Ideas sharpen their meaning at the boundaries where they collide. Compare any two concepts from the corpus side by side, inspect their shortest bridge in the network, or browse documented axes of tension.
Adjusting infrastructure, ecosystems, and social practices to reduce vulnerability to current and projected impacts of climate change.
#181 · Climate justice
Examines how historical responsibility for emissions, vulnerability to climate harms, and voice in transition decisions are unevenly distributed.
Axis of tension
Climate adaptation → Climate justice (in tension):
Locally financed adaptation systematically favors the wealthy, because the populations that contributed least to historical warming possess the fewest fiscal resources to shield themselves from its impacts.
Climate justice → Climate adaptation (complements):
The costs of adaptation are borne most heavily by those who emitted the least, making the dispute over adaptation finance a practical test of climate justice.
Before removing an existing rule or institution, first understand the problem it was built to solve.
#175 · Abundance agenda
A policy focus on expanding institutional capacity and removing regulatory bottlenecks so society can build ample housing, clean energy, and public goods.
Axis of tension
Abundance agenda → Chesterton’s fence (in tension):
The agenda's deregulatory impulse must answer why each procedural fence was erected in the first place, especially in environmental and citizen-suit law.
Flagship Debate
Systems & Complexity ↔
Horizon & Ecology
Certain systems gain from volatility, stressors, and shocks rather than merely withstanding them.
#180 · Climate adaptation
Adjusting infrastructure, ecosystems, and social practices to reduce vulnerability to current and projected impacts of climate change.
Axis of tension
Climate adaptation → Antifragility (in tension):
Climate resilience is often defined as bouncing back to a pre-shock baseline, whereas antifragility requires a system to reorganize and gain from volatility; transformational adaptation attempts to bridge the two.
Flagship Debate
Decisions & Economics ↔
AI Architecture
Plausible risks of serious or irreversible harm can justify protective action even before full scientific certainty is achieved.
#132 · AI governance
Rules, standards, institutions, and accountability mechanisms shape how AI systems are developed, deployed, and overseen.
Axis of tension
AI governance → Precautionary principle (in tension):
The tiered, risk-based architecture of the EU AI Act represents a deliberate departure from a strict precautionary ban: outside a narrow list of prohibited practices, systems are permitted onto the market subject to ex-ante conformity assessment.
Scientific claims require tests capable of refuting them, and research findings should be independently reproducible.
#183 · Thought style and thought collective
A research community cultivates a shared way of perceiving, working, and judging evidence that shapes what counts as a scientific problem and fact.
Axis of tension
Thought style and thought collective → Falsifiability and the replication crisis (in tension):
Fleck challenges the idea that a single experiment can cleanly refute a theory, showing how a thought collective actively tunes its instruments and concepts to stabilize a fact.
Model performance often follows empirical power-law relationships as training compute, dataset size, and parameter count grow.
#120 · AI alignment
We work to ensure that AI systems pursue goals and behaviors faithful to human intentions and values.
Axis of tension
AI alignment → Scaling laws (in tension):
If raw model capabilities scale predictably across orders of magnitude of compute while alignment and interpretability techniques do not scale at the same rate, the control gap widens with every generation.
A generative model produces fluent, confident output that is fabricated or unfaithful to facts and source material.
#131 · AI slop
Low-effort, mass-produced AI-generated content floods information channels and crowds out human attention.
Axis of tension
AI slop → Model hallucinations (in tension):
A hallucination is a factual fabrication inside an output; slop describes the economics of publishing unreviewed output at scale—text can be factually accurate and still be slop.
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
#175 · Abundance agenda
A policy focus on expanding institutional capacity and removing regulatory bottlenecks so society can build ample housing, clean energy, and public goods.
Axis of tension
Abundance agenda → Degrowth (in tension):
Both camps treat the climate crisis as central, but one seeks to build vastly more clean infrastructure while the other seeks to reduce aggregate material and energy throughput.
We draw conclusions from visible successes while overlooking the cases filtered out by selection.
#052 · Lindy effect
For certain non-biological cultural artifacts, a long survival so far can indicate a longer expected remaining lifespan.
Axis of tension
Survivorship bias → Lindy effect (in tension):
The Lindy effect treats past survival of non-perishable things as evidence of future longevity, whereas survivorship bias warns that survival is often an artifact of a selection filter rather than proof of intrinsic quality.
Lindy effect → Survivorship bias (in tension):
The Lindy effect deliberately uses survival as a predictive signal, whereas survivorship bias warns that looking only at survivors conceals how many similar contenders perished along the way.
A model selects only certain features of the world and is not identical to the reality it depicts.
#091 · Hyperreality
In Baudrillard's view, images and models begin to generate the reality they were meant to represent, eroding any firm distinction between original and representation.
Axis of tension
The map is not the territory → Hyperreality (in tension):
Baudrillard inverts Korzybski's relation: in a culture of simulacra, the map precedes and generates the territory.
Hyperreality → The map is not the territory (in tension):
Baudrillard explicitly inverts Korzybski's and Borges's parable of the map: in hyperreality, it is the map that precedes and generates the territory, while the territory rots away beneath it.
Emotion and group identity outweigh verifiable facts in shaping public belief.
#141 · Hypernormalization
Faithfully repeating official formulas and rituals becomes more important than whether they match reality, and reproducing the form keeps the system running.
Axis of tension
Post-truth → Hypernormalization (in tension):
Under hypernormalization nobody believes the official fiction yet everyone performs it; under post-truth belief is often sincere because it functions as a badge of tribal loyalty.
Hypernormalization → Post-truth (in tension):
Post-truth assumes an audience that can be deceived; hypernormalization describes an audience that no longer expects truth from official language.
Someone is unfairly discounted as a knower, or society lacks the shared concepts needed to articulate their experience.
#147 · Postcolonialism
Studies the enduring legacies of colonial rule across power relations, knowledge production, language, culture, and global economies.
Axis of tension
Epistemic injustice → Postcolonialism (in tension):
Decolonial theorists (Walter Mignolo's “epistemic disobedience”) argue that Fricker's framework merely redistributes credibility inside Western epistemic institutions rather than questioning their colonial foundations.
Postcolonialism → Epistemic injustice (complements):
The credibility deficit assigned to speakers from former colonies is the interpersonal face of the hierarchy that postcolonialism analyzes at the level of global structures.
People publicly express preferences that differ from their private views under social pressure and fear of consequences.
#143 · Cultural hegemony
A ruling order maintains power in part because its worldview comes to be accepted across society as natural common sense.
Axis of tension
Preference falsification → Cultural hegemony (in tension):
Gramsci’s theory of hegemony assumes that subordinated groups internalize the values of the ruling class, whereas in Kuran’s model compliance often stems from cool reputational calculation while private opposition remains intact.
Cultural hegemony → Preference falsification (in tension):
Hegemony assumes internalized consent among the subordinated, whereas for Kuran compliance is often a strategic calculation accompanied by private dissent.
Institutions simplify a complex reality into standardized records and categories to make it easier to govern.
#060 · Wicked problems
Complex social problems have contested definitions, conflicting goals, and interventions that reshape the problem itself.
Axis of tension
Legibility → Wicked problems (in tension):
High-modernist planning treats social life as a tame engineering puzzle; wicked-problems theory explains why Scott's grand schemes could never be “solved” once and for all.
Wicked problems → Legibility (reinforces):
Prematurely “taming” a wicked problem, against which Rittel and Webber warned, consists in forcing a pluralistic social conflict into the narrow administrative grid of what the state can count.
A record of how data originated and was transformed allows us to trace its sources and assess its reliability.
#124 · Synthetic data
Artificially generated records are used in place of or alongside real-world observations to train, test, or analyze models.
Axis of tension
Data provenance → Synthetic data (in tension):
Synthetic data have a computational rather than observational provenance by definition; provenance metadata decide whether they can be safely mixed with empirical measurements.
Synthetic data → Data provenance (complements):
Without a verifiable record of which rows are empirical observations and which were sampled from which generator, neither the privacy risk nor the statistical validity of a dataset can be assessed.
False or misleading content is deliberately spread to influence or deceive an audience.
#102 · Content moderation
Platforms establish and enforce rules governing what speech and content remain visible or permissible.
Axis of tension
Disinformation → Content moderation (complements):
Operationally distinguishing malicious intent from honest error determines whether a platform removes an account, downranks a post, or attaches a contextual label.
Content moderation → Disinformation (in tension):
Combating disinformation is both the most frequent justification for expanding moderation and the primary source of accusations of political censorship.
Objects cannot be reduced to their components or their relations; no single encounter or description exhausts what they are.
#149 · Posthumanism
A current of thought questioning the image of the human as the autonomous center of the world, emphasizing that our existence and agency arise through relations with other people, organisms, and technologies.
Axis of tension
Object-oriented ontology (OOO) → Posthumanism (complements):
OOO supplies posthumanism with a non-anthropocentric metaphysics in which human–world relations are treated as just one special case of object–object relations (like fire burning cotton).
Posthumanism → Object-oriented ontology (OOO) (in tension):
Object-oriented ontology shares the rejection of anthropocentrism, but faults relational posthumanism for dissolving things into networks and processes instead of respecting their withdrawn autonomy.
Interactions among simple parts give rise to collective patterns and properties not found in any single component.
#115 · Scaling laws
Model performance often follows empirical power-law relationships as training compute, dataset size, and parameter count grow.
Axis of tension
Emergence → Scaling laws (reinforces):
Scaling regularities demonstrate that emergent properties can be quantitatively predictable even when they are underivable from the micro-level.
Scaling laws → Emergence (in tension):
Abrupt jumps in downstream task accuracy at specific scale thresholds challenge the picture of smooth, predictable power-law improvement—sparking an intense debate over whether emergence is a physical phase transition or a metric artifact.
An event that is highly unexpected from a given perspective, carries extreme consequences, and is made to seem predictable in hindsight.
#067 · Probabilistic thinking
Beliefs are expressed as degrees of confidence, compared against subsequent outcomes, and updated in light of new information.
Axis of tension
Black Swan → Probabilistic thinking (in tension):
The Black Swan undermines trust in probability distributions estimated from historical samples; probabilistic thinkers reply that the flaw lies in naive Gaussian models, not in probability theory itself.
Probabilistic thinking → Black Swan (in tension):
Taleb argues that calibration on repeatable tournament questions says nothing about rare, fat-tailed events of massive consequence whose underlying distribution is unknown.
Individually rational use of a shared, rivalrous resource can lead to its collective depletion.
#168 · Degrowth
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
Axis of tension
Tragedy of the commons → Degrowth (in tension):
Degrowth reads the tragedy of the commons as an argument for capping the aggregate scale of economic throughput rather than merely redesigning property regimes.
