← explore · constellation · places

1. 01 №
   cognitive

   Every time you recall a memory, the brain reconsolidates it into a slightly different form. Karim Nader's 2000 *Nature* paper (Nader, Schafe, LeDoux) demonstrated this in rats: blocking protein synthesis during recall erases the conditioned memory permanently. The implication is that the act of remembering is the act of rewriting.

   Your earliest memory has been overwritten thousands of times. The "original" is gone. What you remember of being four years old is the most recent revision of a copy of a copy of a copy. The instrument you use to verify your past is the instrument that edits it.

   Source. Nader, Schafe & LeDoux, *Nature* 406:722–726 (2000).

2. 02 №
   deep-time

   Voyager 1 carries a gold-anodised aluminium phonograph record designed by a committee led by Carl Sagan. Cosmic-ray erosion at its current trajectory and shielding will leave the record legible for an estimated one billion years — longer than any structure on Earth, longer than the continents in their current arrangement, longer than the human genus has existed by a factor of roughly five hundred thousand.

   The most durable human artifact is leaving. It is on a trajectory that does not return. Whatever recovers it, if anything, will do so in a galaxy that has rotated several times since the species that made it stopped existing.

   Source. NASA JPL Voyager Mission status; Sagan et al., *Murmurs of Earth* (1978); cosmic-ray erosion estimate, NASA Voyager Interstellar Mission documentation.

3. 03 №
   identity

   The adult human body replaces approximately 330 billion cells per day, with near-complete turnover of most tissues on multi-year cycles (Sender & Milo, *Nature Medicine* 2021). With the partial exception of central nervous system neurons and the lens of the eye, essentially none of the matter that constituted "you" a decade ago is still present.

   The continuity you experience as a single self is a pattern maintained across a substrate that is continuously discarded and rebuilt. The Ship of Theseus is not a thought experiment. It is your biography.

   Source. Sender & Milo, *Nature Medicine* 27:45–48 (2021).

4. 04 №
   cognitive

   Benjamin Libet's 1983 experiments (Libet, Gleason, Wright, Pearl) measured a readiness potential in motor cortex roughly 350–550 ms before subjects reported the conscious decision to move. Soon's 2008 *Nature Neuroscience* study using fMRI pushed the lead time to ~7–10 seconds for some abstract choices.

   The decision is detectable in your brain before you experience making it. The methodology has been critiqued — Schurger 2012 reframes the readiness potential as stochastic accumulation — but the basic temporal asymmetry replicates. The thing you call "I decided" is a downstream report.

   Source. Libet et al., *Brain* 106:623–642 (1983); Soon et al., *Nature Neuroscience* 11:543–545 (2008).

5. 05 №
   embodied

   Microplastic particles have been identified in human placental tissue (Ragusa et al., *Environment International*, 2021), in arterial plaque associated with elevated cardiovascular event risk (Marfella et al., *NEJM*, 2024), and in human testicular tissue (Hu et al., *Toxicological Sciences*, 2024). Detection methods are still maturing; concentrations are non-trivial in every cohort surveyed.

   The barrier between the industrial substrate and the developing fetus is not a barrier. The polymer is now part of the inheritance.

   Source. Ragusa et al., *Environment International* 146:106274 (2021); Marfella et al., *NEJM* 390:900–910 (2024); Hu et al., *Toxicological Sciences* 200:235–240 (2024).

6. 06 №
   identity

   Roger Sperry and Michael Gazzaniga's split-brain work, beginning in the 1960s, documented that severing the corpus callosum produces two distinct loci of conscious report inside one skull, each with its own preferences, knowledge, and capacity for decision. Gazzaniga's "interpreter" finding: the left hemisphere will confabulate a coherent verbal explanation for actions initiated by the right hemisphere, with no awareness that it is doing so.

   A single surgical cut produces two minds. Both are you. Neither knows it. The narrator in your head is, on a continuous basis, generating plausible explanations for behaviour it did not author.

   Source. Sperry, Nobel Lecture (1981); Gazzaniga, *The Bisected Brain* (1970); Gazzaniga, *Annual Review of Neuroscience* 12:323–356 (1989).

7. 07 №
   probabilistic

   The Boltzmann brain problem: if the universe has a non-zero probability per unit spacetime volume of producing a self-aware brain-like fluctuation, and the future is infinite, then the overwhelming statistical majority of observers will be momentary fluctuations in dilute equilibrium rather than biological organisms on planets. This is treated as a serious constraint on cosmological models, not fringe metaphysics (Carroll & Chen 2004; Linde 2007; Boddy, Carroll & Pollack 2014).

