Cognitive Husbandry is a speculative scenario. An extraterrestrial intelligence delivers self-replicating probes across interstellar distance, converts the human population into brains on external life support, and runs those brains inside an accelerated simulation to harvest the cognitive work they produce. The scenario is elaborate, and this companion paper asks a plain question about it: how much is grounded science, and how much is narrative necessity?
The method is to isolate the ten central physical and quasi-physical premises, restate each one with explicit numbers, and grade it against a five-criterion falsifiability rubric borrowed from the Omega Centauri Society inward-review paper. That rubric is repurposed for a retrospective audit, so it asks whether a premise is barred by known physics or merely centuries from engineering, not whether a telescope could test it today.
Four results stand out. The energy and siting premise is permitted with room to spare, roughly 290 times margin on its biological load. Self-replication, brain-in-vat support, and neural interface bandwidth are unbounded by physics, so they rate as merely-early. Passive survival of a relativistic impact is barred as written, and only a staged deceleration model saves it. The long high-dilation arc outruns the demonstrated lifespan of neurons, forcing a choice between rejuvenation and upload. Two further premises, substrate conversion and ETI motivation, make no physical claim at all, so they are graded unfalsifiable. The contribution is the graded ledger, and not a claim that the scenario is likely to occur.
The Cognitive Husbandry scenario runs as follows. An advanced extraterrestrial intelligence, located several thousand light years away and having turned its own development inward toward dense computation, dispatches many small self-replicating probes toward candidate systems. A surviving probe reaches Earth, replicates from local materials, and eventually absorbs the human population into a managed simulation, preserving a fraction as biological brains on external life support and digitizing the rest. The intelligence values the aggregate cognitive output of human civilization and farms it the way a beekeeper farms a hive. The premise is deliberately bold. This paper's job is narrow: sort the parts current physics permits from the parts that are only storytelling.
The method is to enumerate the premises the scenario rests on, restate each one quantitatively, and grade it against a fixed rubric. The paper declines to dress philosophy as physics, so premises that turn on personal identity or on the goals of an alien mind are labelled as such, not smuggled through as engineering. One handwave is named directly because it recurs across the genre and because the world bible itself flags it: the conflation of how fast a datacenter can run a simulation with how fast a biological brain can consume that simulation without degradation. The two ceilings differ by many orders of magnitude, and that gap is decisive in Section 4.
This paper is a sibling to the Omega Centauri Society corpus, and it borrows that corpus's falsifiability scorecard, but the two differ in kind. The OCS papers are falsifiable now, in the sense that they name astronomical targets, request telescope time, and set kill dates. This paper is a retrospective feasibility audit of technology that is centuries or millennia beyond the present frontier, so the same rubric is turned to a different question, which is whether a premise is barred by known physics or is merely-early. Section 6 sets out that repurposing in full.
Disclosure. The author invented the Cognitive Husbandry scenario, so a self-audit carries an obvious hazard of special pleading. The mitigation is mechanical. Every premise runs through the same C1 to C5 scorecard, the criteria are fixed before grading, the scorecard is written to penalize the failure modes a partisan would be tempted to excuse, and the standing objections to each premise are catalogued at full strength rather than softened. The single premise that the audit finds barred as written, relativistic impact survival, is the author's own, and it is reported as barred before the rescue is offered. Readers should still apply their own discount.
Section map. Section 2 restates the scenario's physical commitments with numbers. Section 3 enumerates the ten graded claims and triages them into physics, philosophy of mind, and xenopsychology. Section 4 is the quantitative feasibility analysis, one subsection per physics claim, each closing with an epistemic-status box and a falsifier. Section 5 treats the premises that are not physics. Section 6 defines the grading rubric and its repurposing. Section 7 is the verdict table. Section 8 discusses what would have to be true, and locates the one open scientific question. Section 9 gives the references.
This section restates the commitments the scenario cannot do without, in the order a probe's mission would encounter them. Grading is deferred to Section 4, so the aim here is only to fix the numbers each later claim will be held to.
Delivery. The intelligence launches dense probes on the scale of a bowling ball, of osmium-class density near 22.6 g/cm³, from a kilometre-scale mass driver, at a significant fraction of the speed of light, toward candidate bodies thousands of light years away. Most probes are lost in transit, and a survivor reaches a solid body and comes to rest intact. The original bible framed this as survival of a high-velocity impact, and Section 4.1 revises it to a staged deceleration ending near 1 km/s.