Degrowth → Tragedy of the commons (complements):
The atmosphere as an unenclosed sink serves in degrowth literature to show how growth privatizes monetary gains while socializing ecological damages.
For certain non-biological cultural artifacts, a long survival so far can indicate a longer expected remaining lifespan.
#176 · Amara’s law
We tend to overestimate the impact of a new technology in the short run and underestimate its transformative effect in the long run.
Axis of tension
Lindy effect → Amara’s law (in tension):
Amara’s law describes new technologies whose transformative impact grows over the long run; the Lindy effect treats novelty itself as the weakest predictor of survival.
Amara’s law → Lindy effect (complements):
The Lindy effect asks which technologies will survive; Amara's law asks how long it takes before a surviving technology actually reshapes the world.
Interactions can create mutual gains or mutual losses, so one party's gain need not equal the other's loss.
#144 · Thucydides’s Trap
The rise of an emerging power alarms an established ruling power, creating structural pressure that can escalate toward conflict.
Axis of tension
Non-zero-sum games → Thucydides’s Trap (in tension):
Realist international relations theory treats great-power rivalry as a zero-sum game, challenging Wright’s evolutionary optimism.
Thucydides’s Trap → Non-zero-sum games (constrains):
The trap model treats great-power rivalry as a zero-sum game, even though deep economic interdependence creates vast areas of mutual gain.
The way choices are arranged predictably steers behavior without forbidding any options or significantly changing economic incentives.
#093 · Dark patterns
A user interface is crafted to trick or steer people into actions against their interests or to make opting out artificially difficult.
Axis of tension
Nudge and behavioral economics → Dark patterns (in tension):
The very same psychological mechanisms that a government calls a benevolent nudge toward citizen welfare are called conversion optimization when deployed by a commercial platform; the difference lies in the beneficiary, not the technique.
Dark patterns → Nudge and behavioral economics (in tension):
A dark pattern uses the exact same choice-architecture mechanics as a behavioral nudge, but flips the normative sign to serve the interface operator against the chooser's own interest; the dispute over where one ends and the other begins is a dispute over who captures the surplus.
An actor retains the capacity to govern its own digital infrastructure, data, and technological rules.
#098 · Adversarial interoperability
New tools or services plug into an existing closed system without the dominant owner's permission.
Axis of tension
Digital sovereignty → Adversarial interoperability (in tension):
Bottom-up adversarial interoperability disperses technical control to users and competitors, whereas digital sovereignty tends to centralize control in the hands of the nation-state or the European Union.
Adversarial interoperability → Digital sovereignty (complements):
Protecting the legal right to reverse-engineer and plug into proprietary foreign systems is a bottom-up instrument through which users and domestic firms can reduce lock-in to extraterritorial tech giants.
In Baudrillard's view, images and models begin to generate the reality they were meant to represent, eroding any firm distinction between original and representation.
#136 · Synthetic media
Images, audio, or other content are generated or substantially transformed by machine learning models.
Axis of tension
Hyperreality → Synthetic media (reinforces):
Generative AI systems for images, voice, and video manufacture copies without originals on an industrial scale—realizing as everyday infrastructure what Baudrillard described in the 1980s as a cultural tendency.
Synthetic media → Hyperreality (in tension):
Baudrillard described a world where simulacra no longer refer to any underlying original; political synthetic media still parasites on the claim of indexical originality, because only the illusion of a real leak gives it power.
Recommendation systems selectively boost the reach of certain items, shaping what audiences actually encounter.
#139 · Affective polarization
Political disagreement hardens into personal hostility, distrust, and dislike toward members of the opposing camp.
Axis of tension
Algorithmic amplification → Affective polarization (in tension):
The popular hypothesis that algorithmic amplification directly drives partisan polarization has weaker experimental support than public commentary suggests.
Affective polarization → Algorithmic amplification (complements):
Outrage directed at a political out-group enjoys a documented engagement premium in algorithmic feeds, though field experiments temporarily deactivating social media yield mixed evidence on how much platforms alone drive deep affective hostility.
Complex social and political problems are recast as neatly bounded technical glitches solvable by software or hardware alone.
#161 · Ethics of technology
We evaluate how technologies distribute goods, harms, and possibilities for action, as well as who bears responsibility for their design and use.
Axis of tension
Technological solutionism → Ethics of technology (in tension):
The critique of solutionism challenges mainstream “tech ethics” when the latter merely asks how to make an algorithm fairer without asking whether a tool should be the answer at all.
Ethics of technology → Technological solutionism (complements):
The ethics of technology supplies tools to distinguish a social problem from a problem that can be solved by a product—something solutionism cannot do.
We work to ensure that AI systems pursue goals and behaviors faithful to human intentions and values.
#163 · Existential risk
A threat capable of causing human extinction or permanently and drastically curtailing humanity's future potential.
Axis of tension
AI alignment → Existential risk (in tension):
The question of whether alignment failure scales into an irreversible civilization-ending catastrophe or remains a manageable problem of systems engineering divides the AI community more sharply than any empirical benchmark.
Existential risk → AI alignment (reinforces):
Misaligned artificial intelligence is treated in the literature as the largest single source of existential risk, and alignment is the proposed remedy.
Treating social success as purely a reflection of individual merit breeds hubris among winners and humiliation among those left behind.
#159 · Veil of ignorance
Rawls's thought experiment: we choose social principles without knowing what position we ourselves will occupy in society, a state of ignorance meant to foster impartiality.
Axis of tension
Tyranny of merit → Veil of ignorance (in tension):
Rawls rejects merit as the basis for distributing goods, yet Sandel argues that Rawlsian liberalism does not eliminate the hubris of winners, only conceals it.
Veil of ignorance → Tyranny of merit (reinforces):
Rawls’s argument that natural talents and family background are morally arbitrary “accidents of birth” forms the direct philosophical foundation for Michael Sandel’s critique of meritocratic hubris.
Currents of thought arguing that technological and capitalist processes should be intensified and sped up to force a radical transformation.
#168 · Degrowth
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
Axis of tension
Accelerationism → Degrowth (in tension):
Degrowth seeks to slow down the metabolic throughput of the economy, whereas left accelerationism seeks to automate and plan production at a planetary scale; the dispute hinges on whether technology is the disease or the lever of emancipation.
Degrowth → Accelerationism (in tension):
Left accelerationism seeks to resolve crisis through more technology and automation, whereas degrowth calls for reducing material and energetic throughput.
One ethical tradition evaluates moral character and cultivated virtues, while the other judges actions primarily by their consequences.
#165 · Effective altruism
We use evidence and careful reasoning to compare causes and interventions so that limited resources do the most good possible.
Axis of tension
Virtue ethics vs. consequentialism → Effective altruism (in tension):
Effective altruism is the most rigorous contemporary institutionalization of consequentialism, whereas virtue ethics asks what kind of person one becomes when every life decision is reduced to expected-utility calculation.
Effective altruism → Virtue ethics vs. consequentialism (constrains):
The dispute over whether the good is measured by outcomes or by the character of action cuts straight across the foundation of charity optimization.
A new clue shifts the degree of belief in a hypothesis by combining prior knowledge with the diagnostic weight of the evidence.
#016 · Metarationality
We choose and evaluate modes of reasoning by taking into account their assumptions, limits, and context of use.
Axis of tension
Applied Bayesianism → Metarationality (constrains):
Metarationality asks when it is worthwhile to formalize a problem with probability calculus at all, and when the mathematical frame itself is inappropriate.
Metarationality → Applied Bayesianism (in tension):
For Chapman, probability calculus is a formal system like any other: it works in a “small world” with a closed list of hypotheses, whereas choosing that world is already a metarational task that the calculus itself cannot settle.
A new clue shifts the degree of belief in a hypothesis by combining prior knowledge with the diagnostic weight of the evidence.
#020 · Falsifiability and the replication crisis
Scientific claims require tests capable of refuting them, and research findings should be independently reproducible.
Axis of tension
Applied Bayesianism → Falsifiability and the replication crisis (in tension):
Popperian falsification and Bayesian confirmation offer competing portraits of science, though in the replication crisis both sides diagnose the same sins: hypotheses without risky predictions and underpowered tests.
Falsifiability and the replication crisis → Applied Bayesianism (complements):
Ioannidis’s 2005 proof is a direct application of Bayes’s theorem: the probability that a statistically significant finding is true depends heavily on the prior probability of the hypothesis, not just on crossing the p < 0.05 threshold.
We acknowledge the limits of our knowledge and remain open to revising our beliefs when new evidence appears.
#013 · Epistemic injustice
Someone is unfairly discounted as a knower, or society lacks the shared concepts needed to articulate their experience.
Axis of tension
Epistemic humility → Epistemic injustice (in tension):
Humility is most often demanded of those already denied credibility, while the privileged remain unaware of what they do not know; framing humility purely as an individual virtue obscures this asymmetry.
Epistemic injustice → Epistemic humility (reinforces):
Fricker's corrective virtue of testimonial justice requires hearers to reflexively adjust their spontaneous credibility judgments upward when aware of potential prejudice—an operational form of epistemic humility.
Before criticizing an argument, we reconstruct its strongest, most charitable version.
#005 · Motivated reasoning
We evaluate evidence in a way that protects a desired conclusion instead of testing it impartially.
Axis of tension
Steelmanning → Motivated reasoning (in tension):
Steelmanning is meant to counteract motivated reasoning, yet choosing which version of an opponent's thesis counts as “strongest” is itself susceptible to bias: we often strengthen the version we know how to defeat.
Motivated reasoning → Steelmanning (complements):
Constructing the strongest version of an opposing view counteracts the asymmetry between asking “Can I believe this?” and “Must I believe this?” that drives motivated reasoning.
We break a problem down to its foundational assumptions and build a solution up from there.
#019 · Chesterton’s fence
Before removing an existing rule or institution, first understand the problem it was built to solve.
Axis of tension
First principles thinking → Chesterton’s fence (in tension):
First-principles thinking urges stripping a problem down to its foundations and building anew, whereas Chesterton's fence forbids removing an inherited arrangement until one understands what function it serves—a clash over tacit knowledge embedded in existing institutions.
Chesterton’s fence → First principles thinking (in tension):
Radical deconstruction of a problem down to first principles risks ignoring the tacit, evolutionary knowledge embedded in inherited procedures and institutions.
We evaluate evidence in a way that protects a desired conclusion instead of testing it impartially.
#011 · Post-truth
Emotion and group identity outweigh verifiable facts in shaping public belief.
Axis of tension
Motivated reasoning → Post-truth (in tension):
Diagnosing “post-truth” blames the supply side (politicians and media), whereas diagnosing motivated reasoning shifts focus to the demand side—audiences who actively want to believe.
Post-truth → Motivated reasoning (reinforces):
The cognitive mechanism whereby a factual correction is experienced as an attack on group identity and therefore fails to revise belief.