   The argument does not require you to believe you are a Boltzmann brain. It requires you to explain why you are not, given the counting. Most published cosmologies treat the inability to do so as a defect in the cosmology, not in the observer.

   Source. Carroll & Chen, arXiv:hep-th/0410270 (2004); Linde, *JCAP* 0701:022 (2007); Boddy, Carroll & Pollack, *Foundations of Physics* 46:702–735 (2016).

8. 08 №
   computational

   Kosinski, Stillwell & Graepel (*PNAS*, 2013) demonstrated that Facebook Likes alone, processed via logistic regression, predict sexual orientation with 88% accuracy, political affiliation (Democrat/Republican) with 85%, and discriminate Black from White respondents with 95% accuracy. The 2013 paper used ~58,000 users and 170 Likes per user on average.

   The thing you considered private — orientation, politics, ethnicity, neuroticism, parental separation — was inferable from pages you chose to associate with publicly, more than a decade ago, with linear methods that are now considered obsolete. The training data are not coming back.

   Source. Kosinski, Stillwell & Graepel, *PNAS* 110(15):5802–5805 (2013).

9. 09 №
   deep-time

   Tyrannosaurus rex lived approximately 68–66 million years ago. Stegosaurus lived approximately 155–145 million years ago. The temporal distance between T. rex and Stegosaurus is roughly 80 million years; the temporal distance between T. rex and the present is roughly 66 million years. You are closer in time to T. rex than T. rex was to Stegosaurus.

   The dinosaurs are not a single epoch. They are an interval longer than the interval that separates them from us, populated by animals separated from each other by chasms our intuition collapses into a single picture.

   Source. Geologic Time Scale 2020 (Gradstein et al., eds.); Brusatte et al., *Biological Reviews* 90:628–642 (2015).

10. 10 №
    embodied

    Mitochondria carry their own genome — a small circular DNA inherited only through the maternal line — that is descended from a free-living alphaproteobacterium endosymbiosed by an archaeal host roughly 1.5–2.0 billion years ago (Sagan/Margulis, endosymbiotic theory; Roger, Muñoz-Gómez & Kamikawa, *Current Biology* 2017). Every aerobic cell in your body is a chimera maintained across a billion and a half years of co-residence.

    You are not a single organism. You are a stable negotiation between two domains of life that predate the existence of complex cells. The energy that lets you read this is generated by a former bacterium your mother handed you.

    Source. Margulis, *Journal of Theoretical Biology* 14:225–274 (1967); Roger, Muñoz-Gómez & Kamikawa, *Current Biology* 27:R1177–R1192 (2017).

11. 11 №
    ecological

    Hallmann et al. (*PLOS ONE*, 2017) reported a 76% mean decline in flying insect biomass across 63 protected German nature reserves over 27 years of standardised Malaise-trap monitoring. The collapse was not driven by changes in habitat, weather, or land use within the reserves themselves.

    A 27-year longitudinal data set inside protected areas, sampled with the same instrument by the same volunteer entomological society, returned three-quarters less mass. The mechanism is not fully identified. The number is.

    Source. Hallmann et al., *PLOS ONE* 12(10):e0185809 (2017).

12. 12 №
    linguistic

    Russell Hurlburt's Descriptive Experience Sampling method, refined over 40 years, finds that inner speech occupies only a fraction of the moments when people are randomly probed mid-thought — roughly 23% on average across studies, with substantial inter-individual variation. A non-trivial minority of subjects report no inner monologue across hundreds of samples.

    The narrator you assumed was universal is not. Some people walk around without the voice you cannot turn off. They are not impaired. They are running a different operating system.

    Source. Hurlburt & Heavey, *Exploring Inner Experience* (2006); Heavey & Hurlburt, *Consciousness and Cognition* 17:798–810 (2008).

13. 13 №
    probabilistic

    The Doomsday Argument (Carter, 1983; Leslie, *The End of the World*, 1996; Gott, *Nature* 363:315–319, 1993) applies Copernican reasoning to one's birth rank within the human species: assuming you are not unusually early or late among all humans who will ever live, Bayesian update on observed birth rank places a 95% confidence upper bound on remaining human population at ~10–20× the cumulative population to date. The argument is contested; it has not been refuted to consensus.

    The argument's premises are minimal: that you are an observer, that observers are typical of their reference class, that ~117 billion humans have lived to date. The conclusion is uncomfortable enough that the philosophical literature is mostly attempts to dissolve it. The attempts have not succeeded.

