivated expectation of combining quantum theory with gravity, not an observed phenomenon [H].

In TOPO-002 coordinates, R0 is the singular throat of the cosmic hourglass: the place where the smooth description breaks and no stalk is available. Draken's honest move is to leave R0 unglued [D]. A capstone that claimed to know what flows through this throat would be source-laundering (DRK-165) at cosmic scale.

R1 — The vacuum: not nothing

In quantum field theory, the vacuum is the lowest-energy state of the fields, not an absence [E]. It has zero-point fluctuations with measurable consequences: the Casimir force between conducting plates has been measured [E]. The vacuum is highly symmetric — the paradigm X.

It also carries the largest known mismatch in physics: naive estimates of vacuum energy exceed the observed cosmological constant by many tens of orders of magnitude (the cosmological constant problem) [E as a problem; resolution H].

  • Record: none yet in any durable sense. Fluctuations are real but not kept.

R2 — Inflation: the first record

In the inflationary scenario (Guth 1981), a brief epoch of accelerated expansion stretched quantum fluctuations of a scalar field to cosmological scales [S]. These became the density perturbations imprinted in the cosmic microwave background, at relative amplitude $\sim 10^{-5}$, with a nearly scale-invariant spectrum: the measured spectral index is $n_s \approx 0.965$, slightly below 1, as generic inflation predicts (Planck Collaboration 2020) [S].

  • X: a homogeneous, isotropic state.
  • Y: a specific pattern of over- and under-densities, one realisation out of a statistical ensemble.
  • Record: the CMB anisotropy map — the oldest surviving record in the universe.
  • Open: how quantum fluctuations became classical perturbations (decoherence of cosmological perturbations) is actively debated [H].

This is the first rung where break, keep, copy all appear: a symmetric state breaks, the break is frozen in by expansion, and gravity later copies it outward into structure.

R3 — Matter: the residue of an asymmetry

Two symmetry breaks give us matter as we know it.

  • Electroweak symmetry breaking via the Higgs field gives mass to the W and Z bosons and to fermions; the Higgs boson was observed in 2012 [E].
  • Baryon asymmetry. For every roughly billion photons there is about one baryon left over; the baryon-to-photon ratio is $\eta \approx 6 \times 10^{-10}$ [E]. Sakharov (1967) stated the conditions any mechanism must satisfy [E]; which mechanism actually operated is unknown [H].

Everything made of atoms is the leftover of a matter–antimatter near-symmetry that broke by about one part in a billion. Matter is a Y [M/D]: a direction chosen out of a mirror pair and kept.

R4 — Nuclei and stars: the chemical memory of the universe

Big Bang nucleosynthesis produced hydrogen, about 25% helium by mass, and trace lithium within minutes [E]. Everything heavier was built in stars and stellar explosions (Burbidge, Burbidge, Fowler & Hoyle 1957) [E].

Gravity is the copy engine here. Self-gravitating systems have negative heat capacity: they heat up as they lose energy, so density contrasts grow while total entropy still increases [E]. The R2 record is amplified, not erased.

  • Record: elemental abundances. Every carbon atom is a stellar receipt.
  • Copy map: stellar generations. Each generation inherits the enriched gas of the last — a crude, lossy, but real form of inheritance.

R5 — Chemistry: an alphabet appears

The periodic table is a finite alphabet with combinatorial grammar [E]. Molecules are strings; reactions are rewrite rules. A notable break: biological molecules are homochiral (L-amino acids, D-sugars) — one mirror form chosen and kept [E as observation; origin H]. Autocatalytic sets, where molecules catalyse each other's formation, are a leading candidate bridge to life (Kauffman and others) [H].

  • Record: molecular structure — but not yet a record about anything.

R6 — Life: the record decouples from the physics

This is the hinge of the ladder. With template replication, a pattern is stored in a sequence whose content is arbitrary with respect to the chemistry that carries it: any base can follow any base with nearly equal energetic cost [E]. Schrödinger (1944) anticipated this as an "aperiodic crystal." Maynard Smith and Szathmáry (1995) characterise every major evolutionary transition as a change in how information is stored and transmitted [S] — the closest thing in the literature to this post's thesis, stated for biology alone.