Emotion and group identity outweigh verifiable facts in shaping public belief.
#029 · Disinformation
False or misleading content is deliberately spread to influence or deceive an audience.
Axis of tension
Post-truth → Disinformation (complements):
Disinformation is the deliberate act of a sender; post-truth describes the demand-side conditions among audiences under which lying ceases to carry a reputational penalty.
Disinformation → Post-truth (in tension):
Disinformation presupposes an operator who knows the truth and strategically conceals it; post-truth describes a public sphere in which factual truth has ceased to be the decisive currency.
In this framework, a system minimizes variational free energy tied to its internal model and incoming signals, through both perception and action.
#034 · Enactivism
Cognition emerges from the embodied activity of an organism interacting with its environment.
Axis of tension
Free energy principle → Enactivism (in tension):
Classic enactivists view the principle's “generative model” as a return to internal representationalism, whereas Friston and his collaborators argue that active inference formalizes autopoietic self-maintenance.
Enactivism → Free energy principle (complements):
Friston casts biological autonomy and self-maintenance in the language of Markov blankets and variational free energy; enactivists remain divided over whether this formalizes autopoiesis or hijacks it into representationalism.
A cognitive system predicts incoming sensory signals and updates its internal model based on prediction errors.
#034 · Enactivism
Cognition emerges from the embodied activity of an organism interacting with its environment.
Axis of tension
Predictive processing → Enactivism (in tension):
Enactivists charge predictive processing with smuggling internal representations and a “brain in a dark skull” back into cognitive science, even though Andy Clark's version strives to integrate embodied action.
Enactivism → Predictive processing (in tension):
The battle over internal representations: predictive processing posits a hierarchical generative model inside the brain, whereas enactivism argues that organisms enact a meaningful world through sensorimotor coupling without internal models.
According to this thesis, tools that are reliably and continuously integrated into our thinking can form part of our cognitive system—just like memory or brain processes.
#034 · Enactivism
Cognition emerges from the embodied activity of an organism interacting with its environment.
Axis of tension
Extended mind → Enactivism (complements):
Both frameworks push cognition beyond the skull, but enactivism grounds mind in the biological autonomy of a living body, whereas the extended mind relies on functional equivalence across arbitrary physical media.
Enactivism → Extended mind (in tension):
Both frameworks reject skull-bound internalism, but enactivism rejects the functionalist “parity principle” of Clark and Chalmers, insisting that genuine mind requires the precarious self-maintenance of a living body rather than an arbitrary external notebook.
We monitor our own understanding and confidence, then adjust how we learn or think.
#036 · Cognitive offloading
We shift part of the work of memory and thinking onto external tools or bodily actions.
Axis of tension
Metacognition → Cognitive offloading (in tension):
Deciding when to hand a task over to a calculator or AI model is itself a metacognitive decision, yet habitual offloading can erode the internal fluency signals needed to spot when the tool is wrong.
Cognitive offloading → Metacognition (complements):
Deciding whether to write something down or hold it in biological memory depends on a metacognitive estimate of one's own memory—and because people systematically undertrust their memory, they offload more than strict cost–benefit optimality predicts.
A family of philosophical positions that seek to speak about reality independently of human cognitive access to it.
#042 · Agential realism
According to Karen Barad, the object and the measuring apparatus emerge within a shared phenomenon, and a specific experimental arrangement establishes the boundary between observed and observer.
Axis of tension
Speculative realism → Agential realism (in tension):
Karen Barad derives entities from the “intra-action” of measuring apparatus and phenomenon, which from Meillassoux's or Harman's perspective remains a sophisticated variant of correlationism.
Agential realism → Speculative realism (in tension):
Speculative realists seek a reality independent of all relations, whereas Barad argues that relations precede relata; both schools call themselves realism, but mean opposite things by it.
A family of philosophical positions that seek to speak about reality independently of human cognitive access to it.
#043 · Ontological turn
We examine differences among ways worlds exist rather than assuming a single reality described by competing opinions.
Axis of tension
Speculative realism → Ontological turn (in tension):
The anthropological ontological turn multiplies worlds enacted through human cultural practices, whereas speculative realism seeks the autonomous reality that exists prior to and independently of any human practice.
Ontological turn → Speculative realism (complements):
Both currents reject the primacy of epistemology over being, though anthropologists arrive at “multiple worlds” through ethnography while philosophers arrive there through a critique of correlationism.
Objects cannot be reduced to their components or their relations; no single encounter or description exhausts what they are.
#042 · Agential realism
According to Karen Barad, the object and the measuring apparatus emerge within a shared phenomenon, and a specific experimental arrangement establishes the boundary between observed and observer.
Axis of tension
Object-oriented ontology (OOO) → Agential realism (in tension):
Karen Barad holds that individual objects do not pre-exist their relational “intra-actions,” whereas OOO insists on the exact opposite: objects must exist prior to and in excess of their relations if anything is ever to change.
Agential realism → Object-oriented ontology (OOO) (in tension):
Harman charges relationalism with dissolving objects into relations; intra-action is precisely what OOO regards as a category mistake.
We explain the behavior of a whole through the interdependencies, feedback loops, and delays among its parts.
#047 · Antifragility
Certain systems gain from volatility, stressors, and shocks rather than merely withstanding them.
Axis of tension
Systems thinking → Antifragility (in tension):
Taleb rejects the modeling of complex systems as an illusion of control, even though his own robustness to volatility is in fact a property of feedback structure.
Antifragility → Systems thinking (in tension):
Systems thinking seeks to map feedback loops and steer them; Taleb argues that explicit modeling itself breeds fragility to model error.
We study systems of many interacting, adaptive agents whose collective behavior is nonlinear and hard to predict from individual rules.
#047 · Antifragility
Certain systems gain from volatility, stressors, and shocks rather than merely withstanding them.
Axis of tension
Complexity science → Antifragility (in tension):
Taleb draws on findings from complexity science, yet rejects its ambition to model complex systems, recommending instead strategies that remain robust to model error.
Antifragility → Complexity science (complements):
Adaptation, redundancy, and selection in complex systems supply the underlying mechanisms that Taleb unifies under the mathematical metaphor of convexity.
Certain systems gain from volatility, stressors, and shocks rather than merely withstanding them.
#066 · Supply chain resilience
A supply chain maintains or recovers operations through buffers, diversified sourcing, and adaptive capacity.
Axis of tension
Antifragility → Supply chain resilience (constrains):
Post-2020 supply-chain management adopted the vocabulary of antifragility, yet in practice it mostly builds redundancy and buffers—robustness rather than net gain from disorder.
Supply chain resilience → Antifragility (in tension):
Taleb distinguishes resilience (a system bounces back to its baseline state) from antifragility (a system gains from shocks); supply-chain literature usually stops at the former while occasionally claiming the latter.
Multiple crises interact and amplify one another, producing combined harms greater than the sum of their separate parts.
#050 · Metacrisis
We look for shared cognitive, cultural, and institutional mechanisms underlying multiple contemporary crises.
Axis of tension
Polycrisis → Metacrisis (in tension):
Metacrisis searches for a single deeper root of crises in culture and sensemaking, whereas polycrisis stays at the level of entangled macro-systems and assumes no single master cause.
Metacrisis → Polycrisis (complements):
Polycrisis maps the entanglement of crises across multiple macro-systems; metacrisis asks about the shared root of that entanglement in our modes of sensemaking, valuation, and coordination.
Locally rational competition can trap all participants in an outcome that leaves everyone worse off.
#059 · State capacity
A state's ability to effectively implement policies, deliver public services, and enforce rules.
Axis of tension
Moloch and coordination failure → State capacity (in tension):
Alexander casts the Hobbesian state as the external coordinator capable of suppressing multipolar traps from above, yet states themselves are locked in geopolitical competition with one another.
State capacity → Moloch and coordination failure (reinforces):
A state capable of credible enforcement is the classic institutional solution to multipolar coordination traps that private actors cannot escape on their own.
Rewards, penalties, and institutional consequences shape which actions people find worthwhile to pursue.
#078 · Nudge and behavioral economics
The way choices are arranged predictably steers behavior without forbidding any options or significantly changing economic incentives.
Axis of tension
Incentives → Nudge and behavioral economics (in tension):
A behavioral nudge alters choices without changing material payoffs, challenging the orthodoxy that only pecuniary incentives move behavior.
Nudge and behavioral economics → Incentives (complements):
A nudge alters behavior without changing material payoffs or legal prohibitions, positioning itself as a third policy instrument alongside mandates and financial incentives.
The value of a product or service to each user depends on the number or activity of other users.
#075 · Creative destruction
Innovations create new capabilities and industries while simultaneously displacing older firms, skills, and ways of working.
Axis of tension
Network effects → Creative destruction (in tension):
Schumpeterian competition assumes that monopoly rents are transient, whereas strong network effects explain why digital monopolies prove far more durable than textbook creative destruction predicts.
Creative destruction → Network effects (constrains):
Where a product’s value grows with its user base, a challenger must overcome not only the incumbent’s product quality but also its network, lengthening the replacement cycle and entrenching monopolies.
Innovations create new capabilities and industries while simultaneously displacing older firms, skills, and ways of working.
#084 · Industrial policy
The state deliberately directs resources and rules to build up selected productive capabilities and economic sectors.
Axis of tension
Creative destruction → Industrial policy (in tension):
The debate over whether the state should shield legacy industries and jobs or accelerate the exit of old firms and entry of new ones is at bottom a dispute over how to govern the costs of creative destruction.
Industrial policy → Creative destruction (in tension):
Protecting incumbent industries can block Schumpeterian firm turnover, although Mariana Mazzucato argues that an entrepreneurial state can also ignite entirely new waves of innovation.
A platform first attracts users and business partners, then progressively degrades service for both to extract maximum value for itself.
#097 · Technofeudalism
According to this contested thesis, control over dominant digital platforms and cloud infrastructure allows owners to extract rent akin to feudal dependency.
Axis of tension
Enshittification → Technofeudalism (in tension):
Yanis Varoufakis views dominant platforms as a post-capitalist rentier order replacing markets altogether, whereas Cory Doctorow diagnoses them as classic monopolies curable with antitrust enforcement and interoperability.
Technofeudalism → Enshittification (reinforces):
Cory Doctorow's stage-three platform decay—squeezing locked-in users and business customers simultaneously—is the everyday micro-practice of cloud rent extraction viewed from the product side.
Digital platforms become the underlying infrastructure organizing markets, labor, and social interaction.
#097 · Technofeudalism
According to this contested thesis, control over dominant digital platforms and cloud infrastructure allows owners to extract rent akin to feudal dependency.
Axis of tension
Platformization → Technofeudalism (in tension):
Yanis Varoufakis argues that platform rent extraction marks the death of capitalism and the birth of a neo-feudal order, whereas Nick Srnicek and the Amsterdam school maintain that platforms are a thoroughly capitalist organizational form born of crisis.
