Loading
An encyclopedia of futures is mostly an encyclopedia of things not yet known. Every article here is written to say where its own claims run out — which species the result was in, how many people were in the study, which part is argument rather than evidence. Those admissions sit scattered across 229 pages, visible only to whoever happens to be reading that one. This is all of them at once.
Every point in the corpus where an editor stopped to say the evidence is thinner than it looks, or that informed people read it differently.
Showing 223 · drawn from 187 of 229 articles
Every claim about a new capsid's tropism should be read with the question: in which species, and does the receptor it exploits exist in people? Several widely cited enhanced capsids do not work in non-human primates at all.
AAV vectorsExponential curves are easy to find retrospectively because the technologies that failed to accelerate are not the ones anyone plots. Any argument from a set of curves must state how the curves were chosen, and most popular presentations do not.
Accelerating changeSome argue the genetic-divide scenario distracts from the inequality that exists now: a difference of a decade or more in life expectancy between rich and poor districts of the same city, produced by ordinary causes and requiring no new technology. Others reply that a stratification which becomes heritable is categorically different from one that resets each generation, and is worth preventing before it is cheap.
Access and inequalityCardiology supplies the cautionary case. Antiarrhythmic drugs suppressed ventricular ectopic beats, a plausible surrogate for sudden cardiac death, and the Cardiac Arrhythmia Suppression Trial found that they increased mortality. A marker can be strongly associated with an outcome and still be the wrong thing to move.
Aging biomarkersNo regulator recognizes the category, and companies apply the label to everything from a molecule drawn by a generative model to a conventional campaign in which a classifier filtered one plate. The US Food and Drug Administration's 2025 draft guidance addresses AI used to produce evidence submitted in support of a decision, not the provenance of a molecule. There is no registry, no definition, and therefore no denominator against which to count successes.
AI drug discoveryProponents argue the objection misses compounding effects: cheaper cycles mean more targets tested, cleaner selectivity means fewer toxicity failures, and structure prediction opens proteins that had no tractable starting point. Critics answer that none of this has yet changed a phase 2 outcome, that the industry has absorbed several such tool revolutions without moving approval rates, and that the case rests on an argument rather than on a result. Both sides accept that the question is empirically open until a candidate designed this way succeeds or fails in phase 3.
AI drug discoveryConfidence scores estimate how closely a model matches what an experiment would find, not whether the protein does what a paper claims. Predicted structures have already entered the literature as though they were determinations, and a confidently wrong model is harder to catch than an obviously wrong one.
AI protein designThe organization does not assert that its patients can be revived, that current methods preserve a person intact, or that any timeline exists. Its published position is that preservation is worth attempting because the alternative is certain, and because information may survive that present medicine cannot use. Cryobiology societies do not accept even this weaker claim as established.
Alcor Life Extension FoundationPartial reprogramming has extended lifespan in a progeria mouse model, improved tissue repair after injury in normal mice, and restored vision after optic-nerve injury in mice. None of this has been shown in a human. Reprogramming has never been given to a person for an aging indication; the first-in-human intentions companies have stated so far concern eye disease rather than aging itself, and none has published clinical data.
Altos LabsTreated patients in both trials continued to decline: the drugs bought back part of 18 months' worsening, not a plateau. The between-group differences sit below the one-point CDR-SB threshold many clinicians had used for a minimally important difference, and whether the benefit compounds, holds, or fades over five or ten years is unknown; the extension data so far come from sponsors, against external cohorts rather than placebo. Mainstream opinion spans "first disease-modifying success, build on it" to "marginal benefit at real risk and cost", and both positions cite the same trials.
Alzheimer's diseaseProponents point to the causal genetics and to the fact that the only drugs ever to slow the disease are the ones that clear plaques. Critics reply that the slowing is modest, that plaque removal is nearly complete while the clinical effect is not, and that tau tracks symptoms far better. That pattern fits amyloid being an upstream trigger, a partial cause, or a tombstone. The prevention trials now dosing amyloid-positive people before symptoms are the closest thing to a decisive experiment either side has proposed.
Alzheimer's diseaseThe consistency of the harm signal across chemically distinct products is what makes the haemoglobin approach hard to rescue. Any successor has to keep haemoglobin away from endothelial nitric oxide, which in practice means putting it back inside something: a vesicle, a polymer shell or a cell.
Artificial bloodNo timeline for AGI is well supported. Survey medians, scaling extrapolations and insider statements are all weak evidence, and the field's forecasting record, which includes confident predictions of imminence in the 1960s and of impossibility in the 1990s, should discount any specific date, including recent ones.
Artificial general intelligenceNormal growth and myelination in a lamb over four weeks is not evidence of normal human neurodevelopment over the years in which it would be measured. Sheep are precocial, with a shorter gestation and a differently timed brain growth spurt, and the lambs in these studies were euthanised at the end of support rather than followed to adulthood.
Artificial wombSome researchers treat partial support as a stepping stone; others argue the two problems share almost no engineering, and that framing neonatal life support as proto-ectogenesis has attracted attention and opposition the clinical programme did not need.
Artificial wombDe Grey holds that the remaining obstacles are engineering problems with known shapes; most biogerontologists hold that at least three of the seven categories lack any demonstrated repair in a living mammal.
Aubrey de GreyBoth liver programmes are described as single-dose and permanent, and in the cells that take up an edit that is chemically true. The human record behind the durability claim is short: in the published and presented data for VERVE-102 and BEAM-302, about 18 months at most and a year or more in a minority of dosed patients, and nothing yet on the decades over which a cholesterol drug is supposed to work. An edit also cannot be stopped if it turns out to be the wrong trade against a cheap daily tablet, which is the asymmetry that makes the irreversibility of the method a safety question rather than a selling point.
Base editingNone of the studies here implanted an artificial neuron chronically in any animal, let alone a person. The brain-tissue results are acute slice and culture experiments, the in vivo result is an acute experiment in anaesthetized mice, and several "bio-integration" demonstrations used heart cells, epithelial membranes or a plant. Durability is also unmeasured at implant scale: the printed MoS2 neurons were stable for more than a million cycles, which at ten spikes a second is roughly a day of firing.
Bio-interfacing artificial neuronsCritics argue that enhancement changes what it means to be human and should therefore face a presumption against it; proponents reply that the presumption smuggles in the assumption that the current human is well designed, which nobody defends when asked directly.
Bioethics of enhancementAn uncontrolled intervention on one person, with no blinding and an outcome the person is invested in, cannot establish efficacy. The self-experiment tradition in medicine, which includes self-inoculation and self-catheterisation, produced findings because the effects were unmistakable and immediate. Nothing in the enhancement space has that property, so the epistemic value of these demonstrations is close to zero even when the participant is honest.
Biohacking and grindersWhen multiple biological-age metrics are computed on the same cohort, they correlate with each other only weakly, often more weakly than each correlates with chronological age. Belsky and colleagues found this across telomere length, several methylation clocks and clinical composites. Either they are measuring different things, or most of them are measuring noise.
Biological ageMainstream demography disputes the claim that specific regions harbor genuine longevity outliers. Nobody disputes that beans, walking and social connection are good for health. Coverage of the topic routinely treats support for the second claim as support for the first.
Blue ZonesEvery efficacy claim for Blueprint rests on measurements of Johnson alone, with no control group, no blinding and no pre-registered endpoint. Because dozens of drugs, foods and therapies change at once, no result can be attributed to any single one of them, and improvements from a starting point Johnson himself describes as overweight and sleep-deprived would be expected from conventional diet, exercise and sleep alone. The headline measures are surrogates: no epigenetic clock or organ-age score has been validated as a stand-in endpoint showing that an intervention extends healthy life.
Blueprint protocolThe Brain Preservation Prize established that a large mammalian brain can be preserved with synapses visibly intact throughout. It did not establish that memories survive, that the preserved state is sufficient for reconstruction, or that any scanning technology capable of reading a whole human brain is achievable.
Brain preservationNone of these systems transmits meaning. The bit means something only because both participants were told in advance what it would signify. A system with the same architecture and a light bulb in place of magnetic stimulation would work identically, faster, and with less equipment. The neuroscientific content of the demonstrations is that a decoded signal can be delivered as a percept. They do not show that brains can share content.
Brain-to-brain interfacesThe headline results in this field typically come from one to four participants, often the same individuals across multiple papers, with hardware maintained by a research team. Nothing here has been tested at the scale that would establish reliability, and single-participant results have repeatedly failed to generalize.
Brain–computer interfaceSynchron's position is that a device delivered by catheter, with sixteen electrodes, will reach far more patients than one requiring a craniotomy, and that discrete reliable control is enough for most assistive tasks. The counter-position, taken by the penetrating-array developers, is that low-bandwidth control caps the addressable applications at switch-like interaction and rules out speech and dexterous limb control. Both claims are testable and neither has been settled.
Brain–computer interfaceEvery result Johnson reports comes from a single subject with no control, while dozens of interventions change at once and the measurements are selected and published by him or his company. Even a genuine slowdown on an epigenetic clock would not show which intervention caused it, or that it will translate into longer or healthier life: no aging clock is accepted by regulators as a surrogate for those outcomes. Blueprint sells products whose appeal rests partly on these numbers.
Bryan JohnsonNo human study has tested whether caloric restriction extends lifespan, and none realistically can: the trial would run for decades with an intervention almost nobody sustains. Human evidence is limited to surrogate measures over a few years.
Caloric restrictionRestriction gives a large proportional gain in short-lived species and a smaller, contested one in primates. One reading is that it triggers a famine-response programme whose payoff scales with how quickly an organism's reproductive schedule can be deferred; if so, a species that already lives eighty years has little left to gain. The question bears directly on how much weight dietary explanations of exceptional longevity deserve.
Caloric restrictionThe geroscience prediction is that it would, because a younger tissue environment suppresses clonal expansion and a competent immune system removes transformed cells. The counter-argument is that several mechanisms proposed as geroprotective (telomere maintenance, senescence clearance, restored stem-cell proliferation) are the same mechanisms tumours exploit. No human trial has been powered to settle it, and the mouse evidence points both ways depending on the strain and the intervention.