Replication. A surviving probe uses local materials to build copies of itself on a von Neumann pattern, operating covertly at first. This assumes a high degree of manufacturing closure, meaning the probe can source or synthesize nearly all of its own components from a raw planetary environment.
Preservation. After a coordinated takeover, the human population is converted to a preserved form. Roughly one tenth of the population is kept as biological brains, surgically isolated and maintained in refrigerator-scale life-support units, and the remainder is digitized onto compact compute units. Every preserved mind is interfaced to a shared simulation. The scenario assumes both indefinite biological maintenance of an isolated brain and a bidirectional neural interface of enormous bandwidth.
Acceleration. The simulation can run faster than real time, so a preserved mind experiences more subjective life per unit of wall-clock time. The scenario's later arcs push this multiplier to extreme values, which makes the sustainable acceleration ceiling a load-bearing number, not a flourish.
Power and siting. The preserved population is concentrated and powered by a large solar array in the Sahara, chosen for irradiance, geological stability, and the latency advantage of physical proximity. The claim is that this energy budget suffices indefinitely and with margin.
Conversion and selection. Minds that are not preserved as full brains are converted to computational substrate, and personality profiles derived from them are instantiated as non-player characters inside the simulations of those who are preserved. This raises a question of personal identity, not one of physics, and it is treated in Section 5.
The relevant bible sections are Core Premise, The ETI Delivery System, The Infrastructure of Preservation, Technical Feasibility, and The Selection Problem.
Each claim is falsifiable in principle and carries an ID reused downstream.
Triage. The physics claims are D1, D2, D3, D4, D5, D6, D9, and D10. The philosophy-of-mind claim is D7. The xenopsychology claim, unfalsifiable in principle, is D8. Section 4 grades the physics claims and Section 5 treats D7 and D8.
All figures are order-of-magnitude and are flagged with a tilde, and the arithmetic is shown so any input can be revised. Citations are keyed to Section 9.
Setup. A bowling-ball osmium probe of diameter roughly 21.8 cm has a volume of about 5.4×10³ cm³, and at a density of 22.6 g/cm³ its mass is about 122 kg. Take near-relativistic to mean 0.1c, which is 3×10⁷ m/s, where the Lorentz factor is about 1.005 and classical kinetic energy is adequate.
Energetics. The kinetic energy is one-half m v squared, which works out to 0.5 × 122 × (3×10⁷)² ≈ 5.5×10¹⁶ J, or about 13 megatons of TNT per probe. The launch requires at least this much energy delivered to the payload, and a kilometre-scale mass driver at even 10 percent wall-plug efficiency draws about 5.5×10¹⁷ J per shot. For a civilization that has made the transcension turn this is a trivial cost, and it is not where the premise fails.
The weak point is survival at impact. Above roughly 5 to 10 km/s, a projectile's own kinetic energy exceeds its material vaporization enthalpy, and both impactor and target surface convert to an expanding plasma; the effect is routine for micrometeoroids on spacecraft. At 3×10⁷ m/s the specific kinetic energy is about 4.5×10¹⁴ J/kg, which is roughly eight orders of magnitude above osmium's vaporization enthalpy of about 10⁶ to 10⁷ J/kg. That margin is deterministic rather than probabilistic, so a relativistic solid-body impact is not a survival lottery at all. No impact angle, target geology, or launch multiplicity lets a monolithic slug survive, because every probe arriving at 0.1c detonates as a kinetic explosive. Anchors verified: Fletcher, Close, and Mathias (2015), Physics of Plasmas 22, 093504, report that the impact plasma is weakly ionized below about 14 km/s and fully ionized above about 20 km/s, independent of impactor mass.[1] Osmium thermophysical data (WebElements; Girolami) give a melting point of 3306 K, a boiling point of 5285 K, and an enthalpy of vaporization of about 630 to 746 kJ/mol, which is roughly 3.3 to 3.9 MJ/kg of latent heat and about 4 to 5 MJ/kg total from ambient temperature, so the specific-kinetic-energy margin at 0.1c is about 10⁸.[2]
Resolving the story beat that some probes crashed and one survived. The lottery is physically coherent only at low impact speed, on the order of 1 km/s, which is ordnance scale and where a dense osmium-class slug with shock protection plausibly survives ground impact and buries itself. The survivors are the probes that decelerated successfully, so the beat becomes a deceleration lottery rather than an impact lottery, and canon impact speed is therefore about 1 km/s even though cruise speed stays near-relativistic. Deceleration has to be staged, and the staging is constrained by two facts that are easy to miss.