    Source. Carter, *Philosophical Transactions of the Royal Society A* 310:347–363 (1983); Gott, *Nature* 363:315–319 (1993); Leslie, *The End of the World* (1996).

14. 14 №
    cognitive

    H.M. (Henry Molaison, 1926–2008) underwent bilateral medial temporal lobectomy in 1953 for intractable epilepsy. From the day of surgery until his death 55 years later, he could not form new declarative memories. Clive Wearing, infected with herpes simplex encephalitis in 1985, has a working memory of approximately 7–30 seconds; he greets his wife as if for the first time, repeatedly, every day, and has done so for 40 years.

    Both men remained continuously conscious. Both had intact personalities, language, intelligence. Neither could accumulate a life. The architecture that makes "today" connect to "yesterday" is a specific piece of tissue. Remove it and the present becomes the only room.

    Source. Scoville & Milner, *Journal of Neurology, Neurosurgery & Psychiatry* 20:11–21 (1957); Wilson, Kopelman & Kapur, *Neuropsychological Rehabilitation* 18:527–540 (2008).

15. 15 №
    probabilistic

    Nick Bostrom's 2003 simulation argument (Bostrom, *Philosophical Quarterly* 53:243–255) is a disjunctive claim: at least one of (a) civilisations almost always go extinct before reaching simulation-capable maturity, (b) such civilisations almost never run ancestor simulations, or (c) we are almost certainly inside one. The argument's premises are computational and demographic, not metaphysical.

    The argument does not assert (c). It asserts that denying (c) requires committing to (a) or (b). Most readers concentrate on (c) because (a) and (b) are worse. The structure of the disjunction is the part that does not go away.

    Source. Bostrom, *Philosophical Quarterly* 53(211):243–255 (2003).

16. 16 №
    embodied

    The Hallmark Channel of cellular ageing: Hayflick's 1961 finding that normal somatic cells divide a finite number of times (~40–60 population doublings) before entering replicative senescence. Senescent cells do not die. They secrete pro-inflammatory factors (the senescence-associated secretory phenotype, Coppé et al., 2010) that drive tissue dysfunction in neighbouring cells. Ageing is partly the accumulation of cells that have stopped working and will not leave.

    Death is one mechanism. The other is the cells that quit the cooperative but stay in the building, leaking signals that age the neighbours. The body's failure mode is partly social.

    Source. Hayflick, *Experimental Cell Research* 25:585–621 (1961); Coppé et al., *Annual Review of Pathology* 5:99–118 (2010).

17. 17 №
    cognitive

    Simons & Chabris's 1999 "invisible gorilla" study (*Perception* 28:1059–1074) found that roughly 50% of subjects instructed to count basketball passes failed to notice a person in a gorilla suit walking through the centre of the scene, beating their chest for nine seconds. Change blindness studies (Rensink, O'Regan, Clark, 1997) demonstrate that large changes to a scene across a brief visual interruption routinely go undetected.

    The visual experience you trust as a continuous representation of the world is a sparse sketch with confabulated filler. You see what you are looking for. Most of what is in front of you is reconstructed, not perceived.

    Source. Simons & Chabris, *Perception* 28:1059–1074 (1999); Rensink, O'Regan & Clark, *Psychological Science* 8:368–373 (1997).

18. 18 №
    ecological

    The WWF / ZSL Living Planet Index 2024 reports a 73% average decline in monitored vertebrate populations between 1970 and 2020, across ~35,000 population time-series for ~5,500 species. The figure is an average of relative changes across populations, not a headcount of individuals; the methodology has been critiqued for sensitivity to extreme declines (Leung et al., *Nature* 2020), and the headline trend survives reanalysis.

    A single human lifetime. The reduction is not in remote unmonitored fauna. It is in the populations we chose to monitor because we cared about them.

    Source. WWF / Zoological Society of London, Living Planet Report 2024; Leung et al., *Nature* 588:267–271 (2020).

19. 19 №
    cognitive

    The eukaryotic visual system processes incoming light with a latency of roughly 80–150 ms from photon arrival at the retina to conscious perceptual report (Thorpe et al., *Nature* 381:520–522, 1996; Nijhawan, *Behavioral and Brain Sciences* 31:179–198, 2008). The brain compensates by extrapolating forward — the flash-lag effect, the cutaneous rabbit illusion — so that the present you experience is a predicted reconstruction, not a measurement.

    "Now" is a forecast issued by an organ that knows it is running behind. You have never directly perceived the present.

    Source. Thorpe, Fize & Marlot, *Nature* 381:520–522 (1996); Nijhawan, *Behavioral and Brain Sciences* 31:179–198 (2008).