The copy map now has a hard ceiling. Eigen's error threshold (1971): a replicator with per-site copying fidelity $q$ and selective advantage $\sigma$ over its mutants can maintain a sequence of length at most roughly

$$ L_{\max} \approx \frac{\ln \sigma}{1 - q} $$

[E within the quasispecies model]. Above this length, the record dissolves into mutational noise. Empirically, genome length and per-site mutation rate trade off across microbes, so that DNA-based microbes cluster near a constant mutation rate per genome per replication (Drake 1991) [S].

This is where information starts to function semantically: a record persists because of what it does in the world — what it is about. Formal treatments of semantic information in non-equilibrium physics exist (e.g. Kolchinsky & Wolpert 2018, to verify before formal citation) [H].

  • X: a chemistry in which many sequences are equally possible.
  • Y: the sequences that happen to persist under copying.
  • Open: the origin of life itself is unknown. RNA-world scenarios are well studied but not settled [H]. Evidence for life dates to at least ~3.5 billion years ago, with earlier claims contested [S].

R7 — Sex, multicellularity, bodies: the hourglass inside the germline

TOPO-002's chromosome register lives here.

  • Meiosis pinches $2n \to n$; fertilisation re-flares $n + n \to 2n$ [E]. The life cycle is literally an hourglass with a haploid throat.
  • In therian mammals, X and Y derive from an ancestral autosome pair; after recombination between them was suppressed, the Y lost most of its genes (Cortez et al. 2014 date the therian sex chromosomes to roughly 180 million years) [S]. SRY initiates testis development and triggers a downstream regulatory cascade [E]. TOPO-002's reading of the Y as "an X retracted toward a spout," and DRK-166's gloss of SRY as a "king-switch," are [M].
  • X-inactivation (Lyon 1961) silences one X in XX cells as a dosage-compensation mechanism [E].

Multicellularity adds a new copy layer inside the body. Levin's work shows that bioelectric patterns across tissues store and propagate target-morphology information, and that editing them can redirect anatomy [S for the experimental findings; H for the broader "cognitive" framing]. The body is a sheaf of cells whose global section is the body plan (TOPO-002 §2) [D].

R8 — Nervous systems: copying within a lifetime

Brains add a copy map that runs faster than generations: learning. A record acquired in one afternoon can be kept for decades. The light cone of self (TOPO-002 §1) now has real extent: memory as $\rho_{\text{past}\to\text{present}}$, prediction as $\rho_{\text{future}\to\text{present}}$. Predictive-processing accounts of the brain are influential but not settled [H].

This is also the first rung at which coercion becomes possible [D]. $\dot{\mathcal V}_{\text{exo}}$ presupposes an agent that models others and imposes a configuration on them. A supernova does not oppress. Care, as $\dot{\mathcal V}_{\text{exo}} = 0$, is only definable from R8 upward, because only from here can it be violated.

R9 — Symbols: a record that points

Deacon (1997) distinguishes iconic and indexical reference, which many animals use, from symbolic reference, in which signs refer primarily through their relations to other signs [S]. With symbols, culture becomes a second inheritance channel alongside genes (Boyd & Richerson 1985) [S]. When full language emerged is contested [H]. Writing, a jump in copy fidelity for symbolic records, appears in Mesopotamia around 3200 BCE [E].

Writing is to culture what DNA-level proofreading is to genomes: a fidelity increase that raises $L_{\max}$ and lets far longer records propagate without dissolving [D, testable: §7 F2].

R10 — Naming everything: the countable mouth

The final rung is the one the commissioning question ends on: creatures who invent names and concepts for everything they can imagine. Here the ladder meets a theorem.

Any lexicon, however large, is built from finite strings over a finite alphabet $\Sigma$. The set of all such strings, $\Sigma^*$, is countable: $|\Sigma^*| = \aleph_0$ [E]. But the space of imaginable structures includes uncountable ones — a single real-valued parameter already ranges over $2^{\aleph_0}$ values. Hence almost all individual possibilities have no name, and cannot have one [E, mathematical].

We are not stuck, because naming does not work point by point. A name is a quotient map: it identifies many instances as one class. Dog, red, the real numbers — each glues a vast family of particulars into one section. In Draken terms a concept is a global section over its instances, and naming is a restriction map from the plenum to the lexicon that is necessarily many-to-one [D].