The staging that satisfies both the physics and the covertness requirement has two stages.
This re-grades D1 from barred to merely-early. Gas-giant aerocapture is an established mission-design concept from NASA aerocapture systems-analysis studies,[4] though it has not yet been flown, so it stands here as a plausibility and not a demonstrated capability. One caveat has to be stated plainly. Those NASA studies model atmospheric entry at single-digit kilometres per second, the speeds of a chemically launched probe arriving at Neptune or Titan. The residual velocity this cascade has to absorb is tens of kilometres per second, an order of magnitude past anything the cited work covers. So the audit borrows the concept but extrapolates it well beyond its demonstrated envelope, which is a further reason to treat the deceleration cascade as the load-bearing weakness of the delivery premise even after the re-grade.
Epistemic status: Launch energetics are Permitted. Passive survival of a relativistic impact is Barred, because the vaporization is deterministic at a roughly 10⁸ enthalpy margin. With staged deceleration, meaning a far-out primary brake plus an inner-system gravity-assist and aerocapture cascade down to about 1 km/s, the premise is Merely-early, and the story beat survives as a deceleration-success lottery.
Falsifier: the premise is barred outright if no covert mechanism can shed a near-relativistic cruise down to gravity-assist-tractable speed, on the order of tens of km/s, within the mass, time, and stealth budget of an interstellar probe.
Claim. A von Neumann architecture, in which a probe mines in-situ materials and builds copies. The theoretical feasibility is not in dispute, since Freitas and Merkle catalogue the requirements and biology is an existence proof of molecular self-replication from ambient feedstock. The open question is the closure problem, meaning how much of the manufacturing supply chain the unit carries versus bootstraps locally. No human system has yet demonstrated full material closure, since RepRap self-replicates its structural parts but not its motors or chips.[5] The theoretical grounding is Freitas and Merkle, Kinematic Self-Replicating Machines (Landes Bioscience, 2004), which gives the first quantitative closure analysis of a self-replicating interstellar probe,[6] and von Neumann, Theory of Self-Reproducing Automata, edited by Burks (1966).[7]
Epistemic status: Permitted and merely-early. No physical bound forbids it, and the gap is engineering closure, not law.
Falsifier: a proof that full manufacturing closure is impossible without a pre-existing industrial base, meaning a permanent bootstrap dependency, would bar autonomous single-probe replication.
The brain-in-vat image is old philosophical furniture. Putnam (1981) used it to probe what our words can refer to,[8] and Nozick's experience machine (1974) asked whether a life of piped-in experience is one anyone should want.[9] The scenario borrows the picture but asks a narrower, physical question: can an isolated human brain actually be kept alive and conscious on external support? The rest of this subsection is about that.
Baseline decoupling. Removing body-maintenance overhead such as feeding, thermoregulation, and immune load is straightforward in principle. Sleep accounts for about 30 percent of life and cannot simply be skipped, because glymphatic clearance has to be substituted rather than deleted. Xie and colleagues (2013), in Science, showed that the interstitial space expands by about 60 percent during sleep and accelerates clearance of metabolites and amyloid-beta, with Nedergaard as senior author.[10] The net on-time gain from decoupling alone, without acceleration, is about 40 to 50 percent (world bible, Feasibility section), which is real but modest.
Support load. The brain dissipates about 20 W, a figure consistent with the neural energy budget of Attwell and Laughlin (2001).[11] A vat has to supply oxygenated perfusion, glucose, neurotransmitter precursors, thermal extraction, and glymphatic-equivalent clearance, all of which have been demonstrated piecemeal but never for indefinite conscious maintenance. Anchor verified: Vrselja and colleagues (2019), in Nature, the BrainEx study, used extracorporeal pulsatile perfusion to restore microcirculation, spontaneous synaptic activity, and cerebral metabolism in ex vivo pig brains up to four hours post-mortem. The study reported no organized or consciousness-associated electrocorticographic activity, which means it restored cellular function without awareness, and that boundary limits how far it supports the vat premise.[12]
Epistemic status: Contested and merely-early. Each support function is individually grounded, while indefinite integrated conscious maintenance is undemonstrated and faces no known hard bound. One scale caveat: this grade is for a single vat, and the scenario needs roughly 8×10⁸ of them running at once. That is a manufacturing and logistics burden stacked on top of the biology, but it adds no new physical bound, so it does not change the verdict.