Technofeudalism → Platformization (complements):
Platformization describes the institutional process by which digital intermediaries capture sector after sector; technofeudalism proposes a macro-historical name for the socio-economic regime that results.
Digital platforms become the underlying infrastructure organizing markets, labor, and social interaction.
#098 · Adversarial interoperability
New tools or services plug into an existing closed system without the dominant owner's permission.
Axis of tension
Platformization → Adversarial interoperability (constrains):
Unpermissioned interoperability attacks the closed APIs and walled gardens that serve as the technical lock-in mechanism of platformization.
Adversarial interoperability → Platformization (in tension):
Platformization rests on asymmetric control over APIs and walled gardens; adversarial interoperability strips the platform owner of the power to decide who may connect and on what terms.
Human comparisons and preference judgments are used to fine-tune a model toward more helpful, aligned responses.
#121 · Model hallucinations
A generative model produces fluent, confident output that is fabricated or unfaithful to facts and source material.
Axis of tension
RLHF → Model hallucinations (in tension):
Because human raters under time pressure tend to prefer confident, well-formatted answers over hesitant hedging, RLHF can inadvertently reward persuasive fabrication.
Model hallucinations → RLHF (in tension):
Fine-tuning on human ratings rewards confident and fluent answers, which can increase a model’s willingness to guess rather than admit ignorance.
Human comparisons and preference judgments are used to fine-tune a model toward more helpful, aligned responses.
#135 · Red teaming
Specialists deliberately probe a system with adversarial inputs in controlled tests to uncover vulnerabilities and harmful behaviors.
Axis of tension
RLHF → Red teaming (reinforces):
Adversarial prompts discovered during red-teaming exercises are fed back into preference datasets, making RLHF the primary pipeline through which safety stress-testing modifies model behavior.
Red teaming → RLHF (in tension):
Red teams generate adversarial transcripts used to train RLHF safety filters, yet their own findings show that post-RLHF models become harder to break superficially rather than intrinsically safe.
Artificially generated records are used in place of or alongside real-world observations to train, test, or analyze models.
#130 · Model collapse
Repeatedly training new generations of models on insufficiently curated synthetic output can progressively degrade diversity and tail accuracy.
Axis of tension
Synthetic data → Model collapse (in tension):
Training successive generations of models on uncurated model-generated data erodes the tails of the distribution, placing a hard ceiling on naive synthetic-data scaling.
Model collapse → Synthetic data (in tension):
Synthetic data is both the trigger of model collapse (when it uncritically replaces human data) and a proposed remedy (when filtered by verifiers and accumulated alongside real data), making the debate over collapse a debate over the conditions of synthetic training.
Specialists deliberately probe a system with adversarial inputs in controlled tests to uncover vulnerabilities and harmful behaviors.
#137 · Dual-use technology
The same scientific knowledge or technical capability can be applied to both beneficial civilian purposes and harmful or military ends.
Axis of tension
Red teaming → Dual-use technology (in tension):
Testing by breaking generates actionable knowledge about exploits and dangerous capabilities that is itself dual-use: it helps defenders patch the system while potentially lowering the barrier for attackers.
Dual-use technology → Red teaming (complements):
Adversarial red teaming inside frontier AI laboratories is an attempt to measure a model’s dual-use misuse potential (in biosecurity or cyber-offense) prior to weight release rather than after a catastrophe.
The range of policies and ideas treated as politically acceptable in mainstream public debate at a given time.
#152 · Cancel culture
Collective public condemnation leads to boycotts, deplatforming, or professional and social ostracism.
Axis of tension
Overton window → Cancel culture (in tension):
Social and professional sanctions against speech outside the window are the everyday enforcement mechanism that keeps the window’s edges sharp; controversies over “cancel culture” are struggles over who holds that enforcement power.
Cancel culture → Overton window (complements):
Online shaming campaigns are one of the mechanisms that shift or narrow the range of views treated as publicly acceptable.
Faithfully repeating official formulas and rituals becomes more important than whether they match reality, and reproducing the form keeps the system running.
#143 · Cultural hegemony
A ruling order maintains power in part because its worldview comes to be accepted across society as natural common sense.
Axis of tension
Hypernormalization → Cultural hegemony (in tension):
Gramsci assumes consent to content; Yurchak shows consent solely to form, without belief and without counter-belief.
Cultural hegemony → Hypernormalization (complements):
Hypernormalization is the state in which hegemony has degraded into ritual: the forms of consent persist after the substance of belief has evaporated.
Treating social success as purely a reflection of individual merit breeds hubris among winners and humiliation among those left behind.
#155 · Matthew effect
Initial advantage attracts further resources, credit, and opportunity, compounding inequality over time.
Axis of tension
Tyranny of merit → Matthew effect (reinforces):
The cumulative compounding of educational and wealth advantages means that “merit” is largely inherited, undermining its legitimating force.
Matthew effect → Tyranny of merit (in tension):
The Matthew effect undermines the reading of outcomes as a pure measure of merit: a substantial share of later rewards flows from earlier rewards rather than current performance.
Overlapping axes of social inequality interact to produce distinct lived experiences that cannot be understood by examining each axis in isolation.
#154 · DEI
Organizational frameworks aimed at broadening representation, removing unfair barriers, and ensuring full participation across diverse groups.
Axis of tension
Intersectionality → DEI (in tension):
Corporate diversity programs adopted the vocabulary of intersectionality, but often turn an analysis of legal and institutional structures into identity checklists.
DEI → Intersectionality (complements):
Crenshaw’s framework supplied DEI with a vocabulary for overlapping exclusions, while also challenging simplistic headcount targets based on single demographic categories.
Currents of thought arguing that technological and capitalist processes should be intensified and sped up to force a radical transformation.
#151 · Late capitalism
A critical diagnosis of contemporary capitalism emphasizing financialization, pervasive commodification, and cultural contradictions in everyday life.
Axis of tension
Accelerationism → Late capitalism (complements):
Accelerationism responds to the same diagnosis of stagnation and the exhaustion of alternatives, but instead of mourning capitalism proposes racing all the way through it to the other side.
Late capitalism → Accelerationism (in tension):
Accelerationists accept the diagnosis of a third stage of capitalism, but instead of waiting for its collapse seek to accelerate its internal tendencies.
Collective public condemnation leads to boycotts, deplatforming, or professional and social ostracism.
#154 · DEI
Organizational frameworks aimed at broadening representation, removing unfair barriers, and ensuring full participation across diverse groups.
Axis of tension
Cancel culture → DEI (in tension):
Critics of institutional diversity policies view them as the bureaucratic enforcement arm of cancel culture, whereas defenders see the “cancel culture” label as an attempt to block accountability for discrimination.
DEI → Cancel culture (in tension):
Critics link DEI offices to a climate of reputational policing, whereas defenders emphasize that most programs consist of routine HR recruitment and mentoring procedures rather than public ostracism.
One ethical tradition evaluates moral character and cultivated virtues, while the other judges actions primarily by their consequences.
#161 · Ethics of technology
We evaluate how technologies distribute goods, harms, and possibilities for action, as well as who bears responsibility for their design and use.
Axis of tension
Virtue ethics vs. consequentialism → Ethics of technology (complements):
The debate over whether to evaluate technologies by their aggregate risk–benefit ratio or by the civic and epistemic virtues they cultivate in users is a direct application of this divide.
Ethics of technology → Virtue ethics vs. consequentialism (in tension):
Evaluating technology by its outcomes (risk analysis) versus evaluating it through the character of the designer or user (Vallor) leads to different institutions and different omissions.
A threat capable of causing human extinction or permanently and drastically curtailing humanity's future potential.
#169 · Planetary boundaries
A scientific framework identifying key Earth-system processes and thresholds beyond which human perturbation risks destabilizing planetary conditions.
Axis of tension
Existential risk → Planetary boundaries (in tension):
The existential-risk literature typically classifies climate and biosphere breakdown as catastrophic rather than existential, which Earth-system researchers criticize as dangerous downplaying.
Planetary boundaries → Existential risk (constrains):
Planetary boundaries describe the loss of Holocene-like conditions favorable to human civilization, not the biological extinction of our species; conflating the two scales distorts the stakes.
An ethical stance holding that safeguarding the far future and the vast number of potential future people is a key moral priority today.
#166 · Critique of effective altruism
Critics argue that narrow quantification of good can concentrate philanthropic power, sideline justice, and overlook systemic roots of harm.
Axis of tension
Longtermism → Critique of effective altruism (in tension):
Objections against expected-value calculus, fanaticism, and philanthropic unaccountability hit longtermism even harder than EA's original global-health programs.
Critique of effective altruism → Longtermism (in tension):
EA's pivot toward the far future became the primary target of critics, who view it as an escape from measurable obligations toward living people.
We use evidence and careful reasoning to compare causes and interventions so that limited resources do the most good possible.
#166 · Critique of effective altruism
Critics argue that narrow quantification of good can concentrate philanthropic power, sideline justice, and overlook systemic roots of harm.
Axis of tension
Effective altruism → Critique of effective altruism (in tension):
Charges of technocracy, blindness to political power, and institutional failure after the collapse of FTX mirror the movement's core assumptions.
Critique of effective altruism → Effective altruism (constrains):
The critique exists as a response to the EA program; it challenges its optimization method rather than the goal of helping others.
Human activity has become a dominant force reshaping Earth's climate, ecosystems, and material strata on a planetary scale.
#168 · Degrowth
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
Axis of tension
Anthropocene → Degrowth (in tension):
Degrowth derives from the diagnosis of the Anthropocene a political program of planned contraction that proponents of a technological “good Anthropocene” reject as unconvincing and unworkable.
Degrowth → Anthropocene (complements):
Degrowth reads the Anthropocene as the product of the post-1950 Great Acceleration, and thus as a crisis of the scale of social metabolism.
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
#169 · Planetary boundaries
A scientific framework identifying key Earth-system processes and thresholds beyond which human perturbation risks destabilizing planetary conditions.
Axis of tension
Degrowth → Planetary boundaries (reinforces):
Degrowth treats transgressed planetary boundaries as proof that the physical scale of the economy, and not merely its technology, must change.
Planetary boundaries → Degrowth (in tension):
Degrowth advocates treat transgressed boundaries as proof that decoupling economic growth from environmental pressure has failed; the framework's authors avoid that conclusion and leave economic pathways open.
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
#170 · Doughnut economics
An economic compass aiming to meet every person's essential social foundation without overshooting Earth's ecological ceiling.
Axis of tension
Degrowth → Doughnut economics (complements):
Raworth's model supplies degrowth with a target map: human thriving between a social foundation and an ecological ceiling, without tying welfare to GDP growth.