20. 20 №
    deep-time

    Eighteen percent of mammalian species worldwide are at risk of extinction (IUCN Red List, 2024). Approximately 99% of all species that have ever existed are already extinct (Raup, *Extinction: Bad Genes or Bad Luck?*, 1991; Barnosky et al., *Nature* 471:51–57, 2011). Extinction is the rule. Persistence is the rare condition we mistake for the default.

    Every species you can name occupies a thin sliver of present time. The graveyard is the data set. The living are the sampling artifact.

    Source. IUCN Red List of Threatened Species, 2024 update; Raup, *Extinction* (1991); Barnosky et al., *Nature* 471:51–57 (2011).

21. 21 №
    mortality

    Greenberg, Pyszczynski & Solomon's Terror Management Theory (1986–present) is supported by over 500 published experiments (meta-analysed in Burke, Martens & Faucher, *Personality and Social Psychology Review* 14:155–195, 2010). Brief, subliminal reminders of personal mortality reliably shift subjects toward in-group preference, harsher punishment of out-group transgressors, and stronger endorsement of one's existing worldview. Effect sizes are moderate; the pattern is robust across cultures.

    The political opinions you hold are partially a downstream response to a fact you spend most of the day suppressing. The suppression is the mechanism. When the suppression fails, the opinions sharpen.

    Source. Greenberg, Pyszczynski & Solomon (1986–); Burke, Martens & Faucher, *Personality and Social Psychology Review* 14:155–195 (2010).

22. 22 №
    linguistic

    UNESCO's *Atlas of the World's Languages in Danger* estimates that of approximately 7,000 languages spoken today, between 50% and 90% will be extinct or moribund by 2100. A language goes extinct on average every two weeks. Each carries lexical, grammatical, and cosmological structure not redundantly encoded elsewhere.

    The cognitive variety of the species is collapsing at a measurable rate. The thing being lost is the only known instrument capable of describing what it was.

    Source. Moseley (ed.), UNESCO *Atlas of the World's Languages in Danger*, 3rd ed. (2010); Harrison, *When Languages Die* (2007).

23. 23 №
    embodied

    Sender, Fuchs & Milo's revised estimate (*PLOS Biology*, 2016) puts the ratio of bacterial to human cells in the adult body at approximately 1.3:1 — substantially lower than the long-cited 10:1 figure, but still meaning that the majority of cells inside your skin do not share your genome. The collective microbiome contributes roughly 2–20 million unique genes, compared with ~20,000 in the human genome.

    The genetic system that lives in you is more diverse, by orders of magnitude, than the genetic system that is you. The negotiation between them determines metabolism, mood, and immune response. You are an interface.

    Source. Sender, Fuchs & Milo, *PLOS Biology* 14(8):e1002533 (2016); Tierney et al., *Cell Host & Microbe* 26:283–295 (2019).

24. 24 №
    mortality

    Anaesthesia under propofol or sevoflurane produces a state in which the patient has no subjective report of duration, content, or continuity. There is no internal evidence that the consciousness that resumes is the same consciousness that ended. Anaesthesiologist Stuart Hameroff has framed the point clinically: from the patient's first-person frame, six hours of general anaesthesia is indistinguishable from six hours of being dead. The only operational difference is the resumption.

    You have already experienced (multiple times, if you have had surgery) the subjective state of not existing. The fact that you came back from it is not evidence that the not-existing was different in kind from death. It is evidence that this time, the machinery restarted.

    Source. Mashour & Hudetz, *Anesthesiology* 129:1268–1280 (2018); Hameroff, *Anesthesiology* 105:400–412 (2006).

25. 25 №
    identity

    Wegner & Wheatley's 1999 experiments (*American Psychologist* 54:480–492) demonstrated that the subjective sense of having authored an action can be manipulated independently of actual authorship: subjects routinely report having caused movements they did not cause, and deny having caused movements they did, when the timing cues are altered. The feeling of agency is a separate construction layered on top of motor output.

    The sense that you did a thing on purpose is not a direct read of whether you did. It is an inference your brain makes from timing cues, and it can be tricked. The "I" that takes credit is a narrator with weak access to the floor.

    Source. Wegner & Wheatley, *American Psychologist* 54:480–492 (1999); Wegner, *The Illusion of Conscious Will* (2002).