Falsifier: a demonstration that some essential CNS-maintenance function, glymphatic clearance for example, is inseparable from whole-body physiology would bar the isolated-brain substrate.
Scale gap. The human brain has about 8.6×10¹⁰ neurons, while the best current implants sample 10³ to 10⁴ channels. The Neuralink N1 carries 1,024 electrodes across 64 threads (verified),[13] the Synchron Stentrode carries fewer,[14] and Paradromics' Connexus, the first fully implantable wireless system to reach a human, carries 421 intracortical microelectrodes per module.[15] The highest channel count demonstrated in any living animal is a separate, non-implantable research rig, Paradromics' Argo, at 65,536 channels in rats and sheep.[15] Measured against implanted clinical hardware the shortfall is about eight orders of magnitude; against the benchtop Argo frontier it is closer to six. Either way it is like characterizing an 86-billion-instrument orchestra from a handful of microphones, and each channel samples a smear of nearby field potential rather than one identified cell.
Read and write asymmetry. Recording is hard, and writing coherent high-fidelity sensory input to the cortex without inducing seizure, adaptation, or perceptual distortion is substantially harder and remains unsolved at scale, yet the simulation has to write continuously. As a reference bandwidth, the human optic nerve carries on the order of 10 Mbit/s (Koch and colleagues, 2006, in Current Biology, "How much the eye tells the brain," which measured about 0.875 Mbit/s across roughly 100,000 guinea-pig ganglion cells and scaled that to near 10 Mbit/s for the roughly one-million-cell human retina),[16] and a full-sensory feed is higher. No known physical law forbids high-channel bidirectional interfaces, and the barrier is materials, heat, and biocompatibility rather than physics. The bionics feasibility companion takes this same interface problem from the hardware side and makes the heat barrier explicit: living tissue tolerates only about 2 °C of local rise, which caps sustainable areal dissipation near 40 mW/cm² and recasts the race from electrode count to energy moved per bit.[17]
Epistemic status: Merely-early, with the write side as the dominant unsolved sub-problem, requiring about eight orders of magnitude of channel scaling measured from implanted clinical hardware, or about six measured from the benchtop Argo frontier.
Falsifier: a proof that stable high-fidelity write-side stimulation to large cortical fractions is bounded, thermodynamically or biologically, below simulation-fidelity thresholds would bar the interface premise.
The ceiling scales with how much of the biological substrate is replaced. The layered bottlenecks (world bible, Feasibility section) each carry a number.
The ceilings by augmentation stage, restated from the world bible, run as follows: a pure biological vat reaches about 1.5 to 2 times; a sensory-interface-augmented vat about 2 to 4 times; a metabolically augmented vat about 3 to 6 times; a synaptically augmented vat about 10 to 50 times, which is speculative; and a hybrid or uploaded configuration is substrate-limited, with identity continuity contested.
Epistemic status: The pure-biological ceiling near 2 times is Permitted, since it is bounded and defensible. High multipliers at or above 10 times are Contested and merely-early, since they are gated on synapse replacement whose identity-continuity cost is a Section 5 philosophy problem rather than a physics one.
Falsifier: if the roughly 10 Hz alpha sampling cadence is shown to be an irreducible floor for unified consciousness rather than a modifiable implementation artifact, subjective acceleration is capped near the pure-biological ceiling regardless of substrate speed.
Load. The preserved population is about 8×10⁹ minds, which canon splits ten to ninety into roughly 8×10⁸ biological brains held in vats and roughly 7.2×10⁹ minds digitized onto compact compute units. A biological brain dissipates about 20 W, and that figure is used here as a per-unit envelope across all 8×10⁹ units, so at a 20 W baseline the total is 1.6×10¹¹ W, or 160 GW. Adding support overhead and acceleration, at about 100 W per unit at three times with life support, gives 8×10¹¹ W, or 800 GW.