CancerThe causative mutation is untouched. Every red cell a treated patient makes still carries it, and any child they have inherits it exactly as before. Casgevy suppresses the phenotype by reactivating a fetal gene; it does not correct the genome's error, and it changes nothing about inheritance in the way Germline editing would.
CasgevyIt is unsettled whether senescent cells are a primary driver of aging or one downstream consequence among many. Clearance experiments show that they cause specific phenotypes in mice. They do not show that senescence sits upstream of the other hallmarks, and the mouse strains used are not a straightforward model of human aging.
Cellular senescenceA system described as "98 per cent closed" may be quoting water recovery, total mass balance, or oxygen regeneration, and may or may not count food. Because food dominates the unclosed mass in every flown system, headline closure figures that exclude it can differ by an order of magnitude from figures that include it. Comparisons between facilities generally require reading the definitions.
Closed-loop life supportThe medical case for early implantation rests on a developmental window that closes before a child can consent. The cultural case rests on the claim that the intervention alters an identity the child might have chosen. Both are correct about their own premises, which is why the disagreement has not resolved in thirty years.
Cochlear implantEvidence that coffee constituents move markers of aging biology in cultured cells, worms, or mice does not show that the same happens at drinking doses in people, or that it explains the mortality association. And if decaffeinated coffee carries the same association as caffeinated coffee, caffeine-based mechanisms cannot be the whole account even if the association is causal.
Coffee and agingEvery mortality estimate above comes from people who chose how much coffee to drink. Coffee drinkers differ from abstainers in smoking and other habits, and people who become ill or are advised to avoid caffeine may drink less, so illness can lower coffee intake rather than coffee lowering illness. The genetic studies built to sidestep both problems have not found the mortality benefit, and no one has run the randomized trial that would settle the question.
Coffee and agingNo extinct species has been restored. What exists are living animals of extant species carrying a small number of edits derived from an extinct relative's genome. These animals lack the extinct species' full genome, its gut microbiome, its learned behaviour and its ecological context. Whether they are useful is a separate question from whether the name is accurate.
Colossal BiosciencesCompression advocates point to declining disability rates in the late twentieth century. Critics reply that those declines coincided with rising disease prevalence and have since stopped, and that the runners and Adventists who show the clearest compression differ from the general population in ways no policy can replicate.
Compression of morbidityEvery published whole-animal connectome is a snapshot. The fly datasets each come from a single individual fixed at a single moment; the nematode reconstruction was assembled from sections of more than one animal, so it is a composite rather than any one worm's wiring. Variation between individuals of the same species is substantial in mammals and non-trivial even in flies, and synapses turn over on timescales of days.
ConnectomicsA false positive on a cholesterol panel costs a repeat draw. A false positive on a multi-cancer test can commit a well person to imaging, endoscopy and biopsy while the signal's source is hunted, and a true positive for a cancer that would never have caused symptoms commits them to treatment for it. Both the potential benefit and the potential harm are larger than for ordinary panels, and the harms are at present the better characterised of the two.
Consumer blood testingThe strongest case for repeat testing rests on the within-person trend, not the cross-sectional flag. An individual varies around their own set point within a band narrower than the population interval, so a real drift can be visible while every value stays technically normal. Personal reference intervals are a long-standing idea in laboratory medicine, and the same argument underpins Continuous glucose monitoring and Wearable health sensors. What is missing is any trial showing that acting on such trends improves outcomes in healthy people.
Consumer blood testingNo threshold for a post-meal glucose excursion has been validated as a risk factor in a person with normal glucose tolerance. The same meal, eaten by the same person on two days, can produce visibly different curves through sleep, prior activity, stress, sensor site, and simple noise. A movement in the number does not by itself indicate a change in risk, and inferring a health outcome from a curve would require evidence that does not exist yet.
Continuous glucose monitoringThe strongest evidence is in people whose reading changes an insulin dose. In type 2 diabetes managed with diet, Metformin, or a GLP-1 receptor agonist, trials are fewer, effects on HbA1c smaller, and reviewers disagree about whether continuous wear is worth the cost compared with periodic testing. Most of the live clinical argument concerns that boundary, not the wellness market.
Continuous glucose monitoringEach of the cortical prosthesis trials described here has required prior sight: Orion asked for a documented history of useful form vision, ICVP for normal or near-normal vision through at least age ten, and the Spanish trial excludes anyone with under 12 years of good vision. The reason is biological. In people blinded early in life, primary visual cortex becomes active during touch tasks such as Braille reading. A study using magnetic stimulation over the occiput evoked phosphenes in every blind participant with some residual vision but in only 20% of those with none, and found the effect weakest in people without previous visual experience. Whether an implant could give spatial vision to someone born blind has never been tested, so Musk's statement is a forecast rather than a result.
Cortical visual prosthesisClaims that CRISPR is "precise" describe the targeting step, not the outcome. Guide binding is specific; the repair that follows is not, and characterising on-target rearrangements requires long-read or single-cell sequencing that many published studies did not perform.
CRISPR–Cas9Vitrification of small tissue volumes is routine laboratory practice. Long-term storage without chemical change is established physics. Revival of a cryopreserved mammal, recovery of a cryopreserved brain, and repair of ischemic and toxic injury at cellular scale are all undemonstrated, and the last requires technology that does not exist.
CryonicsAlmost no cryobiologist disputes that vitrification suppresses ice, or that −196 °C halts decay. The dispute is about whether the ischemia, toxicity and fracturing that occur in real cases leave enough structure behind to matter, and whether "enough structure" is even a well-defined criterion in the absence of a theory of how memory is physically encoded.
CryonicsKenyon has described cutting sugar and starch from her own diet after work in her laboratory linked glucose to shortened worm lifespan through DAF-16. That inference has not been tested in people. Three decades on, the pathway has yielded human genetics and no human therapy, and the best-supported way to compress late-life decline in humans is still exercise.
Cynthia KenyonKenyon has argued that animals carry a regulatory system that sets the rate of aging and that it can be turned down. Critics read the same experiments as a nutrient-sensing switch into a stress-resistant, low-growth state that happens to postpone death. On that reading the switch is conserved and real, but the claim is a different one. The distinction decides whether "the aging pathway" is a drug target or a metaphor.
Cynthia KenyonResveratrol's lifespan effects have not replicated consistently in mammals. The US National Institute on Aging's Interventions Testing Program, which tests compounds across three genetically heterogeneous mouse cohorts, did not find a lifespan extension for resveratrol, in contrast to its result for Rapamycin.
David SinclairAlmost nobody disputes that NAD+ falls with age or that OSK expression alters epigenetic marks. The dispute is whether these are causes or correlates, and whether a mouse eye result licenses statements about human aging. Sinclair's answer has generally been more confident than his colleagues'.
David SinclairProponents argue that de-extinction develops genetic-rescue tools with immediate use for endangered species, and that the publicity funds conservation science that would otherwise go unfunded. Critics reply that resources spent on proxies could conserve more biodiversity if spent on habitat, and that the promise of reversal weakens the case for prevention.
De-extinctionThe failure was procedural, and it does not show that the technology was implausible in principle. The procedures had no controls and no preclinical foundation, patient selection was driven by the surgeon, institutional incentives favoured a celebrated recruit, and a compassionate-use pathway permitted repeated first-in-human procedures without accumulating evidence. Any regenerative therapy can be run this way, and the safeguards that failed were not specific to tracheas. The structural resemblance to the death of Jesse Gelsinger, which set back Somatic gene therapy by years, and to the He jiankui affair is close: in each case a researcher moved to humans ahead of the evidence, and institutional review either failed or was bypassed.
Decellularized scaffoldsAs of 2026 the only publicly documented births following deliberate genome editing remain the three children of the He Jiankui experiment. Everything else described as a designer baby has involved selecting among embryos, choosing a gamete donor, or, in the distinct case of Mitochondrial replacement therapy, substituting mitochondria rather than editing nuclear DNA.
Designer babiesAnti-doping rules hold athletes responsible for whatever is found in their bodies, regardless of intent. Contaminated or spiked supplements are a recurring source of sanctions, which is why sports bodies point athletes to third-party certification schemes. Those schemes verify identity and screen for contaminants; none of them assesses whether a product works, a distinction routinely blurred in advertising. See Enhancement in sport.
Dietary supplementsA minority position holds that broad-spectrum multivitamins are the one defensible product. It rests on a US trial in male physicians reporting a small reduction in total cancer incidence, and on cognitive substudies of a later cocoa and multivitamin trial reporting modestly slower decline on test batteries in older adults. Critics answer that the effects are small and measured on continuous test scores rather than diagnoses. Defenders answer that the products are cheap and safe and that real-world nutrient intakes are not uniformly adequate. The question is unresolved as of 2026.
Dietary supplementsProponents treat arrival order as a policy variable that funding, regulation and norms can influence. Critics hold that technological development is too coupled for the ordering to be steered: the tools that enable defence usually enable offence, and delaying a capability in one jurisdiction relocates it rather than postponing it.
Differential technological developmentA reconstruction has no memories, only records of memories; no experience, only generated text. Whatever one thinks of Substrate independence, nothing in these systems has the causal relationship to the original brain that even the most permissive theory of Personal identity and continuity requires. Calling them immortality is a category error, not an overstatement of degree.
Digital immortalityNon-disabled people consistently rate life with a given impairment as worse than disabled people rate their own lives, a discrepancy documented in quality-of-life research and known as the disability paradox. Since selection decisions are made by prospective parents who are usually not disabled, and health-economic weightings are derived largely from general-population surveys, the ratings driving both are the ones that adaptation shows to be mistaken.
Disability rights and enhancementThe furthest-advanced construct is DoriVac, a square origami sheet that spaces CpG adjuvant and antigen at controlled distances. In 2026 it matched the approved mRNA vaccines on antibody and T-cell responses in mice and activated dendritic cells in a human lymph-node-on-chip, which is as close to a person as any DNA nanostructure has been. Its mechanism is the TLR9 recognition of bacterial CpG described above, the same property that disqualifies origami as a circulating carrier. A vaccine that works because the immune system notices it therefore says nothing about whether a DNA device can cross a bloodstream intact, which only a first human pharmacokinetic study would establish.