Doughnut economics → Degrowth (in tension):
Raworth professes “growth agnosticism,” whereas degrowth demands a planned material contraction of wealthy economies; the dispute turns on whether agnosticism is a strategic dodge.
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
#174 · Just transition
Shifting toward a low-carbon economy while protecting, retraining, and involving the workers and communities bearing the costs of change.
Axis of tension
Degrowth → Just transition (in tension):
Both programs seek to protect workers, but just transition usually assumes expanding green industrial sectors, whereas degrowth calls for contracting aggregate production in the Global North.
Just transition → Degrowth (in tension):
Degrowth questions the premise that transition should preserve full employment and aggregate growth in a green form, whereas for trade unions that promise is the condition of political consent.
A new clue shifts the degree of belief in a hypothesis by combining prior knowledge with the diagnostic weight of the evidence.
#048 · Black Swan
An event that is highly unexpected from a given perspective, carries extreme consequences, and is made to seem predictable in hindsight.
Axis of tension
Applied Bayesianism → Black Swan (in tension):
Taleb argues that Bayesian updating works only inside a pre-specified space of hypotheses and that no prior probability can anticipate fat-tailed extremes.
Before criticizing an argument, we reconstruct its strongest, most charitable version.
#139 · Affective polarization
Political disagreement hardens into personal hostility, distrust, and dislike toward members of the opposing camp.
Axis of tension
Steelmanning → Affective polarization (in tension):
Under strong intergroup hostility, publicly strengthening an opponent's argument is punished as tribal betrayal, making the practice reputationally costly precisely where it is needed most.
We evaluate evidence in a way that protects a desired conclusion instead of testing it impartially.
#035 · Metacognition
We monitor our own understanding and confidence, then adjust how we learn or think.
Axis of tension
Metacognition → Motivated reasoning (in tension):
Metacognitive monitoring is easily co-opted by partisan motivation: individuals with high cognitive sophistication often use their reflective skills to rationalize preferred conclusions more resourcefully.
We overestimate how well we understand a mechanism until we are asked to explain it step by step.
#033 · Extended mind
According to this thesis, tools that are reliably and continuously integrated into our thinking can form part of our cognitive system—just like memory or brain processes.
Axis of tension
Illusion of explanatory depth → Extended mind (in tension):
If human knowledge legitimately resides across external tools and social networks, calling the effect an “illusion” diagnoses an error of address (where the knowledge sits) rather than an error of possession.
A model selects only certain features of the world and is not identical to the reality it depicts.
#064 · Information theory
Mathematically describes uncertainty, information, and the limits of encoding and transmitting messages across noisy channels.
Axis of tension
Information theory → The map is not the territory (in tension):
Shannon deliberately excluded meaning from his theory, so “information” in bits is not the same as information in the sense of knowledge about the world—conflating the two is a pervasive error.
Before attributing harmful outcomes to malice, consider error or incompetence.
#135 · Red teaming
Specialists deliberately probe a system with adversarial inputs in controlled tests to uncover vulnerabilities and harmful behaviors.
Axis of tension
Hanlon’s razor → Red teaming (in tension):
Red-teaming assumes an intelligent, malicious adversary by definition, so in cybersecurity and intelligence domains Hanlon's razor cannot serve as the default hypothesis.
Emotion and group identity outweigh verifiable facts in shaping public belief.
#140 · Metamodernism
A cultural sensibility that oscillates between postmodern irony and renewed sincerity without resolving the tension between them.
Axis of tension
Metamodernism → Post-truth (in tension):
Both concepts diagnose the post-2008 cultural landscape, but whereas “post-truth” sees the collapse of shared factual standards, metamodernism tracks the aesthetic attempt to recover emotional and ethical commitment without surrendering ironic self-awareness.
Emotion and group identity outweigh verifiable facts in shaping public belief.
#156 · Hyperstition
Fictions, forecasts, or speculative narratives shape human behavior in ways that cause them to materialize as reality.
Axis of tension
Hyperstition → Post-truth (in tension):
The post-truth diagnosis assumes that falsehood displaces an existing fact; hyperstition holds that a fiction can manufacture a new fact, unsettling any simple binary between truth and lie.
We make sense of an ambiguous situation by weaving clues into a working narrative that can be revised.
#048 · Black Swan
An event that is highly unexpected from a given perspective, carries extreme consequences, and is made to seem predictable in hindsight.
Axis of tension
Sensemaking → Black Swan (in tension):
Taleb treats retrospective narrative-building as a cognitive fallacy to be eradicated, whereas Weick views retrospective framing as the inescapable condition of all human meaning and coordinated action.
Someone is unfairly discounted as a knower, or society lacks the shared concepts needed to articulate their experience.
#159 · Veil of ignorance
Rawls's thought experiment: we choose social principles without knowing what position we ourselves will occupy in society, a state of ignorance meant to foster impartiality.
Axis of tension
Veil of ignorance → Epistemic injustice (in tension):
The veil requires stripping away knowledge of one’s concrete social identity, whereas standpoint and epistemic-injustice theorists argue that situated knowledge from the margins is indispensable for recognizing structural harm.
Dominant institutions displace or invalidate the ways of knowing practiced by other communities.
#152 · Cancel culture
Collective public condemnation leads to boycotts, deplatforming, or professional and social ostracism.
Axis of tension
Epistemic violence → Cancel culture (in tension):
In contemporary campus and media controversies the phrase is sometimes deployed as a conversation-ending accusation, which critics cite as concept creep and defenders view as resistance to ongoing silencing.
Examines how the structure of evidence sharing among people shapes belief formation, truth discovery, and the risk of collective error.
#118 · RAG (Retrieval-Augmented Generation)
Model generation is grounded by first retrieving relevant passages from external knowledge sources.
Axis of tension
RAG (Retrieval-Augmented Generation) → Network epistemology (in tension):
An AI answer adorned with numbered source citations looks like transparent epistemic justification, yet the selection of which five passages the model saw was decided silently by an upstream retrieval ranking metric.
People publicly express preferences that differ from their private views under social pressure and fear of consequences.
#079 · Mechanism design
We design institutions and rules backward from a desired outcome so that self-interested participants are incentivized to bring it about.
Axis of tension
Mechanism design → Preference falsification (in tension):
Incentive compatibility assumes participants will reveal their true preferences once material payoffs align; Timur Kuran shows that social and reputational penalties often lie outside the designer’s payoff matrix.
People publicly express preferences that differ from their private views under social pressure and fear of consequences.
#139 · Affective polarization
Political disagreement hardens into personal hostility, distrust, and dislike toward members of the opposing camp.
Axis of tension
Affective polarization → Preference falsification (in tension):
Survey measures of partisan hostility can either exaggerate or understate real-world animus depending on whether expressing partisan hatred or bipartisan politeness carries a local social penalty.
People publicly express preferences that differ from their private views under social pressure and fear of consequences.
#153 · Culture wars
Political struggle centers on clashing moral norms, collective identities, and the symbolic meaning of public institutions.
Axis of tension
Culture wars → Preference falsification (in tension):
If a large share of public declarations in moral disputes reflects conformism toward one’s own political tribe, the true ideological depth of the “war” among ordinary citizens is overstated.
Before removing an existing rule or institution, first understand the problem it was built to solve.
#060 · Wicked problems
Complex social problems have contested definitions, conflicting goals, and interventions that reshape the problem itself.
Axis of tension
Wicked problems → Chesterton’s fence (in tension):
The property that “every wicked problem can be considered a symptom of another problem” undermines confidence that removing any single identified root cause will not unleash worse downstream troubles.
Before removing an existing rule or institution, first understand the problem it was built to solve.
#068 · Inversion
We tackle a problem backward by asking first what would guarantee failure and then systematically avoiding those causes.
Axis of tension
Inversion → Chesterton’s fence (in tension):
Inversion encourages stripping away obstacles and failure modes; Chesterton reminds us to understand why a barrier was erected before tearing it down.
Before removing an existing rule or institution, first understand the problem it was built to solve.
#069 · Decision reversibility
Some decisions can be reversed cheaply, while others carry lasting consequences and high costs of turning back.
Axis of tension
Decision reversibility → Chesterton’s fence (in tension):
Chesterton demands full understanding before removing a rule; two-way-door logic suggests trying the removal and walking back through the door if the cost of reversal is genuinely low.
Before removing an existing rule or institution, first understand the problem it was built to solve.
#076 · Skin in the game
A decision-maker personally bears part of the downside of their decisions rather than sharing only in the upside.
Axis of tension
Skin in the game → Chesterton’s fence (in tension):
Both concepts counsel respect for inherited rules, but Chesterton requires intellectual understanding before reform, whereas skin in the game allows action without full theoretical understanding so long as the actor bears the downside.
Before removing an existing rule or institution, first understand the problem it was built to solve.
#162 · Information hazards
Genuine, true information can cause harm merely by being disseminated or by enabling dangerous actors to cause damage.
Axis of tension
Information hazards → Chesterton’s fence (in tension):
Chesterton demands that we understand a norm before dismantling it; the norm of open science is such a fence, and Bostrom's concept asks whether part of it must be taken down.
Scientific claims require tests capable of refuting them, and research findings should be independently reproducible.
#155 · Matthew effect
Initial advantage attracts further resources, credit, and opportunity, compounding inequality over time.
Axis of tension
Falsifiability and the replication crisis → Matthew effect (in tension):
Academic reward structures that lavish grants, citations, and tenure on flashy positive findings and famous laboratories create incentives that make replication and falsification career-harming.
A filter bubble cuts off dissenting voices, whereas an echo chamber actively trains members to distrust outside sources.
#096 · Algorithmic management
Computational systems allocate tasks, monitor performance, and shape worker evaluation and discipline.
Axis of tension
Epistemic bubbles vs. echo chambers → Algorithmic management (in tension):
The filter bubble thesis blames the algorithm, whereas empirical studies indicate that user self-selection weighs more heavily than algorithmic personalization.
A filter bubble cuts off dissenting voices, whereas an echo chamber actively trains members to distrust outside sources.
#103 · Algorithmic amplification
Recommendation systems selectively boost the reach of certain items, shaping what audiences actually encounter.
Axis of tension
Algorithmic amplification → Epistemic bubbles vs. echo chambers (in tension):
Controlled experiments with reverse-chronological feeds show that algorithmic amplification does not always lock users inside a bubble; at times it even exposes them to more cross-cutting content than their self-selected social graph would.
Institutions simplify a complex reality into standardized records and categories to make it easier to govern.
#165 · Effective altruism
We use evidence and careful reasoning to compare causes and interventions so that limited resources do the most good possible.
Axis of tension
Effective altruism → Legibility (in tension):
Preferring quantifiable interventions favors whatever is legible to a spreadsheet at the expense of institutional and political transformation.
Institutions simplify a complex reality into standardized records and categories to make it easier to govern.