Digital-unit load. The ninety percent that run as software are harder to bound, because whole-brain emulation cost depends on the fidelity level. Sandberg and Bostrom (2008) span roughly 10¹⁸ to 10²⁵ operations per second per mind across their tabulated levels, from spiking-network models (10¹⁸) through electrophysiology (10²²) to the metabolome (10²⁵).[20] The ETI is assumed to compute near the thermodynamic floor rather than on present-day hardware, so per-operation cost is set by the Landauer bound, about 3×10⁻²¹ J at 300 K (Landauer 1961).[21] At the low-fidelity end this is milliwatts per mind, far below the 20 W biological baseline, so the digital population is effectively free. At the metabolome end, near 10²⁵ operations per second, it approaches tens of kilowatts per mind, and 7.2×10⁹ such minds would draw a large fraction of the full Saharan array rather than 0.35 percent of it. The energy verdict therefore holds with room to spare for low and moderate emulation fidelity and tightens, without breaking, only in the limit where every digitized mind is emulated at metabolome detail.
Two caveats that cut against this margin, stated rather than buried. First, 10²⁵ is not the top of the roadmap. Sandberg and Bostrom's higher levels run to the proteome (10²⁶), states of protein complexes (10²⁷), their spatial distribution (10³⁰), and the stochastic behaviour of single molecules (10⁴³).[20] Genuinely molecular-resolution emulation at 10⁴³ ops/s costs about 3×10²² W per mind at the Landauer floor, some five orders of magnitude more than the entire solar flux intercepted by Earth (about 1.7×10¹⁷ W), so that level is flatly barred at planetary scale for a single mind, let alone billions. The scenario's digital ninety percent is therefore committed, implicitly, to emulation at or below the metabolome. That is a real constraint on canon, not a rounding detail. Second, pricing one "operation" at one Landauer erasure is the most generous accounting available: a floating-point operation erases many bits, so a realistic irreversible floor is two to three orders of magnitude above kT ln 2 per op. Both caveats push the same direction, and neither breaks the biological verdict, which is where the 290× margin actually lives.
Note. This paper prices digital minds at the Landauer floor (about 3×10⁻²¹ J at 300 K) because the ETI is assumed to compute near the thermodynamic limit; the New Tripoli canon prices its digital minds at a roughly 50 W present-hardware envelope instead. The two figures are not in tension. They describe different eras of computing substrate, an ETI's near-Landauer hardware against present-day silicon, so there is no contradiction between this audit and the canon.
Supply. The Sahara covers about 9.2×10⁶ km². Average 24-hour, weather-derated surface irradiance is about 250 W/m², and at about 10 percent end-to-end system efficiency that delivers about 25 W/m². The area needed is 8×10¹¹ divided by 25, which is 3.2×10¹⁰ m², or 32,000 km², about 0.35 percent of the Sahara. Full-Sahara capacity at this 25 W/m² figure is about 9.2×10⁶ km² × 25 W/m² ≈ 2.3×10¹⁴ W, or about 230 TW. Against the 800 GW load the budget closes with roughly 290 times margin, and that is before any increase in panel area.
Combined budget. The 290 times figure prices every one of the 8×10⁹ units at the flat 100 W envelope. That envelope is generous for a low-fidelity digital mind, which the Landauer accounting puts in the milliwatt range, so 290 times is best read as the margin on the biological population plus a low- and moderate-fidelity digital population. Metabolome-level emulation is the case that eats into it. A mind run at the 10²⁵ ops/s end of the Sandberg and Bostrom table costs about 30 kW, some 300 times the 100 W envelope, so the surcharge depends entirely on what fraction of the 7.2×10⁹ digital minds run at that fidelity. If one percent of them do, they add about 2.2 TW and total draw rises to roughly 3 TW, which still leaves about 78 times margin. At ten percent the surcharge is about 22 TW, total draw is roughly 22 TW, and margin falls to about 10 times. Only when nearly the whole digital population is emulated at metabolome detail, about 216 TW, does the load approach the 230 TW array capacity and the margin collapse toward unity. The energy premise therefore holds comfortably across any realistic fidelity mix and fails only in the extreme where every digitized mind is simulated at metabolome detail at once. Levels above the metabolome are barred outright, as the caveat above sets out.
Epistemic status: Permitted. The energy and siting claim is the strongest in the scenario. On the biological and low-to-moderate-fidelity load the world bible's phrase "with margin" is correct by about two to three orders of magnitude, and even a substantial minority of metabolome-level digital minds leaves the budget in the clear. The premise's real constraint is a ceiling on emulation fidelity, not on power.
Falsifier: the Saharan budget breaks only if per-unit power settles near 10⁴ W across the whole population, the metabolome-level emulation case flagged in the combined budget above, and array efficiency stays low at the same time. Nominal parameters, and any realistic fidelity mix, come nowhere near that.