DNA nanotechnologyIf TRIAD reports longer-lived treated dogs, it will be the strongest lifespan result for any drug outside the laboratory, but in a species with its own metabolism, diseases, and dosing window. No rapamycin trial in humans has tested survival or healthspan, and a canine result cannot substitute for one.
Dog Aging ProjectThe costs of restricting research are diffuse and counterfactual, consisting of vaccines not developed and variants not anticipated, and the benefits are counterfactual as well. Neither side can produce the comparison that would settle the question, which is why two decades of argument have not converged.
Dual-use research of concernOpponents of abortion have described ectogenesis as a resolution that satisfies both sides; abortion-rights scholars generally read the same scenario as a mechanism for compelling transfers. The technology cannot settle which reading prevails, because the disagreement was never about whether an alternative existed.
EctogenesisNon-equilibrium thermodynamics places constraints on what physical systems can do. It contains no premise about what anyone ought to do. Every version of the e/acc argument that appears to derive an imperative from entropy production relies on an unstated value claim to get from one to the other.
Effective accelerationismPGT-A reliably measures chromosome copy number in the sampled cells. Whether acting on that measurement produces more babies is a separate claim. It is what patients are paying for, and the randomised evidence does not support it.
Embryo selectionThe 2012 work was biochemistry in a tube. Editing inside human cells was reported months later by other groups, among them one led by George church. Both the credit dispute and the patent litigation turn on whether the tube experiment already contains the invention or whether making it work in a eukaryotic nucleus was the inventive step. No answer is accepted by both sides.
Emmanuelle CharpentierOne camp holds that tacit laboratory skill, working materials, and the difficulty of aerosolisation remain formidable, and points to the RAND null result and to the consistent failure of non-state attempts. The other holds that these barriers are exactly the ones automation, protocol publication, and contract research organisations are dismantling, and that a red-team exercise with a fixed scenario cannot measure a capability that has not been reached yet. The disagreement concerns the shape of a trend rather than any observed event, and evidence is unlikely to settle it before the fact.
Engineered pandemicsThe claim that natural short sleepers suffer no cost rests on small numbers of people, self-report, and limited cognitive testing over short periods. No cohort has been followed for decades with the outcome measures that would settle it: dementia incidence, cardiovascular events, mortality.
Engineered sleep reductionEnhancement advocates read familial short sleep as proof that sleep need is a modifiable set-point with unused headroom. Sleep researchers more often read it as proof that the set-point is under tight genetic control and that a rare variant shifts it slightly, in the way rare variants shift height. On that reading the trait is heritable and not, in any practical sense, engineerable.
Engineered sleep reductionSome argue that if enhancement is coming regardless, a society should provide it universally and cheaply, converting a positional race into a level baseline. Others reply that this simply relocates the race to whatever remains scarce, and that the terminal state of a fully enhanced population is the same ranking with higher costs.
Enhancement arms raceNo epigenetic clock has been shown to track a causal driver of aging rather than a downstream correlate. Methylation changes could be the mechanism, a readout of the mechanism, or an incidental consequence of cell-turnover history, and the regression cannot tell these apart.
Epigenetic clocksNo experiment has separated epigenetic drift as a cause of aging from epigenetic drift as a consequence of it. Reprogramming can improve function and still leave the causal question open, because the intervention that rewrites methylation also changes transcription, metabolism, and chromatin structure at the same time.
Epigenetic reprogrammingErasing an installed mark has been demonstrated in cultured cells and in animals. Whether a clinician could reliably reverse a silencing therapy in a patient, in the tissue where it was delivered, has not been shown.
Epigenome editingThe one formal adjudication came in 2006, when the National Research Council's triennial review of the American nanotechnology programme examined molecular manufacturing and declined to take either side: attainable error rates, speeds and thermodynamic efficiencies of such systems "cannot be reliably predicted at this time" from theory, and the committee recommended funding experimental demonstrations that link the models to results. No demonstration of positional mechanosynthesis of a useful structure has been produced since. The question Smalley raised therefore remains open on the terms the review set: whether an experiment can be made to match the calculations, which already exist.
Eric DrexlerOrd's figures are stated credences from one careful analyst rather than scientific estimates, and they are frequently quoted without that qualification. Their most defensible use is comparative: they say which risks their author thinks deserve the most attention, not how likely the world is to end.
Existential riskOutcome data come from case reports, single-centre series, registries that depend on what centres report, and reviews that fill gaps from press reports and personal communications. Counts also disagree: one review counted seven face graft losses among 46 patients by April 2023, a multicentre study six among 48 by September 2023. The prosthesis comparison cited above involved 12 people. As of 2026 no clinical method regrows or engineers a functioning human hand or face; Limb regeneration and Tissue engineering cover that research, and Osseointegration describes prostheses anchored directly to bone.
Face and hand transplantationFor someone missing a hand, the alternative is a prosthesis with no drug risk. Transplant recipients who had used prostheses have described them as heavy, ill-fitting and difficult to use, and the National Academies noted in 2025 that prostheses do not yet integrate sensation as a hand does, a gap that Neuroprosthetics research aims to close. A Walter Reed representative told the National Academies committee that veterans with amputations may be reluctant to accept the risks of lifelong immunosuppression; the report adds that a transplant disqualifies a service member from deployment. One written submission predicted that improving myoelectric prostheses would, once their cost falls, make hand transplantation an ethical dilemma. The 2016 comparison concluded that transplantation should stay restricted, with people missing both hands the best candidates.
Face and hand transplantationConfident long-horizon technology predictions have been wrong in both directions and at roughly comparable rates. Nuclear energy, space settlement and machine translation were all over-forecast in the mid-twentieth century; the internet, genomic sequencing costs and protein structure prediction were all under-forecast. The lesson is that specific dates carry almost no information, rather than that optimists are wrong.
Future of humanityNo positive test for gene doping in an elite athlete has been publicly reported, more than two decades after the prohibition and several years after the first approved test entered the laboratory guidelines. Whether that reflects absence of the practice or the difficulty of catching it is unresolved, and anti-doping scientists generally argue that both contribute.
Gene dopingBurkina Faso's government terminated all Target Malaria activities on its territory on 22 August 2025, eleven days after a small-scale release of non-gene-drive male-biasing mosquitoes. Nothing molecular failed and no drive was involved: a national authority withdrew consent for a release of organisms engineered so that they could not persist. The staged approach assumes each step earns the next; in the country where it had run furthest, the step that had just succeeded technically ended the programme instead.
Gene drivesA mouse lives under thirty months, weighs thirty grams, and can receive a vector dose per kilogram that is impossible to manufacture for an adult human. Every result in this section is a result about a small, short-lived, inbred animal treated in a specific-pathogen-free facility.
Gene therapy for agingEvery projection about generation ships assumes that human gestation and infant development proceed acceptably in whatever gravity the vessel provides. No mammalian evidence exists at any gravity level between zero and one. A ship providing artificial gravity by rotation avoids the question. A ship without it stakes the mission on an untested developmental biology.
Generation ship biologyYeast is where the two halves of recoding come apart, one of them finished and the other not. The Sc2.0 design changes every TAG stop codon in the genome to TAA, and all sixteen synthetic chromosomes have now been built to it, so the rewriting half exists in a eukaryote. The freeing half does not: eukaryotic release factor eRF1 reads all three stop codons, so it cannot be deleted the way bacterial RF1 was once its codon was cleared, and re-engineering it to ignore TAG remains a proposal. The chromosomes are also still in separate strains. The largest consolidated strain reported carries seven and a half of them, rather more than half the genome, and combining the rest into one cell is the project's stated open problem. Until that strain exists and its eRF1 is altered, the eukaryotic project has printed a new design without yet changing the code.
Genetic code expansion and recodingLow case volume is often cited as showing that genetic discrimination is rare. It is equally consistent with discrimination that cannot be detected: an applicant declined cover is rarely told which item in the file was decisive, and an underwriting model that weights a family history does not announce itself.
Genetic discriminationClaims that cognitive enhancement by embryo selection is now practical rest on population-level score accuracy applied to a within-family decision. Mainstream statistical genetics regards that substitution as invalid, and no company offering the service has published outcome data on selected children.
Genetic enhancement of cognitionEditing a small number of trait-associated genes into a living relative produces a proxy, not a resurrected species. Church has said this himself, but press coverage of Colossal's announcements frequently has not, and the 2025 dire wolf claim drew sharp criticism from paleogeneticists on exactly this point.
George ChurchLong-term dosing produced thyroid C-cell tumours in rats and mice, which is the basis for a boxed warning and a contraindication in people with a personal or family history of medullary thyroid carcinoma. Rodent C-cells express the GLP-1 receptor far more abundantly than human ones do, and no corresponding human effect has been demonstrated. The warning reflects an open question; the finding has not been shown to transfer to humans.
GLP-1 receptor agonistsIf the whole benefit runs through weight and glucose, these are excellent drugs for two diseases and say nothing about aging. If part of it is direct receptor signalling in vessels, kidney, and brain, they belong in the geroprotector conversation. The Alzheimer's failure is evidence for the first reading. Nobody has run the trial that would settle it, and movement in an Epigenetic clock reading or another composite aging measure would not settle it either, because a change in a prediction is not a change in an outcome.
GLP-1 receptor agonistsReduced to a headline, the study reads as a way to flush Alzheimer's plaque from the brain in minutes. It does not show that. The 10-minute effect was faster tracer drainage in healthy mice. The amyloid reduction followed a week of daily sessions in a mouse model and was a fall in staining intensity, not removal; tau was not measured. AQP4 was redistributed towards blood vessels and more abundant, and channel opening was not measured. No person was exposed.