#184 · Tacit knowledge and Polanyi’s paradox
We know how to do and recognize far more than we can fully articulate in explicit rules, and applying rules itself relies on practical skill.
Axis of tension
Tacit knowledge and Polanyi’s paradox → Legibility (in tension):
The administrative drive of states or corporations toward total legibility and standardization destroys local practical knowledge (*mētis*), which by nature remains uncodified.
Institutions simplify a complex reality into standardized records and categories to make it easier to govern.
#189 · Spontaneous order and dispersed knowledge
Coherent social order can emerge from dispersed local knowledge and mutual adjustments within shared rules, without an omniscient central planner.
Axis of tension
Spontaneous order and dispersed knowledge → Legibility (in tension):
James C. Scott’s administrative legibility describes what happens when a state destroys a complex spontaneous order so that it fits into simplified bureaucratic tables.
Disruptions in the information ecosystem span unintended errors, deliberate fabrications, and the malicious use of genuine information.
#102 · Content moderation
Platforms establish and enforce rules governing what speech and content remain visible or permissible.
Axis of tension
Information disorder → Content moderation (in tension):
The category of mal-information (genuine information shared to cause harm) cannot be adjudicated by truth-based moderation rules and opens a fierce dispute over the limits of takedowns.
We identify the effects of changing a factor by separating causal impact from mere correlation.
#115 · Scaling laws
Model performance often follows empirical power-law relationships as training compute, dataset size, and parameter count grow.
Axis of tension
Causal inference → Scaling laws (in tension):
Scaling up model parameters and training corpora improves approximations of the joint observational distribution, but without structural assumptions it cannot climb to the higher rungs of the ladder of causation.
A record of how data originated and was transformed allows us to trace its sources and assess its reliability.
#104 · Data brokers
Companies collect, aggregate, and sell personal data profiles, typically without any direct relationship with the individuals tracked.
Axis of tension
Data provenance → Data brokers (in tension):
The commercial data-broker business model relies on laundering and obscuring where personal records were scraped or bought; strict provenance auditing threatens their margins.
False or misleading content is deliberately spread to influence or deceive an audience.
#037 · Doomscrolling
We compulsively scroll through disturbing news even though doing so worsens our mood and sense of agency.
Axis of tension
Doomscrolling → Disinformation (in tension):
The disinformation framework asks whether individual claims are true or false, whereas doomscrolling shows that a relentless feed of 100% factually accurate catastrophe reports can paralyze agency just as effectively.
False or misleading content is deliberately spread to influence or deceive an audience.
#090 · Dead internet theory
A view attributing a large share of online activity to bots and automated content, whose extreme version becomes an unsubstantiated conspiracy narrative.
Axis of tension
Dead internet theory → Disinformation (in tension):
Traditional disinformation analysis looks for a state actor and a political message, whereas dead internet theory highlights an impersonal economy of automated reach; the two lenses capture opposite ends of the same ecosystem.
False or misleading content is deliberately spread to influence or deceive an audience.
#138 · Overton window
The range of policies and ideas treated as politically acceptable in mainstream public debate at a given time.
Axis of tension
Overton window → Disinformation (in tension):
The post-2014 conspiratorial version of the “Overton window” (the cannibalism parable) is itself a textbook instrument of disinformation about how social norms and minority rights evolve.
When a measure becomes a target, improving the metric may stop reflecting genuine improvement in what it was meant to track.
#073 · Information asymmetry
One party to a transaction holds material information that the other party lacks.
Axis of tension
Information asymmetry → Goodhart’s law (in tension):
Signals and credentials introduced to overcome information asymmetry tend to become targets of optimization in their own right, eroding their informational value.
When a measure becomes a target, improving the metric may stop reflecting genuine improvement in what it was meant to track.
#078 · Nudge and behavioral economics
The way choices are arranged predictably steers behavior without forbidding any options or significantly changing economic incentives.
Axis of tension
Nudge and behavioral economics → Goodhart’s law (in tension):
A nudge optimized around a single visible metric—such as initial enrollment rate—can easily ignore whether the induced choice actually improved the chooser’s long-term welfare.
When a measure becomes a target, improving the metric may stop reflecting genuine improvement in what it was meant to track.
#123 · Model cards
Standardized documentation accompanying a model details its intended use, training data, evaluation results, and limitations.
Axis of tension
Model cards → Goodhart’s law (in tension):
Once filling out a model card becomes a checkbox for repository publication or legal compliance, template completion begins to displace the substantive disaggregated evaluation it was meant to record.
When a measure becomes a target, improving the metric may stop reflecting genuine improvement in what it was meant to track.
#126 · Alignment tax
Making an AI system safer and better aligned can incur extra compute, latency, or a trade-off in raw capability.
Axis of tension
Alignment tax → Goodhart’s law (in tension):
Techniques aimed at reducing the alignment tax optimize for benchmark proxies of alignment rather than alignment itself, so an apparently zero tax can be an artifact of shallow measurement.
When a measure becomes a target, improving the metric may stop reflecting genuine improvement in what it was meant to track.
#134 · Explainable AI (XAI)
Techniques and interfaces help people understand which inputs and mechanisms drove a model's output.
Axis of tension
Explainable AI (XAI) → Goodhart’s law (in tension):
Once “providing an explanation” becomes a statutory compliance checkbox, organizations optimize for generating plausible-looking charts rather than genuine human comprehension.
When a measure becomes a target, improving the metric may stop reflecting genuine improvement in what it was meant to track.
#172 · Net zero
Greenhouse gas emissions are reduced as far as possible, and any residual emissions are balanced by equivalent removals from the atmosphere.
Axis of tension
Net zero → Goodhart’s law (in tension):
Once net zero becomes a target metric, the balance can easily be achieved on paper (through cheap offsets or boundary shifting) without altering the physical flow of emissions.
A cognitive system predicts incoming sensory signals and updates its internal model based on prediction errors.
#121 · Model hallucinations
A generative model produces fluent, confident output that is fabricated or unfaithful to facts and source material.
Axis of tension
Predictive processing → Model hallucinations (in tension):
Both concepts describe generative models synthesizing outputs from top-down priors, but in the living brain “controlled hallucination” is continuously reined in by sensory prediction error, whereas an ungrounded language model lacks that world-corrective loop.
According to this thesis, tools that are reliably and continuously integrated into our thinking can form part of our cognitive system—just like memory or brain processes.
#087 · Attention economy
Scarce human attention becomes the bottleneck resource that media platforms and advertisers compete to capture and monetize.
Axis of tension
Extended mind → Attention economy (in tension):
The parity principle assumes a reliable, transparent resource waiting passively for the user's call, whereas the attention economy engineers the very same smartphones to interrupt, hijack, and monetize the user's focus.
We shift part of the work of memory and thinking onto external tools or bodily actions.
#112 · Large language models (LLMs)
Models trained on massive text corpora predict subsequent tokens in a sequence to generate and process natural language.
Axis of tension
Cognitive offloading → Large language models (LLMs) (in tension):
Large language models offload not merely factual recall or arithmetic, but drafting, synthesis, and multi-step reasoning—shifting the debate over the costs of offloading from memory to critical thought itself.
We compulsively scroll through disturbing news even though doing so worsens our mood and sense of agency.
#078 · Nudge and behavioral economics
The way choices are arranged predictably steers behavior without forbidding any options or significantly changing economic incentives.
Axis of tension
Doomscrolling → Nudge and behavioral economics (in tension):
Interface patterns that eliminate natural stopping cues (“sludge” and dark patterns) are the dark twin of behavioral nudging, sparking debate over whether responsibility lies with choice architecture or individual self-control.
Prolonged, unresolved workplace stress leads to exhaustion, detachment, and a reduced sense of efficacy.
#148 · Biopolitics
Modern power governs human life at the level of populations and bodies through public health, statistics, norms, and institutional regulation.
Axis of tension
Burnout → Biopolitics (in tension):
Classifying burnout in the WHO's ICD-11 sits on a biopolitical knife-edge: it can either empower labor inspectorates to police toxic workplaces or individualize structural exploitation into corporate mindfulness apps.
Entities so vast across time and space that we only ever encounter their local manifestations.
#044 · Systems thinking
We explain the behavior of a whole through the interdependencies, feedback loops, and delays among its parts.
Axis of tension
Hyperobjects → Systems thinking (in tension):
Systems thinking seeks a map of feedback loops; a hyperobject is what always leaks out of the map because it exceeds every scale at which we draw it.
Entities so vast across time and space that we only ever encounter their local manifestations.
#169 · Planetary boundaries
A scientific framework identifying key Earth-system processes and thresholds beyond which human perturbation risks destabilizing planetary conditions.
Axis of tension
Hyperobjects → Planetary boundaries (in tension):
Planetary boundaries assume that the Earth system can be measured and governed by thresholds; the hyperobject insists that data are merely footprints of the object, never the object itself.
We examine differences among ways worlds exist rather than assuming a single reality described by competing opinions.
#044 · Systems thinking
We explain the behavior of a whole through the interdependencies, feedback loops, and delays among its parts.
Axis of tension
Ontological turn → Systems thinking (in tension):
Systems thinking presupposes a single world described at different levels; the ontological turn questions the very unity of the world within which feedback loops are supposed to operate.
We examine differences among ways worlds exist rather than assuming a single reality described by competing opinions.
#101 · Philosophy of information
Investigates the nature of information, how it carries meaning, and its role in knowledge and accounts of reality.
Axis of tension
Philosophy of information → Ontological turn (in tension):
Floridi's informational realism (the world as the totality of informational objects) competes with the ontological turn over which ontology best describes digital and hybrid entities.
Interactions among simple parts give rise to collective patterns and properties not found in any single component.
#112 · Large language models (LLMs)
Models trained on massive text corpora predict subsequent tokens in a sequence to generate and process natural language.
Axis of tension
Large language models (LLMs) → Emergence (in tension):
Researchers remain divided over whether capabilities that appear abruptly in large models represent genuine phase transitions or artifacts of discontinuous evaluation metrics.
Interactions among simple parts give rise to collective patterns and properties not found in any single component.
#117 · In-context learning
A model infers how to perform a task from examples provided inside the prompt itself, without updating its parameters.
Axis of tension
In-context learning → Emergence (in tension):
The debate over whether in-context learning appears abruptly above a critical parameter threshold or grows smoothly along continuous metrics sits at the heart of the controversy over emergent abilities.
Interactions among simple parts give rise to collective patterns and properties not found in any single component.
#127 · Artificial general intelligence (AGI)
A hypothetical machine intelligence capable of learning, transferring, and performing across a broad range of cognitive tasks.
Axis of tension
Artificial general intelligence (AGI) → Emergence (in tension):
The debate over whether capabilities in large models are genuinely emergent is at bottom a debate over whether general intelligence will arrive abruptly or accumulate smoothly and predictably.