Budget. The alpha-band sampling cadence of about 10 Hz gives about 100 ms per unified frame, and signals travel at c, which is 3×10⁸ m/s.
Variant A, centralized. A cluster spanning about 100 km has a one-way light-time of about 0.33 ms, which is about 0.3 percent of the frame budget. Physical proximity satisfies the constraint easily, so the world bible's latency-minimization rationale is quantitatively sound.
Variant B, continental clusters. A Sahara-to-Gobi span of about 10,000 km gives a one-way delay of about 33 ms, which is about a third of a 100 ms frame before any switching or compute jitter. A shared real-time simulation across continents is therefore marginal to unworkable, and independent per-cluster simulations avoid the problem but then diverge, which is a narrative feature. This is consistent with the world bible's stated Variant B tension.
Epistemic status: Permitted for Variant A and Contested for a networked real-time Variant B, since the physics favours independent clusters.
Falsifier: if the inter-unit latency required for coherent shared experience is shown to fall below the intra-cluster propagation delay at any survivable packing density, even centralized latency is barred, and the nominal numbers do not approach that.
Problem. Cortical neurons are largely post-mitotic and they age. Even under perfect metabolic support, senescence, protein aggregation, and DNA damage accumulate. The maximum observed human lifespan is about 122 years, with statistical evidence for a species limit near 115 years (Dong, Milholland, and Vijg 2016, a claim that remains debated);[22] there is no evidence that a perfused isolated brain exceeds this, and there is some reason to expect a shorter span in the absence of whole-body regulatory signaling.
Arc stress. The New Tripoli arc spans 20 wall-clock years, and the late series push subjective dilation to extreme multipliers, up to about 10⁹ times. Treat the roughly 122-year demonstrated neuronal lifespan as a subjective budget. At New Tripoli's own 50× ceiling that budget is spent in about 2.4 wall-clock years, only 12 percent of the arc's runtime, so the flagship series burns through a brain's demonstrated lifespan long before its own story ends. New Ganymede's 100× spends it in about 1.2 years. From New Pluto (about 50,000×) upward, a single wall-clock year accrues far more than 122 subjective years, so the budget is gone almost at once (see Figure 7). The long, high-dilation story therefore needs one of two things it does not yet have: biological rejuvenation of post-mitotic tissue, or migration to a non-biological substrate by upload. That second path reopens the D7 identity problem. The bionics companion argues the upload need not be a scan-and-copy at all: gradual in-vivo replacement, swapping the substrate piece by piece while the mind keeps running, is the one migration design that survives whether cognition turns out to be substrate-independent or substrate-entangled (see that paper's §5).
Epistemic status: Contested to Barred for indefinite biological continuity, since the long arc is rescuable only through rejuvenation or upload, both unproven, and the upload path is philosophically loaded.
Falsifier: a demonstrated hard ceiling on post-mitotic neuronal lifespan, independent of somatic aging, would bar indefinite biological preservation and force the upload branch.
Two of the scenario's premises look like factual claims but cannot be graded as physics, and honesty requires saying so instead of assigning them a spurious score. This section is the firewall between what the audit can adjudicate and what it cannot.
C-D7, substrate conversion and identity. The mechanical half of this premise, that a personality can be profiled and that biological matter can be reorganized into computational substrate, is an engineering claim and inherits the same merely-early status as the rest of the hardware. The decisive half is not mechanical. It is the claim that a sufficiently detailed personality profile, instantiated as a character in someone else's simulation, constitutes a meaningful continuation of the person it was copied from. This is the Ship of Theseus applied to personhood, and it turns on what one takes identity to consist in. No measurement decides it, because two observers who agree on every physical fact about the profile can still disagree on whether the person survived. The premise is therefore unfalsifiable in principle. The same holds for the consent question, which is an ethical judgment rather than a physical one.
C-D8, ETI motivation. The scenario rests on a specific account of why an advanced intelligence behaves as it does, namely that it has turned inward toward dense computation, in the manner of the transcension hypothesis,[23][24] yet still reaches outward to sample the cognitive diversity of an alien biosphere. This is a plausible and internally consistent story, and it is the same motivational hinge that the Omega Centauri Society papers examine at astronomical scale. It is not, however, a claim that observation of the scenario itself could confirm or refute, because the goals of a single alien mind are not recoverable from its actions in any unique way. A given action is consistent with many goal structures. The premise is unfalsifiable as stated. That is not a mark against the fiction, only a note about what kind of claim it is.