Glymphatic systemProponents point to periarterial tracer influx, which even critical reviewers accept, the AQP4-knockout results, the sleep effect in mice, and human MRI in which an intrathecal tracer reaches the whole brain. Critics reply that solute spread in rodent tissue fits diffusion, that AQP4 dependence has not replicated across laboratories, that a study measuring removal from tissue found the opposite sleep effect, and that influx is not clearance. The two camps use different tracers, injection sites, anaesthetics and definitions, so most experiments cited by one side do not test the claim the other side makes.
Glymphatic systemGrey goo cannot happen without molecular assemblers, and no assembler has been built or shown to be buildable. Every assessment of the scenario is therefore conditional on a technology whose feasibility is itself disputed. That makes the risk unusual, because the debate about its plausibility sits downstream of a debate about chemistry that has never been settled experimentally.
Grey gooThe standard method for turning a microbiome association into a causal claim is to transfer a human community into a germ-free rodent and see whether the phenotype travels with it. A review of that literature found that about 95% of published human-microbiota-associated rodent studies reported the phenotype transferring, a rate the authors argue is implausibly high and better explained by publication and design bias than by the microbiome causing that many human diseases. The method most causal microbiome claims rest on is the one least able to return a negative result.
Gut microbiomeMembership of the list carries no causal weighting. The framework says which processes belong on the list, but not which contribute most to mortality, in what order they act, or which would be worth intervening on first. Several authors argue that this is the framework's central omission, not a detail to be filled in later.
Hallmarks of agingThe 2017 announcement is a case study in how a procedural rehearsal becomes a claimed achievement. Joining tissues between two cadavers demonstrates that the anatomy can be dissected and sutured. It says nothing about circulation, consciousness, immune tolerance or neurological function, which are the entire content of the proposal.
Head transplantationRepeated cold-water immersion immediately after resistance training reduces long-term gains in muscle size relative to active recovery, an effect reported across several controlled studies and pooled analyses; the strength penalty is found less consistently than the hypertrophy one. The likely reason is that cold suppresses the inflammatory and anabolic signalling the training stimulus depends on. An intervention that blunts an adaptive response cannot be assumed to be adaptive itself, which matters wherever recovery practice meets athletic preparation.
Heat and cold exposureMany poikilotherms live longer in the cold, a cold-sensing channel mediates the effect in nematodes, and transgenic mice running a slightly lower core temperature outlived controls without eating less. Sauna epidemiology points the other way. These cannot both be general laws about body temperature, and the field has not reconciled them. The obvious reconciliation is that a transient thermal stress and a sustained thermoregulatory set point are different variables, but no human data distinguish them.
Heat and cold exposureSeveral groups have announced human cloning attempts or successes since 2001, including a religious organisation that claimed a birth in 2002. None provided verifiable evidence. As of 2026 no cloned human birth has been documented.
Human cloningNo published work has produced a model of a whole human body, personalised to an individual, that predicts responses across organ systems. Roadmaps showing one within a decade or two extrapolate from organ-scale successes, and those successes are concentrated where the physics is simple and the anatomy carries the prediction. Cardiology and pharmacokinetics are not a representative sample of physiology.
Human digital twinsOne camp argues that only mechanistic models can answer interventional questions, because a model fitted to observed care cannot say what would happen under care that was never given. The other argues that mechanistic models will always be underdetermined in an individual, and that large learned models trained across millions of records will predict better in practice even if they explain nothing. Hybrids are common in the literature but not yet in anything a regulator has cleared.
Human digital twinsFew participants in this debate hold that enhancement is intrinsically wrong or that it is unconditionally permissible. The live disputes are about who bears the risk, whether consent is meaningful under competitive pressure, and whether the state should fund, permit, discourage or prohibit each modality separately.
Human enhancementNo published work demonstrates accurate, complete, non-mosaic editing of a human embryo verified to the standard that clinical use would require. The 2020 international commission treated establishing such a demonstration as a precondition even to begin discussing translation.
Human germline editingOne camp holds that defining a translational pathway is the responsible response, because a route that is legal, licensed, and monitored is safer than a prohibition that pushes work into unregulated clinics. The other holds that specifying a pathway is itself a form of endorsement that makes eventual use likelier, and points to the Asilomar conference precedent, where a self-imposed moratorium was lifted once containment measures were agreed.
Human germline editingDARPA's Biostasis programme has paid for metabolic-suppression work since 2018, and its published results run through Xenopus tadpoles and then out of the animal entirely. The delta-opioid compound SNC80 cut oxygen consumption in explanted porcine hearts to under half of vehicle control over six hours of perfusion, and reversibly suppressed metabolism in human gut and liver organ chips. No intact large mammal has been placed in a hypometabolic state by any of this, and both the perfused heart and the chip leave out what makes a whole body difficult: a defended set point, a circulation, and a brain. What would settle whether the torpor programme is latent in large mammals is a measurement of suppressed oxygen consumption, rather than lowered temperature alone, in a living pig or non-human primate.
Human hibernation and torporProponents treat the interface as the limiting factor and the brain as an underused source. The competing view holds that conscious cognition is intrinsically low-rate, in which case a thousand-channel implant would deliver the same ten bits per second with more surgery. No experiment has yet distinguished these, because no interface has come close to saturating even the narrow channel.
Human–AI mergerThe word "immunosenescence" implies degeneration, and the field is not unanimous that this is the right description. On the alternative reading the aged immune system has been remodelled, not broken. It is optimised for the antigens it has already met at the expense of those it has not, a reasonable allocation for an organism past reproduction. Cytomegalovirus gives the argument its sharpest form: the virus correlates with mortality in Swedish cohorts but not in some other populations, which suggests its effect depends on context and is not a fixed cost.
ImmunosenescenceA clinic's headline rate depends on which patients it accepts and how it counts. Rates per embryo transferred exclude cycles cancelled before transfer; rates per cycle started include them. Neither figure describes an individual's chance, which is dominated by age and by cause of infertility.
In vitro fertilisationEvery complete demonstration of in vitro gametogenesis is in mice. Reports described in press coverage as human breakthroughs have concerned earlier stages of the pathway, ovarian support cells used to improve conventional egg maturation, or nuclear-transfer approaches that are not gametogenesis from pluripotent cells at all.
In vitro gametogenesisThe point the two published cohorts agree on is that an in-body CAR can be made and can clear marrow disease. The point they do not settle is whether it lasts. In the cohort with the longest follow-up, median progression-free survival was 4.0 months and three of four patients died after extramedullary relapse; the authors conclude that the limited persistence they observed means the platform needs further optimisation. Nine patients across two trials cannot distinguish a platform problem from a disease problem.
In vivo CAR-T therapySeveral widely reported "stem cell" trials use embryonic stem cells rather than iPSCs, including prominent Parkinson's and type 1 diabetes programmes. The two behave similarly in the dish but differ entirely in supply chain, ethics and immune matching. Coverage frequently conflates them.
Induced pluripotent stem cellsThe mechanistic case for inflammaging as a driver is strong in mice. In humans it rests largely on association, and inflammatory markers are downstream of nearly everything that goes wrong with an aging body. Sceptics argue that IL-6 is an excellent prognostic indicator because it integrates many forms of damage, which is a different claim from saying that lowering it would help.
InflammagingIsomorphic's published evidence is a peer-reviewed structure-prediction model and benchmark results reported by the company for a proprietary engine. No Isomorphic molecule has entered a clinical trial, the company has not publicly named a development candidate, and the partners' milestone payments depend on results not yet achieved.
Isomorphic LabsPublic credit for CRISPR is contested well beyond Doudna and Charpentier. Francisco Mojica, Rodolphe Barrangou, Philippe Horvath, Virginijus Šikšnys and Feng Zhang all made claims to foundational contributions. The 2020 Nobel recognized two people for a discovery with a long chain of contributors, a familiar structural problem with the prize.
Jennifer DoudnaDisability-rights scholars argue that selecting against a trait expresses a judgement about people living with it. Savulescu's reply relies on a welfarist account of disability, in which a condition is a disability if it reduces expected wellbeing in the circumstances, not because it deviates from a norm. Critics say this relocates the judgement rather than removing it. See Disability rights and enhancement.
Julian SavulescuEvery clinical claim for Flow rests on studies funded by Kernel and written mainly by its employees. The headline cognitive-impairment result comes from one cross-validated cohort of 101 people with no external test set, and none of Kernel's published papers validates the brain-age score sold in clinics. The open question is not whether the helmet records a real haemodynamic signal but whether the classifiers hold up in independent cohorts.
Kernel (neurotechnology company)Every human klotho finding is a correlation. Low serum klotho in an unhealthy person is at least partly a consequence of reduced kidney function, so the mortality association is compatible with klotho being a marker rather than a driver. Assay method matters too: immunoprecipitation-based measurement tracks kidney function considerably better than the commercial ELISAs most cohort studies use, which complicates comparison across papers. Nothing short of administering klotho and measuring an outcome will separate the readings, and that experiment has not been reported in people.
KlothoKlotho-null mice retain phosphate to a degree that is lethal on its own. Crossing them with mice lacking the renal phosphate transporter NaPi2a lowered serum phosphate, restored fertility and body weight, suppressed ectopic calcification and prolonged survival; feeding those double-knockouts a high-phosphate diet brought the premature-aging features back. On that reading the 1997 phenotype is severe mineral disease that mimics old age, not accelerated aging. Proponents of the longevity interpretation reply that phosphate toxicity is itself one of the mechanisms by which tissues age, which keeps klotho central without requiring it to act through any separate anti-aging pathway.
KlothoThe bladder claim rests on seven patients reported in 2006. When the same autologous cell-seeded scaffold was carried into a phase 2 study in children and adolescents with spina bifida, it improved neither bladder compliance nor capacity at 12 or 36 months, and serious adverse events of bowel obstruction or bladder rupture occurred in four participants; the investigators concluded that the serious adverse events surpassed an acceptable safety standard. Engineered airway grafts, the other line of this work to reach patients, ended in retracted papers and institutional findings of scientific misconduct. Neither failure was a vascularization failure, which is the obstacle usually named: one construct did not reproduce its own earlier case series, and the other reached patients ahead of the animal evidence that would have predicted the outcome.