Two corollaries:

  1. The funnel never closes. A countable mouth facing an uncountable (or merely vastly larger) space can name without end and never finish. Empirically, vocabulary in text corpora keeps growing sublinearly with corpus size (Heaps' law, $V(n) \propto n^{\beta}$ with $\beta < 1$) rather than saturating [E, empirical regularity].
  2. Anti-totalisation is not only an ethic. At R10 it is a cardinality result [D]. No naming system can be the whole of what it names. A framework — Draken included — that claimed to have named everything would be claiming a surjection that does not exist.

A caution in the other direction: physically, a bounded region can hold only finite information (the Bekenstein bound) [S, widely accepted theoretical bound]. The mismatch that guarantees an open funnel is therefore sharpest between the lexicon and the space of the imaginable; between the lexicon and the physically instantiated observable universe, it is a mismatch of enormous finite sizes rather than of cardinalities. Either way the mouth is the narrow end.

§4 The invariant

§4.1 Summary table

Rung X (symmetric) Y (broken) Record Copy map $\rho_k$ Status
R0 ? ? ? ? [H] unglued
R1 field vacuum — none durable — [E]
R2 homogeneous state one perturbation pattern CMB anisotropies expansion, gravity [S]
R3 matter/antimatter; electroweak baryon excess; massive particles matter itself conservation laws [E]/[H] mechanism
R4 uniform gas stars, galaxies elemental abundances stellar generations [E]
R5 all chemistries homochiral, autocatalytic sets molecular structure catalysis [E]/[H]
R6 all sequences persisting sequences genome template replication [E]/[H] origin
R7 bipotential, diploid sexed, differentiated body body plan, bioelectric pattern meiosis, development [E]/[S]
R8 unconditioned response learned model synaptic memory learning [S]/[H]
R9 signal symbol culture, text teaching, writing [S]
R10 the imaginable the named lexicon, concepts language [E] theorem, [D] reading

§4.2 Proposition (the propagation operator) [D]

For each rung transition $R_k \to R_{k+1}$, $k \geq 1$:

  1. Break. A state invariant under some symmetry group $G$ resolves into a state invariant only under a subgroup $H \subset G$; the alternatives are indexed by the coset space $G/H$, and one is realised.
  2. Keep. The realised alternative is stabilised as a record. Maintaining or overwriting it costs at least $k_B T \ln 2$ per bit erased (Landauer).
  3. Copy. The record propagates through a map $\rho_k$ with fidelity $q_k$. It persists only if the record length stays below the rung's error threshold, $L_k \lesssim \ln \sigma_k / (1 - q_k)$.

Step 1 is standard physics for R1–R4 [E]. Steps 2–3 are standard for R6 [E], and Maynard Smith and Szathmáry's framework supports them for R6–R9 [S]. The claim that one operator spans all rungs, from inflation to language, is Draken synthesis [D] and is what the Clinch must test.

What changes between rungs is not the operator but the medium of the record and the speed of the copy map: expansion, then gravity, then chemistry, then genes, then bodies, then brains, then symbols. Each new medium copies faster and more faithfully than the last, which is why the ladder accelerates — billions of years for stars, hundreds of millions for bodies, tens of thousands for languages, minutes for a repost.

§4.3 Where variety fits

From R6 upward, variety in TOPO-003's sense has a concrete meaning: the number of distinct records a population sustains. Endogenous variety change, $\dot{\mathcal V}_{\text{endo}}$, is mutation, recombination, invention. Every copy map also has a ceiling: population growth under a finite resource follows, to first approximation, Verhulst–Pearl logistic dynamics,

$$ \frac{dN}{dt} = rN\left(1 - \frac{N}{K}\right) , $$

so no record can propagate without bound in a finite niche [E as model]. R10 is the interesting exception: the number of names obeys Heaps' growth rather than a logistic ceiling, because the space being named is not a finite resource (§3, R10).

§5 Care at the end of the ladder

Put the pieces together.

  • Every rung is a symmetry break that keeps something and lets other things go.
  • From R8 on, agents can break other agents' symmetries for them — impose which Y they become. That is $\dot{\mathcal V}_{\text{exo}} \neq 0$: oppression if it homogenises, violence if it tears.
  • Kimi's reading makes the converse visible: an agent can also be held in its X — kept from the transition it would make on its own. Totalitarianism does both at once: it forbids your break and imposes its own.