26. 26 №
    deep-time

    Approximately 95% of the mass-energy content of the universe is not directly observed. Dark matter (~27%) is inferred from gravitational effects on galactic rotation curves (Rubin & Ford, *Astrophysical Journal* 159:379, 1970) and lensing. Dark energy (~68%) is inferred from the accelerating expansion measured via Type Ia supernovae (Riess et al. 1998; Perlmutter et al. 1999, Nobel 2011). Neither has been detected directly in any laboratory experiment. The standard model accounts for ~5% of what is there.

    The physics that describes the matter you can see is a footnote. Twenty-six years after the dark energy detection, the dominant constituent of the cosmos has no candidate particle, no laboratory signature, no agreed mechanism. The cosmos is mostly a placeholder term in an equation that fits the data.

    Source. Rubin & Ford, *ApJ* 159:379 (1970); Riess et al., *AJ* 116:1009 (1998); Perlmutter et al., *ApJ* 517:565 (1999); Planck Collaboration, *A&A* 641:A6 (2020).

27. 27 №
    computational

    Large language models trained on web-scale corpora (Common Crawl, C4, The Pile, RedPajama) include billions of pages of public writing — blog posts, forum threads, code commits, archived personal sites — without per-author consent or opt-out at training time. Bender, Gebru, McMillan-Major & Mitchell (the "stochastic parrots" paper, FAccT 2021) documented that the scale of these corpora makes auditing what is in them effectively impossible. Anything you wrote on the open web before ~2021 is, with high probability, weight in a model somewhere.

    The version of you that exists as text on the public internet has been compiled into instruments you will never see, deployed to generate text in your absence, and operates under no obligation to disclose your contribution. You are training data. You have been for years.

    Source. Bender, Gebru, McMillan-Major & Mitchell, *FAccT 2021*; Dodge et al., *EMNLP 2021* (C4 documentation); Gao et al., arXiv:2101.00027 (The Pile, 2020).

28. 28 №
    mortality

    Hospice nurses and palliative-care physicians report unusually consistent rates of perceptual phenomena in dying patients — visions of deceased relatives, sensed presences, lucid awareness preceding death by hours or days (so-called "terminal lucidity" in dementia patients, documented by Nahm, Greyson, Kelly & Haraldsson, *Archives of Gerontology and Geriatrics* 55:138–142, 2012). Reported rates of end-of-life experiences range from ~50–80% across cohorts; cross-cultural stability is high. Mechanism is unknown. Reporting is subject to selection and confirmation bias.

    The phenomenon is documented across cultures, traditions, and sceptical clinical settings. The clinicians who see it are generally not the ones who arrived predisposed. The fact that no mechanism is identified is not the same as the fact that nothing is happening.

    Source. Nahm, Greyson, Kelly & Haraldsson, *Archives of Gerontology and Geriatrics* 55:138–142 (2012); Fenwick & Brayne, *American Journal of Hospice and Palliative Medicine* 28:7–15 (2011).

29. 29 №
    identity

    The vertebrate central nervous system does not regenerate. Cardiac muscle, hepatic tissue, skin, bone, intestinal epithelium, and haematopoietic stem-cell lineages all sustain ongoing turnover or repair across adult life. Cortical and most spinal neurons do not (Bhardwaj et al., *PNAS* 103:12564–12568, 2006, using ¹⁴C dating from atmospheric weapons-test pulses, found that the vast majority of cortical neurons are as old as the person). The substrate that hosts the self is the one substrate that does not get replaced.

    Every other tissue in you renegotiates its membership. The cells that hold your memories are the same cells you had as a child, ageing alongside you, never substituted. When they go, what they held goes with them.

    Source. Bhardwaj et al., *PNAS* 103(33):12564–12568 (2006); Spalding et al., *Cell* 153:1219–1227 (2013).

30. 30 №
    probabilistic

    The "Great Silence" is not the Fermi paradox in general but its sharpest form: SETI surveys (Breakthrough Listen, 2016–present; Allen Telescope Array; historical Project Phoenix) have now examined thousands of nearby stars across multiple frequency bands without detecting a technosignature. The Drake equation terms most updated by exoplanet science (η_⊕, planet frequency) have moved upward since 1995. The silence is therefore not explained by scarcity of targets — it is a positive observational result against a null model in which technological civilisations are common and durable.

    The hush is data. It rules out specific cosmologies. Either civilisations are rare, short-lived, undetectable, or carefully quiet — and each of those four explanations has an unpleasant implication for the long-run prospects of the one civilisation we can observe in detail.

    Source. Hanson, 'The Great Filter' (1998); Sandberg, Drexler & Ord, 'Dissolving the Fermi Paradox', arXiv:1806.02404 (2018); Price et al., *AJ* 159:86 (2020) (Breakthrough Listen).