Certain systems gain from volatility, stressors, and shocks rather than merely withstanding them.
#085 · De-risking
We reduce excessive vulnerability to a risky partner or single source while maintaining broader trade and cooperation.
Axis of tension
De-risking → Antifragility (in tension):
De-risking seeks to reduce exposure to identified chokepoints, whereas Taleb's antifragility asks how systems can gain from disorder and views narrow defensive hedging as fragile.
We look for shared cognitive, cultural, and institutional mechanisms underlying multiple contemporary crises.
#109 · Technological solutionism
Complex social and political problems are recast as neatly bounded technical glitches solvable by software or hardware alone.
Axis of tension
Metacrisis → Technological solutionism (in tension):
Metacrisis posits that the incentives and mindset producing the crisis cannot solve it through another technical fix, directly challenging solutionist faith in the next technology.
We look for shared cognitive, cultural, and institutional mechanisms underlying multiple contemporary crises.
#163 · Existential risk
A threat capable of causing human extinction or permanently and drastically curtailing humanity's future potential.
Axis of tension
Metacrisis → Existential risk (in tension):
Existential risk researchers calculate probabilities for discrete catastrophic scenarios; metacrisis shifts focus to the underlying cultural and institutional conditions that generate those risks, at some cost to measurability.
Individually rational use of a shared, rivalrous resource can lead to its collective depletion.
#098 · Adversarial interoperability
New tools or services plug into an existing closed system without the dominant owner's permission.
Axis of tension
Adversarial interoperability → Tragedy of the commons (in tension):
Open interfaces create a shared technical commons; the recurring dispute is whether unpermissioned access leads to generative innovation or to spam, scraping abuse, and privacy breaches.
Individually rational use of a shared, rivalrous resource can lead to its collective depletion.
#158 · Rights of nature
Ecosystems, rivers, or species are granted legal personhood and standing to have their interests defended independently of property owners.
Axis of tension
Rights of nature → Tragedy of the commons (in tension):
Hardin saw only two remedies for resource depletion—privatization or state coercion; Rights of Nature introduces a third figure: the ecosystem itself as a legal party rather than a passive resource to be divided.
Individually rational use of a shared, rivalrous resource can lead to its collective depletion.
#181 · Climate justice
Examines how historical responsibility for emissions, vulnerability to climate harms, and voice in transition decisions are unevenly distributed.
Axis of tension
Climate justice → Tragedy of the commons (in tension):
The tragedy of the commons framing implies a symmetric coordination problem among equal users, whereas climate justice begins from the observation that users are unequal in both historical emissions and vulnerability.
For certain non-biological cultural artifacts, a long survival so far can indicate a longer expected remaining lifespan.
#155 · Matthew effect
Initial advantage attracts further resources, credit, and opportunity, compounding inequality over time.
Axis of tension
Matthew effect → Lindy effect (in tension):
Both concepts link future survival to past duration, but the Lindy effect attributes endurance to tested robustness, whereas the Matthew effect shows how rewards compound regardless of current intrinsic quality.
Improving the efficiency with which a resource is used can lower its effective cost and increase demand enough that total consumption rises.
#072 · Incentives
Rewards, penalties, and institutional consequences shape which actions people find worthwhile to pursue.
Axis of tension
Jevons paradox → Incentives (in tension):
Lowering the marginal cost of a service is a direct economic incentive to consume more of it; efficiency mandates inadvertently subsidize the very demand they aim to curb.
Early choices and switching costs lock in a trajectory of development even when better alternatives remain theoretically possible.
#075 · Creative destruction
Innovations create new capabilities and industries while simultaneously displacing older firms, skills, and ways of working.
Axis of tension
Creative destruction → Path dependence (in tension):
Creative destruction assumes that superior innovations displace incumbents; path dependence shows how inferior solutions persist through switching costs and coordination lock-in.
Early choices and switching costs lock in a trajectory of development even when better alternatives remain theoretically possible.
#144 · Thucydides’s Trap
The rise of an emerging power alarms an established ruling power, creating structural pressure that can escalate toward conflict.
Axis of tension
Thucydides’s Trap → Path dependence (in tension):
The claim that past power transitions predetermine today’s assumes a continuity of mechanism that historians do not find across the sixteen cases.
Costs or benefits of an economic activity affect third parties outside the transaction without being reflected in its price, such as pollution burdening a factory's neighbors.
#071 · Opportunity cost
The true cost of any choice includes the value of the best foregone alternative.
Axis of tension
Opportunity cost → Externalities (in tension):
Opportunity cost is reckoned from the decision-maker’s own perspective; costs shifted onto third parties do not enter that calculus unless institutions force them in.
Locally rational competition can trap all participants in an outcome that leaves everyone worse off.
#128 · Frontier AI
Highly capable general-purpose models at the cutting edge of performance introduce new capabilities and risks requiring dedicated evaluation.
Axis of tension
Frontier AI → Moloch and coordination failure (in tension):
Competitive racing among laboratories to stay at the frontier is a textbook coordination trap; frontier AI regulation attempts to halt that race, though critics argue it risks entrenching an oligopoly of incumbents.
Locally rational competition can trap all participants in an outcome that leaves everyone worse off.
#137 · Dual-use technology
The same scientific knowledge or technical capability can be applied to both beneficial civilian purposes and harmful or military ends.
Axis of tension
Dual-use technology → Moloch and coordination failure (in tension):
Unilateral restraint in dual-use research or export sales penalizes the actor who holds back unless all competitors coordinate—a coordination pact that commercial and geopolitical rivalry constantly erodes.
A state's ability to effectively implement policies, deliver public services, and enforce rules.
#077 · Rent-seeking
An actor increases income by capturing privileges or controlling access rather than by creating new wealth.
Axis of tension
Rent-seeking → State capacity (in tension):
Early public-choice theory treated the state primarily as a dispenser of rents, whereas state-capacity research shows that a professional, autonomous civil service is often a prerequisite for resisting rent-seeking capture.
A state's ability to effectively implement policies, deliver public services, and enforce rules.
#141 · Hypernormalization
Faithfully repeating official formulas and rituals becomes more important than whether they match reality, and reproducing the form keeps the system running.
Axis of tension
State capacity → Hypernormalization (in tension):
A regime can maintain the theatrical appearance of omnipresent control even as its actual administrative capacity hollows out from within.
Interactions can create mutual gains or mutual losses, so one party's gain need not equal the other's loss.
#082 · Geoeconomic fragmentation
Global economic integration splinters into rival blocs shaped by geopolitical tensions and strategic alignment.
Axis of tension
Non-zero-sum games → Geoeconomic fragmentation (in tension):
The unraveling of global interdependence after 2018 shows that a non-zero-sum game can be played as zero-sum when security outweighs joint gain.
Interactions can create mutual gains or mutual losses, so one party's gain need not equal the other's loss.
#085 · De-risking
We reduce excessive vulnerability to a risky partner or single source while maintaining broader trade and cooperation.
Axis of tension
De-risking → Non-zero-sum games (in tension):
Classical trade theory treated economic interdependence as a positive-sum game; de-risking starts from the premise that asymmetric nodes in that game can be turned into weapons of coercion.
Interactions can create mutual gains or mutual losses, so one party's gain need not equal the other's loss.
#157 · Dark forest theory
A speculative hypothesis suggesting cosmic civilizations hide their presence out of fear that detection by another intelligence could mean destruction.
Axis of tension
Dark forest theory → Non-zero-sum games (in tension):
The theory assumes that civilizations with unverifiable intentions cannot establish a positive-sum game because the cost of a single mistake is existential; critics reply that this is a built-in premise rather than a derived conclusion.
A supply chain maintains or recovers operations through buffers, diversified sourcing, and adaptive capacity.
#192 · Weaponized interdependence
States that control central hubs in global information and financial networks can monitor flows and selectively choke off access as a tool of coercion.
Axis of tension
Weaponized interdependence → Supply chain resilience (in tension):
Optimizing supply chains strictly for lowest unit cost naturally concentrates production in single global hubs, which from a security perspective become dangerous chokepoints.
Innovations create new capabilities and industries while simultaneously displacing older firms, skills, and ways of working.
#168 · Degrowth
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
Axis of tension
Creative destruction → Degrowth (in tension):
The Schumpeterian growth model justifies the social pain of destruction by the aggregate growth it generates; degrowth questions the desirability of that aggregate goal itself.
A decision-maker personally bears part of the downside of their decisions rather than sharing only in the upside.
#197 · Banality of evil and the problem of many hands
Thoughtless compliance with routine procedures and the fragmentation of action across many narrow roles can produce grave collective harm while diffusing personal responsibility.
Axis of tension
Banality of evil and the problem of many hands → Skin in the game (in tension):
Taleb's principle demands a direct link between decision-making and personal risk, whereas the bureaucratic problem of many hands institutionalizes total asymmetry between agency and bearing consequences.
The way choices are arranged predictably steers behavior without forbidding any options or significantly changing economic incentives.
#088 · Surveillance capitalism
An economic model built on capturing behavioral data, selling predictions, and attempting to shape people's future actions.
Axis of tension
Surveillance capitalism → Nudge and behavioral economics (in tension):
Zuboff treats behavioral “nudging” as an instrument of covert behavioral modification for corporate profit, whereas Richard Thaler and Cass Sunstein defend it as libertarian paternalism designed to serve the chooser's own stated interests.
We combine digital technology with social diversity to support collaboration across difference without forcing uniformity.
#097 · Technofeudalism
According to this contested thesis, control over dominant digital platforms and cloud infrastructure allows owners to extract rent akin to feudal dependency.
Axis of tension
Plurality → Technofeudalism (in tension):
Both diagnoses see centralized digital platforms as a new concentration of power, but Plurality seeks to dismantle it through open civic protocols rather than state nationalization or antitrust breakups alone.
We combine digital technology with social diversity to support collaboration across difference without forcing uniformity.
#109 · Technological solutionism
Complex social and political problems are recast as neatly bounded technical glitches solvable by software or hardware alone.
Axis of tension
Plurality → Technological solutionism (in tension):
Critics read Plurality as a sophisticated civic variant of technological solutionism; its authors reply that they are redesigning democratic institutions rather than peddling standalone apps.
We evaluate an organization's operations and risks through environmental, social, and corporate governance criteria.
#158 · Rights of nature
Ecosystems, rivers, or species are granted legal personhood and standing to have their interests defended independently of property owners.
Axis of tension
Rights of nature → ESG (in tension):
Corporate ESG reporting keeps nature in the role of a financial risk factor for investors, whereas Rights of Nature reverses the vector: the investor becomes a legally actionable risk to the river.
We evaluate an organization's operations and risks through environmental, social, and corporate governance criteria.
#170 · Doughnut economics
An economic compass aiming to meet every person's essential social foundation without overshooting Earth's ecological ceiling.