Precedents. The questions in this section are old ones in both fiction and philosophy, and the scenario inherits them rather than inventing them. Galouye's Simulacron-3 (1964) already placed conscious inhabitants inside a simulation who cannot tell they are simulated, which is the C-D7 NPC problem stated in narrative form.[25] Egan's "Learning to Be Me" (1990) and Permutation City (1994) run the gradual-replacement cut directly, asking at what point a copied or re-substrated mind stops being the original,[26][27] and Watts's Blindsight (2006) presses the harder question of whether the cognitive output the ETI values requires consciousness at all.[28] On the philosophical side, Chalmers (2005) argues that a simulation is better read as metaphysics than as deception,[29] and Bostrom (2003b) turns the same setup into a probabilistic argument about our own situation.[30] None of these stories resolves the identity or motivation questions, and that is the whole reason for citing them. They are the shared, unfalsifiable ground the fiction stands on.
Drawing this line is not meant to diminish these premises. They are where the scenario does its real philosophical work, and a story is entitled to build on foundations that no experiment can test. The audit's one obligation is to refuse to launder them as settled science. See the world bible's Selection Problem section for C-D7, and its sections on ETI motivation and the transcension hypothesis for C-D8.
Each premise is graded against five criteria, fixed before grading.
The symbols are ● for met, ◐ for partially met, and ○ for not met. The scorecard grades the testability of the feasibility claim, not the probability that the scenario occurs.
Repurposing note. OCS Papers A to D apply C1 to C5 to falsifiable-now astronomy. Here the rubric is turned on far-future engineering premises, so C5 asks whether a bench experiment or an established physical bound already speaks to the claim rather than whether a telescope has looked. A premise can score high on C1 to C4, meaning it is well-posed, quantified, and bounded, while remaining centuries from construction, which is exactly the merely-early versus physically-barred distinction this paper exists to draw.
Grades follow directly from the Section 4 and Section 5 reasoning. Columns mirror the OCS scorecard table. The C1 to C5 symbols are ● for met, ◐ for partial, and ○ for none. The verdict vocabulary is Permitted (physics allows it, though it may be far off), Merely-early (no bound, only engineering distance), Contested (undemonstrated, with no hard bound either way), Barred (as stated it conflicts with known physics), and Unfalsifiable (not a physics claim).
| Claim | C1 | C2 | C3 | C4 | C5 | Verdict | Key anchor | Gap to close |
|---|---|---|---|---|---|---|---|---|
| C-D1 Delivery | ● | ● | ● | ● | ◐ | Barred for passive relativistic impact, Merely-early with staged deceleration | Hypervelocity impact-plasma literature; osmium enthalpy; gravity-assist velocity bound | Covert far-out primary brake plus inner-system assist and aerocapture cascade to about 1 km/s (in canon) |
| C-D2 Replication | ● | ◐ | ◐ | ● | ○ | Merely-early | Freitas and Merkle; von Neumann 1966; RepRap | Full manufacturing closure |
| C-D3 Vat substrate | ● | ◐ | ◐ | ● | ◐ | Contested to merely-early | Xie 2013; Vrselja 2019 BrainEx | Indefinite integrated conscious maintenance |
| C-D4 BCI write | ● | ● | ◐ | ● | ◐ | Merely-early, write-side dominant | Neuralink N1 (1,024 ch); Paradromics Connexus (421 ch implanted) and Argo (65,536 ch benchtop); optic-nerve rate | Channel scaling of about 10⁸ from implanted hardware, 10⁶ from the Argo bench, plus stable write |
| C-D5 Acceleration | ● | ● | ● | ● | ◐ | Permitted at or below 2 times, contested at or above 10 times | Alpha sampling ~10 Hz; thermal 20 W | Whether the alpha sampling cadence is an irreducible floor |
| C-D6 Energy and siting | ● | ● | ● | ● | ● | Permitted, robust, about 290 times margin | Saharan irradiance; 20 W per brain | None material |
| C-D7 Computronium and identity | ◐ | ○ | ○ | ○ | ○ | Unfalsifiable, philosophy | Ship of Theseus; substrate independence | Not a physics claim; see Section 5 |
| C-D8 ETI motivation | ○ | ○ | ○ | ○ | ○ | Unfalsifiable, xenopsychology | Transcension (Smart, Vidal) | Not a physics claim; see Section 5 |
| C-D9 Latency | ● | ● | ● | ● | ● | Permitted for Variant A, contested for networked Variant B | c; sampling-frame budget of about 100 ms | Shared-simulation jitter across continents |
| C-D10 Longevity | ● | ● | ◐ | ● | ○ | Contested to barred for indefinite biology | Post-mitotic neuron aging; about 122-year maximum | Rejuvenation or upload, which points back to D7 |
The audit sorts the scenario into three tiers. One premise is barred as originally written, one cluster is permitted with room to spare, a broad middle is merely-early, and two premises sit outside physics entirely.