Lab-grown organsRepugnance has historically attached to interracial marriage, homosexuality, dissection and vaccination. Critics including Martha Nussbaum, Arthur Caplan and Steven Pinker argue that disgust is a poor moral instrument because it attaches so easily to the unfamiliar, and that Kass supplies no criterion for distinguishing wise repugnance from prejudice. Kass replies that repugnance is evidence to be examined rather than a verdict, and that admitting only what can be formalized would arbitrarily discard most of moral experience.
Leon KassThe limit is a property of cells in culture. Most somatic cells in a living body never approach their replicative capacity, and senescent cells accumulating in aged tissue arise largely through stress and damage rather than through counting divisions. That the limit exists is not in dispute. The claim that it sets the length of a human life is much larger, and the evidence for it is indirect.
Leonard HayflickOne camp holds that mammals retain the genetic machinery for regeneration and merely fail to activate it, so the problem is one of triggering. Another holds that the salamander programme depends on features mammals do not have, including specific gene families expanded in the axolotl and a wound response that does not fibrose; on that view the problem is one of construction rather than triggering. The fingertip result supports the first view weakly; the failure of every attempt at larger scale supports the second.
Limb regenerationIt does not mean an individual stops dying, that lifespan becomes unbounded, or that any therapy has been demonstrated. A population could be at escape velocity while its members still died of accidents, infection, and residual disease, and any individual could miss the threshold by being too old when it arrives.
Longevity escape velocityAlmost no biogerontologist disputes that escape velocity is arithmetically coherent, or that it would follow from sufficiently rapid rejuvenation. The dispute is entirely over whether the required rate is remotely achievable, and whether de Grey's damage taxonomy is complete enough to serve as a plan and not just a sketch.
Longevity escape velocityEvery claim in this area concerns dogs, and as of November 2026 no drug had been shown in a controlled trial to extend dogs' lives. If STAY reported a survival benefit, it would establish that a drug extended lifespan in one population of senior companion dogs, not that it restored youth; a human programme would still have to begin again at first-in-human safety, against an indication no regulator accepts. The natural human experiment on the IGF-1 axis is instructive: people with growth hormone receptor deficiency have strikingly little cancer and diabetes, without a demonstrated increase in lifespan.
LoyalProponents argue that a positive canine survival result would be the first controlled demonstration that a drug slows aging in a large outbred mammal sharing the human environment, and would pressure regulators to define a human aging endpoint. Skeptics reply that dogs metabolize drugs differently, and that the human bottleneck was never a lack of evidence from another species: the problem is that "aging" has no legal status as a condition to treat.
LoyalOn a functional account such as global workspace theory, an architecture with the right selection and broadcast structure qualifies whatever it is made of. On integrated information theory the physical organization of the hardware decides, and conventional computers have the wrong organization. So the same system is conscious on one leading theory and not on another, with no experiment between them. This is the dispute set out under Substrate independence.
Machine consciousnessEvery placebo-controlled trial of cognition or sleep was paid for by a company with rights to the ingredient: Neurocentria, Wu-xi Magtein Bioscience, AIDP or Threotech. No independent group has published a replication of the cognitive effect, and the brain-uptake mechanism that distinguishes MgT from cheaper salts has been shown only in rats. The largest trial's primary result sits just under the significance threshold, which its methods set as one-tailed, with no correction reported for the many outcomes tested.
Magnesium L-threonateThe "limit" position holds that biology truncates the tail: something goes wrong at an age no current intervention touches. The "no limit" position holds only that the data cannot distinguish a truncated tail from a very thin one, not that people will live to 150. Almost nobody argues that current medicine can produce a 130-year-old.
Maximum human lifespanA microrobot carries a very small amount of drug. Delivering a clinically meaningful dose requires either enormous numbers of devices or a payload so potent that leakage would be dangerous. Critics argue that for most indications a well-formulated particle, as covered in Targeted drug delivery, reaches the same tissue at a fraction of the complexity, and that Lipid nanoparticles have already carried nucleic acid cargoes into human cells at industrial scale without moving at all. The counterargument is that microrobots do things formulations cannot: apply mechanical force, move against flow, and be positioned rather than distributed.
Medical microrobotsWhether atomically precise diamondoid machinery can be built at all is unresolved rather than settled in either direction. The mainstream chemistry community largely stopped engaging with the question after the early 2000s; proponents read that as neglect, critics as a verdict. What is not contested is that nobody has built one.
Medical nanorobotsNo study has restored a memory that was lost, improved memory outside a laboratory task, produced a benefit lasting beyond the testing session, or demonstrated anything in a person with a chronic implanted device. Every human result comes from electrodes placed for a different clinical purpose and removed within days or weeks.
Memory prosthesisDecoders output the most probable interpretation under a trained model, which is a different thing from a fact about the mind. A confident reconstruction from a mismatched model is confidently wrong, and the subject has no way to contest it. The same failure mode discredited earlier physiological lie detection.
Mental privacyTwo 2019 trials in older adults found that metformin blunted the mitochondrial adaptation to aerobic training and the hypertrophic response to resistance training. If that result generalizes, a healthy older adult taking metformin may be trading away part of the benefit of the best-evidenced geroprotector available for an unproven one.
Metformin and the TAME trialVery few neuroscientists claim that brains do something physically uncomputable, and the mainstream objection is not vitalist. David Chalmers, who defends the functionalist premise the whole proposal rests on, still treats human uploading as a technology for some distant century rather than a foreseeable one. The disagreement is about recoverable parameters, not about whether brains are physical.
Mind uploadingSome researchers argue that publicising a hazard nobody was pursuing creates interest in it, and that the risk assessment is speculative enough that a moratorium chills legitimate chemistry. The authors' reply is that the technical barriers give an unusually long lead time, and that a governance regime is far easier to build before anyone has invested a career in the work than after.
Mirror lifeThe donor contributes 37 genes out of roughly 20,000, none of them influencing the traits people associate with parenthood, so the phrase overstates the biological relationship. It understates the legal point, though. The modification is heritable, because a girl born after the procedure passes the donor mitochondria to her own children, and that is why it required primary legislation.
Mitochondrial replacement therapyNeither party demonstrated a tooltip, attempted a mechanosynthetic reaction, or specified a decisive experiment. Chemists have generally treated Smalley's objections as sufficient reason not to pursue the idea; proponents treat the absence of an experimental refutation as leaving the question open. Both positions are compatible with the evidence, which is thin on the specific question of programmed covalent transfer under positional control.
Molecular assemblerThe proposal concerns preventing catastrophic harm by a small number of highly motivated actors. The evidence concerns millisecond-scale changes in association-test scores among undergraduates. No argument has been offered that bridges the two, and the distance is not a matter of degree.
Moral enhancementThe sharpest bioconservative statement denies the relevance of the autonomy argument instead of contesting it. Michael Sandel argues that the vocabulary of autonomy and rights is "ill equipped" for enhancement precisely because it "cannot explain our moral hesitation about people who seek genetic remedies or enhancements for themselves"; his objection is to the "drive to mastery" a choice expresses, not to who makes it. Against Sandberg's claim that the decision belongs to the person in the body, Sandel's reply is that the decision's location was never the question. The two disagree about whether consent exhausts the ethics, and no refinement of consent procedures can address that.
Morphological freedomFollistatin gene-transfer products sold outside regulatory oversight have not been through controlled trials, have no published safety follow-up, and use vectors whose expression cannot be switched off. The published human data on gene transfer with AAV include serious immune and hepatic events at high systemic doses.
Myostatin inhibition"NAD+ falls with age" is repeated as settled. NAD+ degrades within minutes of tissue collection, making measurement difficult, and while declines are well documented in aged mouse tissues, human data are mixed, with several studies failing to find an age-related fall in skeletal muscle. The premise of the whole intervention is less secure than its marketing implies.
NAD+ precursorsThe preprint of the 2023 study was called "Increased hyaluronan by naked mole-rat HAS2 extends lifespan in mice". The version published in Nature is called "improves healthspan in mice". The lifespan numbers are in both. The framing changed in review, a fair signal of how much weight a 4% median gain in one mouse strain will carry.
Naked mole-ratFinding cancer earlier does not by itself help patients. Lead-time bias makes survival from diagnosis look longer whenever diagnosis moves earlier, even if death occurs at the same moment. Length bias means screening preferentially catches slow-growing tumours, which are the ones least likely to kill. Only a reduction in mortality, measured against an unscreened control arm, distinguishes benefit from statistical artefact.
Nanoscale diagnosticsNo intervention has been shown to give people a flat mortality curve, and current biological-age measures could not demonstrate one if it existed. A 2023 consensus framework from the Biomarkers of Aging Consortium stated that "there are not yet any formally validated or likely surrogate endpoint biomarkers of aging" and that no aging biomarker had been approved by US regulators for clinical use. Claims that a person can be kept at a Biological age of 25-35 rest on animal studies and surrogate markers such as the Epigenetic clock, not on human outcomes.
Negligible senescenceIn 1998 Christof Koch wagered David Chalmers a case of wine that a clear neural correlate of consciousness would be identified within 25 years. Koch conceded in 2023. The concession was collegial and the science had advanced considerably, but a leading experimentalist had judged the specific target unmet.
Neural correlates of consciousness"Neural dust" invites comparison with the devices imagined under Medical nanorobots, which are proposed to be a thousand times smaller and to move under their own power. The demonstrated motes are passive millimetre-scale electronics, closer in kind to a very small pacemaker than to anything nanoscale. The mismatch between the name and the hardware has repeatedly led to the work being reported as more advanced than it is.
Neural dust and ultrasonic implantsParticipant counts, performance figures, and adverse-event descriptions for Neuralink's studies have come predominantly from company statements, livestreams and social media. In October 2025 Bloomberg reported that a paper on the first three PRIME participants had been submitted to the New England Journal of Medicine. As of October 2026 no peer-reviewed clinical report from any Neuralink trial has been published, and the speech results exist only as a conference abstract, so durability, complication rates, and decoder performance cannot yet be assessed independently.