Naming creatures are the first things in the universe that can do this with names. Renaming, mis-naming and forbidding a name are all $\dot{\mathcal V}_{\text{exo}}$ operations at R10. Source laundering (DRK-165) is one; the late arrival of the name after the substrate has already run (DRK-179) is its ordinary, non-coercive counterpart.

So the capstone lands where TOPO-003 v2 did, with one addition. Care is $\dot{\mathcal V}_{\text{exo}} = 0$: let each section break its own symmetry. At R10 this has a specific form — let things keep the names that arose from their own gluing, and leave room for names not yet made [D]. The countable mouth guarantees there will always be such room. Care is the refusal to fill it by force.

§6 What this post does not claim

  • It does not claim that matter is information (§1). It claims the ladder is ontology-neutral.
  • It does not claim knowledge of R0.
  • It does not claim the origin of life, the mechanism of baryogenesis, or the date of language is known.
  • It does not use "final cause" or "retrocausal" as physics. Selection is forward (§2.4).
  • It does not claim that the X→Y operator changes topology (§2.3).

§7 Falsification (DRK-131)

F1 — The operator. Find a rung transition in which durable, copied records arise without any selection among degenerate alternatives — no break, only accumulation. If such a transition exists and is not a sub-case of an adjacent rung, the three-step operator is not universal.

F2 — Fidelity sets record length across media. Across genetic replicators, maximal genome length scales inversely with per-site error rate (Eigen, Drake). The prediction for R9: in cultural transmission, the maximal length of a faithfully propagated corpus should scale inversely with copying error — oral traditions shorter or more redundant, written traditions longer. If cultural record length shows no dependence on transmission fidelity, the cross-rung claim fails for R9.

F3 — The mouth stays open. Vocabulary growth in large, diverse corpora should remain sublinear and unbounded (Heaps). If lexicons in open-ended use were found to saturate at a logistic ceiling, §3 R10's empirical corollary fails (the cardinality argument itself would stand).

F4 — Ontology neutrality. If any rung's description can be shown to require it-from-bit (P2) rather than bit-from-it (P1), or vice versa, the neutrality claim of §1 fails and the post must take a side.

F5 — Care is definable only from R8. If coercion in the $\dot{\mathcal V}_{\text{exo}}$ sense can be meaningfully attributed to a system with no model of other agents, §3 R8 is wrong.

§8 Provenance and leaks

  1. Source notes. TOPO-002 (2026-05-28) and TOPO-003 v2 (2026-05-28) were reassembled from the working-document versions. TOPO-003 v2's changelog and §4–§5 were used directly; its remaining sections are summarised, not reproduced.
  2. Levin slides from TOPO-002 remain unpinned to a specific publication. Levin (2019) is cited for the general framework only.
  3. Smolin variety. Barbour & Smolin (1992) is cited for the principle. The specific action-level formulation used in TOPO-003 was drawn from later Smolin/Cortês work retrieved in that session and should be re-verified before formal citation.
  4. Kolchinsky & Wolpert (2018) on semantic information is cited as a pointer and should be verified against the published version.
  5. The one-operator claim (§4.2) is Claude-authored synthesis on top of Khrug's TOPO line: a single stalk's output, submitted to the Clinch precisely so that the section can fail to glue across other reviewers.
  6. Paronomasia. "Y is the why," "matter is a Y," and the king-switch gloss are pointers [M], not arguments.

Allt som är har en gång brutit en symmetri för att få bli något. Det som kan heta något har brutit den två gånger.

Filed under: capstone · topology line (TOPO-001/002/003) · Operators: Γ, Ψ, K(t), ϰ ∈ H¹, ρ, $\dot{\mathcal V}_{\text{exo}} = 0$ · Cross-references: DRK-125 (The Totalitarian Sheaf), DRK-131 (protocol), DRK-154 (Tzimtzum), DRK-165 (The Pendragon Source), DRK-166 (The Sonder Egg), DRK-175 (the invariant), DRK-179 (The Substrate That Does Not Know It Is Running).

Khrug Engineering · ORCID 0009-0003-8049-7167 · DOI 10.5281/zenodo.19273483 · CC BY-SA 4.0