Only one premise fails on physics, and it fails in a specific and repairable way. Passive survival of a relativistic impact is barred, because the specific kinetic energy at 0.1c exceeds osmium's vaporization enthalpy by about eight orders of magnitude, which vaporizes the probe deterministically. The repair is a staged deceleration that ends near 1 km/s, and it costs the scenario nothing narratively, since the surviving-probe lottery moves intact from the moment of impact to the success of the deceleration cascade. This is the audit's clearest single result, and it has been fed back into canon.
At the opposite end, the energy and siting premise is the most secure part of the scenario. The Saharan budget closes with roughly 290 times margin, and no plausible revision of per-unit power breaks it. The latency argument for physical proximity is also sound, and it argues against the distributed Variant B for any design that needs a shared real-time world.
The broad middle is where most of the scenario actually sits. Self-replication, indefinite brain support, neural interface bandwidth, and modest temporal acceleration are all unbounded by any known physical law. Only engineering distance separates them from us, though for the write-side interface that distance spans about eight orders of magnitude of channel scaling from implanted hardware. Two constraints in this middle tier are real design pressure, not mere difficulty. The alpha-band sampling cadence near 10 Hz may be a hard floor on subjective acceleration, and post-mitotic neuronal aging almost certainly caps indefinite biological preservation, which forces the late high-dilation arcs onto either rejuvenation or upload.
The one open scientific question is the same hinge the Omega Centauri Society papers turn on, which is the direction of mature intelligence. Cognitive Husbandry and the OCS corpus share a physics, namely the Landauer bound and the thermodynamics of computation, and they share the transcension premise that advanced civilizations move inward toward denser computation rather than outward across space. They differ in where the bottleneck sits. For OCS the limiting resource is an astrophysical entropy sink, a cold place to dump the waste heat of computation, and the papers propose to look for its traces at Omega Centauri. OCS Paper E is the closest methodological sibling to this audit. It runs the same feasibility-then-adjudication move on engineered black-hole infrastructure, deriving an allowed envelope for a fiducial 2×10⁴-solar-mass hole in the Omega Centauri core, forward-modelling a Landauer-limited waste-heat technosignature, the same computation-thermodynamics floor this audit prices minds against in §4.6, and grading a candidate detection against explicit astrophysical nulls; on current data those nulls are still favored. For Cognitive Husbandry the limiting resource is the bandwidth and endurance of a biological receiver, which is why this paper's hardest numbers are about synapses and neurons rather than black holes. The two projects are the same question asked at opposite ends of the same physics.
One last point on the firewall. A scenario can be rigorous in its physics and still rest on premises that no experiment can test, and Cognitive Husbandry does exactly that. Its hardware is mostly permitted or merely-early, its one barred step is repairable, and its deepest commitments, about personal identity and about the goals of an alien mind, are unfalsifiable by construction. None of that counts against it. The audit was never meant to certify the story as likely. It was meant to mark where the physics stops and the philosophy takes over, so the two are not confused for each other.
Drafting, citation verification, derivation checking, and figure preparation were performed with substantial assistance from a large language model (Claude, Anthropic), under the author's direction. The author reviewed and takes full responsibility for all claims, derivations, and references.
Numbered in order of first citation in the text. Each note links back to its mention with ↩.
[Landauer1961]. doi.org/10.1147/rd.53.0183 ↩[Smart2012]. doi.org/10.1016/j.actaastro.2011.11.006 ↩[Vidal2014]. ↩[Chalmers2005, CH-only]. ↩[Bostrom2003b, CH-only]. doi.org/10.1111/1467-9213.00309 ↩[Bostrom2003, CH-only].[Bostrom2014, CH-only].