NeuralinkAdvocates argue that neural data is categorically different because it is generated involuntarily and can reveal states the subject has not chosen to express. Critics reply that the same is true of a consumer heart-rate monitor, and that treating the brain as special repeats the error diagnosed in genetic exceptionalism (see Genetic discrimination).
NeurorightsNone of the variants identified in centenarian cohorts has been shown to cause exceptional longevity, and no drug derived from one has extended human healthy life. The cohorts also carry a survivorship problem: people who reach 95 are, by construction, unrepresentative in ways no covariate captures.
Nir BarzilaiCritics generally accept that a multi-disease endpoint is what geroscience needs and doubt that metformin should carry it. Their argument is that a null result would be read as evidence about aging as an indication rather than about one weak biguanide, so a cheap drug with thin preclinical support puts the precedent at risk. Barzilai's answer has been that metformin was chosen for a safety record accumulated in millions of older people and for a price that would make a positive result usable everywhere.
Nir BarzilaiA typical enhancement finding is a within-session improvement of a few percent on one task, in around twenty people, relative to a sham the participants may have identified. It is not evidence that a device makes anyone smarter, and it does not transfer to untrained tasks.
Non-invasive neuromodulationIn the United States, Dietary supplements are not reviewed for efficacy before sale, and enforcement is post-market. A compound withdrawn from a supplement after a warning letter can reappear under a different name. Consumers of this market are not protected by the evidence standards that apply to the prescription drugs in the same category.
NootropicsOptogenetics transformed animal neuroscience within a decade, yet after twenty years its clinical footprint is one published case report in a rare form of blindness. The mechanism is not in doubt. The obstacle is that every therapeutic use requires solving gene delivery and light delivery at the same time, in a specific tissue, with acceptable immunology.
OptogeneticsPress coverage of bioprinting routinely reports the geometry as though it were the achievement. Printing the shape of an organ is a solved problem, and a desktop machine can do it in a day. The unsolved problem, and one that is not close to being solved, is producing tissue with the correct cell types, in the correct densities, arranged around a perfused vascular tree, to a mechanical and metabolic standard that keeps a human alive.
Organ bioprintingProponents of regulated compensation argue that the current system already pays everyone in the operating room except the person supplying the organ, and that a legal market with price floors and follow-up care would be safer than the black market that exists anyway. Opponents argue that payment converts a gift relationship into a transaction that will fall hardest on the poor, that valuation of body parts is corrosive regardless of safeguards, and that the Iranian experience shows exploitation rather than a solution. The disagreement is primarily about values rather than about the elasticity of supply.
Organ shortageBone-to-metal integration is reliable; skin-to-metal integration is not. Every proposal for permanently percutaneous hardware, whether limb anchors, transcutaneous power leads or the skull pedestals used in academic Brain computer interface research, inherits the same unsolved problem, and it is a principal reason implanted systems are pushed toward fully wireless designs.
OsseointegrationAzoulay and colleagues caution that the welfare implications of their result are ambiguous: gatekeeping by established scientists may help a young field build on shared assumptions, and their evidence covers only the academic life sciences. Whether institutional renewal depends on mortality itself, or only on the mobility that mortality happens to produce, remains open. No human population in which senior figures stop aging exists to test it.
Overpopulation and life extensionMost enhancement debates assume the modified person benefits. Pantropy assumes a colony benefits from having members suited to it. Those interests coincide only if the person wants to stay. The disability-rights literature summarised in Disability rights and enhancement has long argued that framing a body as suited or unsuited to an environment obscures who set the environment, and that argument transfers directly.
PantropyHumans already occupy Antarctica, the deep ocean and orbit without any genetic change, using suits, habitats and supply chains. Critics argue this is the decisive precedent: engineering an environment is fast, reversible, and improves with each iteration, while engineering a genome is slow, irreversible in a population, and improves only across generations. Pantropists reply that habitat dependence is a permanent tax of its own, and that a population which can never step outside amounts to a permanently sustained expedition rather than a settlement.
PantropyThe intuitive reading of parabiosis is that young blood supplies something restorative. The alternative is that old blood contains inhibitory factors and the young partner simply dilutes them. Replacing about half the plasma of old mice with saline and albumin, with no young blood at all, produced improvements comparable to parabiosis in several tissues. If dilution is the mechanism, the therapeutic target is removal, not transfusion.
Parabiosis and young bloodBoth bounds above are set by mice. The first human exposure to a partial reprogramming therapy, OSK for optic neuropathies, dosed its first participant in June 2026 in a phase 1 safety study confined to the eye. A study of that size and design can detect an inflamed retina. It cannot address the failure mode that decides whether the technique is usable at all, a single over-reprogrammed cell seeding a tumour years later, and only multi-year follow-up in a treated human compartment will move that question.
Partial reprogrammingIf Parfit is right, an upload that inherits full psychological continuity carries everything worth caring about, and its being a copy is not a further loss. If he is wrong and identity requires something more, such as a continuously running physical process, then destructive scanning is death with a convincing survivor. The engineering cannot decide between these. See Teleportation problem.
Personal identity and continuityThe August 2026 release states that two endpoints were met and supplies no numbers. Recurrence-free survival is not survival, and the companies say the trial continues in order to evaluate the latter. The nearest published comparison is unhelpful in a specific way: at five years the phase 2b's exploratory overall-survival hazard ratio was 0.471, but on fourteen events, with a confidence interval running from 0.165 to 1.345. That interval is consistent with a large benefit and with none.
Personalized mRNA cancer vaccinesMelanoma carries one of the highest mutation burdens of any common cancer, meaning more candidate neoantigens per tumour, and already responds to checkpoint blockade better than most. Proponents argue the mechanism is antigen-agnostic and should transfer wherever there are enough mutations to choose from. Sceptics answer that it is being tested first where the immune system has the most to work with, and that the tumours responsible for most cancer deaths are the low-mutation ones where it has the least. The two positions agree about melanoma and disagree about lung and kidney cancer, which is why those readouts matter more than the one just delivered.
Personalized mRNA cancer vaccinesA report telling a couple that one embryo carries, say, sixty per cent less risk of schizophrenia than another describes a shift in a probability that was already low, computed from a model that has never been validated against the outcome of a selected birth. Proponents argue that a real if small expected benefit is still a benefit. Critics reply that the reported percentages imply a precision the underlying scores do not have.
Polygenic embryo screeningBostrom's definition implies a crossing point, since a capacity either exceeds the current human maximum or does not. Critics inside transhumanism note that capacities are continuous and technology-relative: literacy, spectacles and antibiotics have all raised human maxima, so the reference class keeps moving and the threshold recedes as it is approached.
PosthumanWhether Sunstein's paralysis argument refutes the principle or only a caricature of it is itself disputed. Daniel Steel's book-length defence recasts the weak principle as a meta-rule (scientific uncertainty must not by itself justify inaction) and binds stronger versions to consistency and efficiency constraints under which a precaution cannot be both recommended and forbidden by the same rule. Reviewers of that repair answer that permitting different versions in different contexts revives the incoherence objection one level up. What divides the camps is whether the principle is a decision rule at all or a constraint on other rules. No regulatory outcome can settle that, because each side classifies the same outcome differently.
Precautionary principleThe entire clinical evidence for prime editing is one trial with two participants, whose blood stem cells were edited outside the body and returned after busulfan conditioning; restored NADPH oxidase activity was maintained to six and four months at the last reported follow-up. No prime editor has been delivered inside a human body. Every in vivo correction is still in mice, in liver and a small number of other tissues. What would settle the platform's case is durability measured in years and a second indication, and as of 2026 neither exists.
Prime editingDefenders argue that a small expected benefit is still a benefit, and that the principle does not require large effects to hold. Opponents reply that a principle whose real-world application yields an uncertain fraction of a standard deviation, while restructuring how parents relate to their children, is a poor trade, and that the moral cost is incurred whether or not the prediction pays off.
Procreative beneficenceNo randomised controlled trial has ever been run in progeria, and with roughly 160 identified patients worldwide one is unlikely ever to be run. Every survival estimate for lonafarnib compares treated children with untreated children identified separately, a design that cannot exclude differences in ascertainment, diagnosis date, or standard of care. The effect is the best-supported claim in the field, and it still rests on no randomised comparison.
ProgeriaOne camp treats the syndrome as normal aging compressed: it shares genome instability, telomere attrition, premature senescence and stem-cell depletion with ordinary aging, and progerin itself is made at low levels in unaffected people through sporadic use of the same splice site. The other treats it as a segmental disease that copies some of aging's output through a mechanism aging does not use: no raised cancer risk, no cognitive decline, and a distinctive vascular lesion rather than the usual atherosclerotic one. Nothing in the genetics settles which reading is right, and the therapeutic stakes differ. On the first reading a progeria cure is a lead on aging; on the second it is only a cure for progeria.
ProgeriaThe 2024 rejection was not a safety refusal. The advisory committee's objections centred on functional unblinding, on the impossibility of separating the drug from an unstandardized psychotherapy the agency does not regulate, and on missing data about durability and abuse potential. In the same month the journal Psychopharmacology retracted three papers reporting earlier MDMA trial data, citing undisclosed unethical conduct at one site.
Psychedelic therapyOne camp holds that the acute experience is the therapy, that its intensity predicts outcome, and that a non-hallucinogenic analogue would be a different drug. The other holds that the experience is a side effect of engaging a plasticity mechanism, and that the intensity-outcome correlation is what expectancy alone would produce. No human trial has separated them.
Psychedelic therapyTwo large randomised evaluations of workplace wellness programmes converged on the same answer. In a trial at a US retailer, employees offered the programme reported more regular exercise and more active weight management at 18 months, while clinical measures, health spending and employment outcomes did not differ significantly from controls. The Illinois study raised screening rates but found no significant effect on medical spending, other health behaviours or productivity, and it added a result that complicates every observational claim here: employees who chose to enrol were already healthier and lower-spending before the programme began, so such schemes look effective in uncontrolled data largely through selection. Advocates answer that the interventions tested were weak. That is a fair point about those trials, but it does not show that a stronger programme works.
Quantified selfFaster propulsion reduces integrated dose linearly and requires no biology. Critics of biological hardening argue that any resource spent on editing humans would buy more risk reduction spent on propulsion, shielding mass, or mission architecture. Proponents reply that transit time has a floor set by orbital mechanics and that permanent settlement, unlike a mission, has no exposure endpoint at all.
Radiation tolerance in humansA minority of physicians prescribe rapamycin off-label for longevity, and a self-selected community of users takes it. In a 2023 survey of 333 adults with a history of off-label use, about nine in ten current users took it once a week, and mouth ulceration was the only condition reported significantly more often than by 172 non-users. The authors themselves note that self-reported data from a possibly self-selected group are less reliable than a randomized trial. No dose, schedule or monitoring standard has been validated for healthy adults.
RapamycinPopular claims that drugs such as rapamycin can hold people at a fixed biological age go well beyond the evidence. In mice, the best-studied mammal, rapamycin lengthens life, but treated mice still die of the same pattern of diseases as untreated ones, and in one broad study it improved few of the aging traits measured. The one dog lifespan trial has not reported, no primate survival result has been peer-reviewed, the marmoset clock analysis detected no significant change in epigenetic age, and no human trial has measured aging or lifespan as an outcome.
RapamycinThe supplement regimen at the base of the argument is not backed by clinical evidence for life extension in humans. Trials of Nad precursors, Metformin and antioxidant supplementation have not demonstrated an effect on human lifespan, and exercise remains the only well-evidenced geroprotector.
Ray KurzweilEvery intervention above is measured against surrogates: anti-Müllerian hormone, antral follicle count, oocyte yield. None of these has been validated as predicting the outcomes people care about: a healthy birth or a later menopause. The obstacle is the one described in Aging biomarkers, and the correlation-versus-causation trap is the one that limits the Epigenetic clock. It is arguably more acute here, because a definitive trial would have to run for a decade.
Reproductive longevityProponents treat the respirocyte as an engineering target awaiting a fabrication technology. Most biomedical engineers treat it as an illustration of what atomically precise manufacturing would permit, with no bearing on what can be attempted now. Both readings accept that the device cannot currently be built; they differ on whether the design constitutes progress toward building it.
RespirocytesLong-term safety beyond two years is unknown, and the cardiovascular question is open: TRIUMPH-3's hazard ratio for heart attack, stroke and cardiovascular death was 1.12 with a confidence interval from 0.64 to 1.96. Whether the pulse increase seen in phase 2 and again in the kidney trial matters clinically is a question only longer trials can answer. The main open questions rest on TRIUMPH-Outcomes, whose primary completion is estimated for February 2029.
RetatrutideReported successes are usually object localization, motion detection and letter identification under favourable conditions. Users typically keep their cane and continue to rely on other senses. Press coverage that describes patients as "seeing again" overstates every result published to date.
Retinal implants and visual prosthesesEvery mechanism Retro works on rests on rodent evidence. Autophagy enhancement, partial reprogramming and heterochronic blood exchange have all produced functional improvements in mice. None has been shown to extend healthy life in humans, and the company's ten-year target is a goal rather than a projection supported by data.
Retro BiosciencesProponents argue that life extension adds an option and removes none: anyone may still decline. Critics reply that options with strong social defaults are not neutral, and that the pressure not to burden others, which currently pushes toward accepting death, would push toward accepting treatment with equal force and less scrutiny.
Right to die and the duty to liveSupporters hold that a quantitative design that survives physical scrutiny is a stronger claim than an assertion that revival is impossible, and that the burden has not been met on the other side. Critics hold that an analysis of surfaces, tissues and failure modes that have never been produced cannot be checked at all, and that engineering which cannot be falsified does not count as engineering.
Robert FreitasThe best-documented results come from briefly trained judges on a flat competition course, and from a user study of 35 people aged 20-34 in rooms about 30 m apart. None of the evaluations cited here measures accuracy, error rates or operator fatigue for avatar robots doing real hazardous work over full shifts. Company accounts are unreviewed, and Mundane's comes without performance numbers.
Robotic avatarsThe competition report's authors wrote that it remains open how much human-likeness avatars need or should have, while noting that head movement, gaze, gestures and an animated face helped recipients perceive the operator as present. The joint analysis found that simpler robots often performed reasonably well, perhaps because they were easier to control, and that the more complex humanoid systems were at elevated risk of catastrophic failure. Neither report settles how human-like an avatar should be.
Robotic avatarsEWGSOP, EWGSOP2, the Asian and international working groups, and the US FNIH project draw different cut-offs on different measures, and prevalence in the same population can shift several-fold depending on which is applied. Trials recruited under one definition cannot be cleanly compared with trials recruited under another, and a regulator asked to approve a sarcopenia drug has no consensus endpoint to approve it against. The 2024 GLIS consensus settled the concept (muscle mass plus strength, with function relegated to an outcome) but deliberately deferred the operational cut-offs, so the harmonisation the pipeline needs is still pending.
SarcopeniaNo validated assay reports senescent-cell burden in a living human across tissues. Without one, trials cannot confirm that a drug engaged its target in the organ of interest, dose selection is guesswork, and a null result cannot be distinguished from an underdosed one. This is the same surrogate-endpoint gap that constrains the whole of geroscience measurement.
SenolyticsThe damage-repair position holds that intervention should target the accumulated products of aging, which can be enumerated, rather than the metabolism that produces them, which is not fully understood. Most biogerontologists regard the seven-category list as an assertion rather than a finding, and note that the boundary between damage and normal biological variation is not sharp. Both sides agree that damage accumulates; the disagreement is over research strategy.
SENS Research FoundationDeroy and Auvray argue that sensory substitution is better described as a learned perceptual skill, closer to reading than to seeing, and that it does not create a new modality with its own phenomenal character. Defenders point to distal attribution and to automaticity as evidence that something more than inference is occurring. Task performance cannot settle the dispute, since both accounts predict the same behaviour. The dispute is the clearest case in enhancement research where the question concerns experience rather than capability.
Sensory augmentationBoth wings accept Good's argument. They differ on the conditional probability of a good outcome given a fast transition. The safety wing treats that probability as low by default and the alignment problem as unsolved; the acceleration wing treats it as high, on the grounds that capability and benevolence tend to correlate and that delay has its own body count. Neither number is measurable, and the disagreement has proven resistant to evidence.
SingularitarianismStructured exercise has a randomized functional endpoint in older adults. Caloric restriction has decades of rodent lifespan data and controlled human trials on intermediate measures. Sleep has neither: no comparable mammalian lifespan experiment has manipulated sleep as the independent variable, and no human trial has been powered on events. The absence does not show that sleep is unimportant; it shows that most public advice about sleep is more confident than the evidence warrants.
Sleep and longevityA 2024 study in Nature Neuroscience, also in mice, used a different tracer approach and reported that clearance from brain tissue was reduced during sleep and under anaesthesia rather than increased. The two have not been reconciled; they differ in tracer, anaesthetic, and in what counts as clearance. Fluid-dynamics modelling has separately questioned whether bulk flow of the magnitude the glymphatic model requires is physically plausible. The popular version of the mechanism is stated more firmly than the literature behind it supports.
Sleep and longevityEvery figure in this article comes from a cohort selected for above-average health, screened repeatedly, and numbering in the hundreds across six decades. Anything but a large effect is often undetectable, and null results are rarely informative. In space physiology the mechanism is frequently better understood than the epidemiology.
Space medicineThe 2021, 2023, 2024 and 2026 results each come from a single implanted participant. The 2025 inner-speech study analysed four BrainGate2 participants (T12, T15, T16 and T17). The decoders behind those results were each trained on one person's own attempted speech. Models trained across people are newer: a 2026 reanalysis of two public datasets aligned each person's cortex into a shared space with a per-subject transform, and a 2026 preprint that pooled six intracortical participants reported that a shared model beat every individual one and adapted to a new user from fewer than 200 sentences, reaching a word error rate below 7 percent. BrainGate2 is a pilot clinical trial and the company studies are early feasibility studies, so the published figures describe individuals rather than a cohort, and there is no denominator for how often an implant reaches conversational performance. A trial that reported performance for every participant it enrolled would supply one.
Speech neuroprosthesisElon Musk wrote on 30 October 2026 that Neuralink's first product, Telepathy, "enables high-speed communication with computers or humans simply by thinking" and that its "[o]utput bandwidth will vastly exceed the fastest human speech or typing in a few years." In the speech study his post quotes, the implant decodes attempted or imagined speech from speech motor cortex, not thought in general. Natural conversation runs at roughly 150-160 words per minute. Peer-reviewed decoding has reached a median 78 with a quarter of words wrong, the longest home use averaged 56, and Neuralink gives its own median participant about 50, so matching ordinary conversation would take roughly a two- to threefold gain, and exceeding the fastest speakers more than that; no published result shows either. The trial behind the post lists safety, not speed, as its primary outcome.
Speech neuroprosthesisA large international market sells "stem cell therapy" for aging, joint pain and neurological disease, typically using autologous adipose-derived cells that have not been shown to engraft or differentiate. Regulators including the US Food and Drug Administration have taken enforcement action against several such providers. Three patients suffered severe and permanent bilateral vision loss after intravitreal injection of adipose-derived cells at one Florida clinic, a case series reported in the New England Journal of Medicine in 2017.
Stem cell exhaustionProposals for a bright line fall into three families. A provenance test asks whether fertilisation occurred, which is clear but arbitrary. A morphological test asks whether the structure has the parts of an embryo, which is vague. A potentiality test asks whether it could develop into a person if transferred. Most philosophers regard potentiality as the morally relevant criterion, but it is the hardest to apply, since testing it directly requires the transfer nobody permits.
Stem-cell-based embryo modelsSynchron argues that a device a vascular surgeon can implant in an ordinary catheter lab scales to a patient population that craniotomy never will, and that reliable simple control covers most of the clinical need. Critics respond that sixteen field-potential channels cap the device below the applications that justify an implant at all, and that the right comparison is with non-invasive eye trackers rather than with penetrating arrays.
StentrodeShowing that a treatment lowers a methylation age reading does not show it slowed aging. The reading could shift without any change in function, and running the same sample twice can move the estimate by more than a year on some platforms, which is large relative to reported intervention effects. Horvath was senior author on the TRIIM study, in which nine men receiving growth hormone alongside two other drugs showed reduced clock readings; the study had no control group, and it remains among the most cited human results in the area.
Steve HorvathBoth camps accept every physical fact about a functional duplicate and divide over what those facts establish. Chalmers argues that a gradual replacement preserving functional organisation cannot change experience without producing a subject systematically wrong about its own qualia; Searle answers that organisation is not the property doing the work, and that a system defined by the formal relations among its states has no semantics of its own. Neither position predicts an observation the other denies, which is why no result from neuroscience has moved anyone: a settled theory of consciousness would have to arrive first, and its verdict on non-biological systems would then be inherited rather than tested.
Substrate independenceSupporters argue that synthetic human chromosomes would give a clean testbed for regulatory genomics and virus-resistant manufacturing cells, and that the work stops at cell culture. Critics respond that the tools do not stop where the stated goal does, and that a capability to write human chromosomes eventually meets the questions raised by Germline editing and by Synthetic embryos whether or not anyone intends it to. Neither side disputes that the capability is decades away.
Synthetic genomesCoverage of the 2016 experiment frequently described human cells as having been made "part tardigrade" or as gaining the animal's survival abilities. Expressing one protein of roughly 400 amino acids in a cell line is a routine transfection experiment. It changes one property, partially, in vitro. It does not make a cell, still less a person, tolerant of anything a tardigrade tolerates.
Tardigrade genes and human cellsWhether EPR is a usable clinical phenomenon remains disputed. Some groups argue it is real but patient-specific and could be exploited by selecting patients whose tumours show high permeability, using the imaging and circulating-marker methods described in Nanoparticle diagnostics. Others argue that the effect is too weak and too inconsistent in humans to build a therapeutic strategy on. No approved nanomedicine's benefit is clearly attributable to EPR.
Targeted drug deliveryNo system has improved its own architecture without human direction. No theory predicts the feedback coefficient of such a loop. No consensus definition of general intelligence exists that would let anyone say when the threshold has been crossed. The singularity remains a conjecture with a long intellectual pedigree and no experimental support.
Technological singularityProponents typically defend the horizon claim, which is nearly unfalsifiable, and critics typically attack the accelerating-returns claim, which is the weakest. The mechanism claim, that recursive self-improvement has a strong enough feedback coefficient to run away, is the one that decides the question, and there is no direct evidence for or against it.
Technological singularityA high TRL implies past demonstration; it implies nothing about years to deployment. Some TRL 6 technologies reach routine use in three years and some never do. Any article on this wiki that pairs a TRL with a horizon is pairing an assessment with a guess, and the two should be read differently.
Technology readiness levelConsumer telomere tests, sold through the same direct-to-consumer testing channels as broad blood panels, report a length and an implied "cellular age". No professional body endorses telomere length as a clinical measure of aging outside the diagnosis of telomere biology disorders, and in most head-to-head comparisons methylation-based clocks outperform it as mortality predictors.
Telomeres and telomeraseSupporters treat the trillion-dollar figures as evidence that aging research is systematically underfunded relative to its expected return. Critics note that the same method applied to any large mortality reduction produces similar numbers, so the figure demonstrates that mortality is valuable rather than that geroscience will deliver. Neither side's figures show a fiscal saving: the one US simulation cited here with a budget line finds higher total entitlement spending unless eligibility ages rise.
The longevity dividendBernard Williams argued in 1970, a decade before Parfit's book, that the split cannot be blamed on exotic technology. Told he will be tortured after his memories and character are erased and replaced with another person's, a person's fear does not diminish, because "undergoing physical pain in the future is not excluded by any psychological state I may be in at the time"; yet the same case, presented as a body-swap between two people, pulls most readers toward the psychological criterion, and Williams closed by admitting he did not know which verdict to accept. The two presentations remain the cleanest statement of why no experiment can arbitrate: both camps agree on every physical fact and disagree only over which description a person facing the machine should use.
The teleportation problemCritics from both religious and secular positions have described transhumanism as a technological eschatology: resurrection by cryopreservation, transcendence by uploading, salvation by superintelligence. Transhumanists generally reply that their claims are empirical and falsifiable in a way that religious ones are not. The comparison still has force, because the movement's central promises cannot be falsified on any near timescale.
TranshumanismAnnouncements report births, and they report graft losses, failed embryo transfers and recipients who never became pregnant far less consistently. The published denominator (how many women were screened, transplanted, and left without a child) is incomplete. No randomised or matched comparison against surrogacy or adoption exists, and none is likely. A live birth demonstrates that the procedure can work; it does not establish how often it does.
Uterus transplantationCritics argue that the procedure subjects a healthy woman to a long operation, a second woman to immunosuppression, and a fetus to exposure to those drugs, for an outcome that is gestation and not survival. In their view this sits outside what transplantation has previously justified, and surrogacy or adoption achieves parenthood without any of it. Supporters answer that the same reasoning would forbid living kidney donation to a patient who could dialyse, that surrogacy is unavailable or illegal for many candidates, and that carrying a pregnancy is what these women are asking for and is not interchangeable with obtaining a child. The Montreal Criteria were an early attempt to fix the boundary, and professional bodies have generally held that the procedure is defensible under research protocols with independent donor advocacy.
Uterus transplantationNo published model of any human cell predicts its behaviour across DNA, RNA, protein and regulation together and has been shown to be right about a condition it was not trained on. What exists is a set of narrow predictors, most of them reading one modality, evaluated retrospectively on data that was collected before the model saw it. Fabian Theis, a computational biologist at Helmholtz Munich, told Nature in June 2026 that he did not think anyone would sensibly claim to have built a virtual cell unless they needed to sell a start-up.
Virtual cell modelsOne camp holds that only mechanistic models answer interventional questions, because a model fitted to observed data cannot say what happens under a condition never observed, and that interpretability is what makes a wrong prediction useful. The other holds that mechanistic models will stay underdetermined wherever parameters cannot be measured, and that learned models will predict better in practice even while explaining nothing. Hybrid systems that layer learned components onto mechanistic scaffolds are common in preprints and rare in anything validated. The Virtual Cell Challenge result, in which statistical methods held their own against pure learning, is currently read as support by both sides.
Virtual cell modelsNon-invasive optical glucose measurement has been announced repeatedly for decades and demonstrated by nobody. Products making the claim are sold anyway, and a person with diabetes acting on a fabricated number can be harmed within hours. This is the sharpest example of a general pattern in consumer diagnostics, where the marketed capability runs ahead of the physics.
Wearable health sensorsDevice makers cite cases where a notification preceded a diagnosis, which is real and unsystematic. Cardiologists point out that the randomized tests of intensive rhythm monitoring have raised diagnosis and treatment rates without demonstrating fewer strokes. Both positions are consistent with the evidence, because the trials run so far, on implants and on watches alike, have counted diagnoses rather than events.
Wearable health sensorsEstimates from within the emulation community typically fall in the second half of this century. Kenneth Miller, a computational neuroscientist at Columbia, argued in 2015 that even mapping the relevant structure of a human brain plausibly requires centuries at any realistic rate of improvement. The disagreement concerns how many unknown parameters the translation step hides, not the physics.
Whole brain emulationConstructs that genuinely propagate a design through a population, such as a Gene drive, do it by copying DNA, which is the mechanism missing here. Without heredity there is no variation for selection to act on, so the process cannot evolve. The runaway scenarios of the Grey goo literature also require a closure over materials and energy that nothing in this work approaches.
XenobotsCritics accept the observations and dispute the interpretation. Dissociated amphibian cells were known to reaggregate and sort by tissue type in the 1950s, and ciliated epithelium swims because that is what ciliated epithelium does. On the sceptical reading, a xenobot is a motile explant with a haircut, and the language of body plans, goals and reconfigurable organisms is doing work the data do not require. Levin's side replies that a stable, reproducible, functional morphology absent from the frog's life cycle is exactly what needs explaining. The two camps disagree about the explanatory frame; neither disputes the images.
XenobotsOne reading treats pig organs as a bridge, keeping patients alive until Lab grown organs, Organ bioprinting or the reseeded matrix scaffolds of Tissue engineering mature. Another treats them as the destination, on the argument that growing a vascularised human kidney from Induced pluripotent stem cells remains far harder than editing a pig, and that a manufacturable animal organ is the realistic supply solution for this century.
XenotransplantationRestoring measured function in three domains over one year would be the strongest human evidence any geroprotective intervention has produced. It would not show that lifespan is extended, that mortality falls, or that the effect persists after treatment stops. Function and survival are correlated but not the same, and the competition's timeframe cannot address the second.
XPRIZE HealthspanEverything above is uncertainty in the subject. This is uncertainty in the coverage — what these articles are missing, declared by the articles themselves. A reference work that publishes its own gaps is easier to trust than one that does not appear to have any.
Nothing on this page was written for it. Each entry is a callout an editor placed inside an article, at the exact sentence where the evidence stopped supporting the claim. Collecting them required no new judgement — only reading the corpus as one document instead of 180.
The result is a view of the field that no individual article can give and that secondary coverage almost never attempts. Popular writing about human futures aggregates the findings; the caveats are what get dropped in the retelling. Here they are the only thing kept.
Two entries deserve different weight. Thin evidence means the claim may well be true and has not been shown — a mouse result, twelve participants, a surrogate endpoint. Disagreement means the evidence is in and competent people still differ, which is a more durable condition and rarely resolves by waiting.
An article with no entry here is not necessarily an article with nothing uncertain in it. For a person or a completed event, having no caveat is usually correct. For a technology, it is more often a gap in the writing than a fact about the world, which is why the count sits in the backlog below rather than being presented as a clean bill of health.
The reverse also holds. An article carrying three cautions is not a worse article than one carrying none. Under this site’s editorial rules it is very likely a better one.