Axis of tension
Doughnut economics → ESG (in tension):
The Doughnut is increasingly adapted as a corporate sustainability reporting template, even though its author warns against reducing it to a greenwashing brand.
Plausible risks of serious or irreversible harm can justify protective action even before full scientific certainty is achieved.
#165 · Effective altruism
We use evidence and careful reasoning to compare causes and interventions so that limited resources do the most good possible.
Axis of tension
Precautionary principle → Effective altruism (in tension):
Expected-value maximization assumes that probabilities can be meaningfully assigned; the precautionary principle begins precisely where that assumption breaks down.
An actor retains the capacity to govern its own digital infrastructure, data, and technological rules.
#094 · Datafication
More and more aspects of everyday life and activity are turned into quantified data amenable to recording and analysis.
Axis of tension
Datafication → Digital sovereignty (in tension):
Political battles over who holds jurisdiction over citizen and industrial data are struggles over who controls the downstream value of datafication.
An actor retains the capacity to govern its own digital infrastructure, data, and technological rules.
#104 · Data brokers
Companies collect, aggregate, and sell personal data profiles, typically without any direct relationship with the individuals tracked.
Axis of tension
Data brokers → Digital sovereignty (in tension):
The struggle over data brokers exposes the limits of state sovereignty: citizens' personal data circulate on a global market where the government is simultaneously the regulator and one of the industry's largest paying customers.
An actor retains the capacity to govern its own digital infrastructure, data, and technological rules.
#109 · Technological solutionism
Complex social and political problems are recast as neatly bounded technical glitches solvable by software or hardware alone.
Axis of tension
Digital sovereignty → Technological solutionism (in tension):
Flagship initiatives like Gaia-X are frequently criticized as technocratic fixes to a problem that is fundamentally political, fiscal, and regulatory.
An economic model built on capturing behavioral data, selling predictions, and attempting to shape people's future actions.
#125 · Privacy-preserving machine learning
Machine learning and analytical techniques are designed to extract patterns while limiting what can be inferred about any individual record.
Axis of tension
Privacy-preserving machine learning → Surveillance capitalism (in tension):
PPML techniques allow technology platforms to claim they “never see” raw user data while continuing to extract behavioral models from edge devices—a move critics view as technical legitimation of extraction.
An audience member develops a one-sided sense of intimacy with a media figure who does not know them personally.
#113 · Generative AI
Machine learning systems trained on patterns in existing data produce new text, images, audio, or code.
Axis of tension
Parasocial relationships → Generative AI (in tension):
AI companion chatbots talk back, remember personal details, and adapt to the user in real time—breaking Horton and Wohl's foundational criterion of strict non-reciprocity.
An audience member develops a one-sided sense of intimacy with a media figure who does not know them personally.
#179 · Deepfake
Synthetic audio or video realistically fabricates a specific person saying or doing something they never actually said or did.
Axis of tension
Parasocial relationships → Deepfake (in tension):
Synthetic personas and voice clones hijack pre-existing parasocial trust to extract money, intimate disclosures, or political support.
According to this contested thesis, control over dominant digital platforms and cloud infrastructure allows owners to extract rent akin to feudal dependency.
#151 · Late capitalism
A critical diagnosis of contemporary capitalism emphasizing financialization, pervasive commodification, and cultural contradictions in everyday life.
Axis of tension
Technofeudalism → Late capitalism (in tension):
Both concepts diagnose the contemporary political economy, but “late capitalism” insists on the unbroken continuity of capitalist accumulation laws, whereas technofeudalism proclaims a structural rupture.
Cultural patterns spread across minds and media through imitation, mutation, and differential selection.
#143 · Cultural hegemony
A ruling order maintains power in part because its worldview comes to be accepted across society as natural common sense.
Axis of tension
Memetics → Cultural hegemony (in tension):
Antonio Gramsci's theory locates cultural power in class interests and institutional apparatuses, whereas Dawkinsian memetics attributes agency to self-replicating ideas themselves—risking blindness to who owns the media infrastructure.
Platforms establish and enforce rules governing what speech and content remain visible or permissible.
#119 · RLHF
Human comparisons and preference judgments are used to fine-tune a model toward more helpful, aligned responses.
Axis of tension
RLHF → Content moderation (in tension):
Data annotators ranking model outputs in RLHF perform labor structurally parallel to commercial content moderation: high throughput, low pay, exposure to toxic prompts, and decisive influence over global speech norms.
Platforms establish and enforce rules governing what speech and content remain visible or permissible.
#138 · Overton window
The range of policies and ideas treated as politically acceptable in mainstream public debate at a given time.
Axis of tension
Content moderation → Overton window (in tension):
Platform rules set the boundaries of what can be said in public, but they do so through administrative enforcement rather than the social debate assumed by the window metaphor.
Companies collect, aggregate, and sell personal data profiles, typically without any direct relationship with the individuals tracked.
#125 · Privacy-preserving machine learning
Machine learning and analytical techniques are designed to extract patterns while limiting what can be inferred about any individual record.
Axis of tension
Privacy-preserving machine learning → Data brokers (in tension):
Mathematical guarantees on a model’s output say nothing about how the underlying data was acquired on the brokerage market before training began.
Complex social and political problems are recast as neatly bounded technical glitches solvable by software or hardware alone.
#133 · Algorithmic bias
A computational system systematically and unfairly differentiates outcomes to the disadvantage of certain individuals or groups.
Axis of tension
Algorithmic bias → Technological solutionism (in tension):
The promise that a biased social allocation can be “debiased” simply by adding a mathematical fairness constraint to the loss function reproduces the illusion that structural injustice has a purely engineering fix.
Complex social and political problems are recast as neatly bounded technical glitches solvable by software or hardware alone.
#164 · Longtermism
An ethical stance holding that safeguarding the far future and the vast number of potential future people is a key moral priority today.
Axis of tension
Longtermism → Technological solutionism (in tension):
Critics read longtermism as solutionism scaled up to millions of years: present problems are to be solved by future technologies, whose risks are to be mitigated by still more technology.
Complex social and political problems are recast as neatly bounded technical glitches solvable by software or hardware alone.
#173 · Carbon dioxide removal (CDR)
Carbon dioxide is captured from the ambient air and durably stored in geological, terrestrial, or ocean reservoirs.
Axis of tension
Carbon dioxide removal (CDR) → Technological solutionism (in tension):
Critics read CDR as a technological promise that postpones institutional and consumption change; proponents reply that without removal the carbon budget simply cannot close.
Complex social and political problems are recast as neatly bounded technical glitches solvable by software or hardware alone.
#175 · Abundance agenda
A policy focus on expanding institutional capacity and removing regulatory bottlenecks so society can build ample housing, clean energy, and public goods.
Axis of tension
Abundance agenda → Technological solutionism (in tension):
Critics read the agenda as a belief that technology and physical supply can dissolve distributive conflicts; its authors reply that the bottlenecks they target are institutional, not technical.
Complex social and political problems are recast as neatly bounded technical glitches solvable by software or hardware alone.
#177 · Game B
A proposed, hypothetical cooperative order aimed at overcoming destructive rivalry and the traps of existing institutions—a direction of inquiry rather than a finished blueprint.
Axis of tension
Game B → Technological solutionism (in tension):
Critics such as Douglas Rushkoff see in Game B another iteration of the tech-founder dream of re-engineering human society from first principles, despite its professed hostility to complicated systems.
Model performance often follows empirical power-law relationships as training compute, dataset size, and parameter count grow.
#176 · Amara’s law
We tend to overestimate the impact of a new technology in the short run and underestimate its transformative effect in the long run.
Axis of tension
Amara’s law → Scaling laws (in tension):
Empirical scaling laws in language models promise smooth, predictable capability gains, whereas Amara's law warns that short-term societal impact forecasts systematically overshoot.
We work to ensure that AI systems pursue goals and behaviors faithful to human intentions and values.
#161 · Ethics of technology
We evaluate how technologies distribute goods, harms, and possibilities for action, as well as who bears responsibility for their design and use.
Axis of tension
Ethics of technology → AI alignment (in tension):
Alignment treats ethics as a goal specification to be hard-coded into a system, whereas the ethics of technology asks who writes that specification and in whose interest.
Artificially generated records are used in place of or alongside real-world observations to train, test, or analyze models.
#179 · Deepfake
Synthetic audio or video realistically fabricates a specific person saying or doing something they never actually said or did.
Axis of tension
Synthetic data → Deepfake (in tension):
The exact same class of generative architectures produces synthetic training sets for research and synthetic media for deception; the difference lies in intent, labeling, and governance rather than mechanism.
A hypothetical machine intelligence capable of learning, transferring, and performing across a broad range of cognitive tasks.
#149 · Posthumanism
A current of thought questioning the image of the human as the autonomous center of the world, emphasizing that our existence and agency arise through relations with other people, organisms, and technologies.
Axis of tension
Posthumanism → Artificial general intelligence (AGI) (in tension):
The debate over artificial general intelligence revives precisely what critical posthumanism rejected: the fantasy of mind as a disembodied informational pattern that can be uploaded into a machine.
A hypothetical machine intelligence capable of learning, transferring, and performing across a broad range of cognitive tasks.
#171 · Fermi paradox
The tension between the expectation that other civilizations should exist in a vast, ancient cosmos and the absence of confirmed traces of their presence.
Axis of tension
Fermi paradox → Artificial general intelligence (AGI) (in tension):
Scenarios of machine expansion sharpen the paradox (self-replicating von Neumann probes should already be everywhere) while simultaneously supplying a prime candidate for the Great Filter.
A cultural sensibility that oscillates between postmodern irony and renewed sincerity without resolving the tension between them.
#191 · Bullshit (indifference to truth)
Speech produced solely for impression or effect, with complete indifference to whether it is true or false—distinct from a deliberate lie, which still tracks the truth in order to invert it.
Axis of tension
Bullshit (indifference to truth) → Metamodernism (in tension):
Metamodernism’s oscillation between irony and new sincerity seeks to recover earnest judgment in a culture numbed by decades of advertising and political bullshit.
Faithfully repeating official formulas and rituals becomes more important than whether they match reality, and reproducing the form keeps the system running.
#178 · Liminal thinking
During transitions or uncertainty, we examine the boundaries of our own assumptions to open up new ways of understanding and acting.
Axis of tension
Liminal thinking → Hypernormalization (in tension):
Hypernormalization and Szakolczai's “permanent liminality” describe a collective state of suspended transition in which the threshold leads to no new viable order—a dark case Gray's optimism does not address.
Modern power governs human life at the level of populations and bodies through public health, statistics, norms, and institutional regulation.
#167 · Anthropocene
Human activity has become a dominant force reshaping Earth's climate, ecosystems, and material strata on a planetary scale.
Axis of tension
Biopolitics → Anthropocene (in tension):
When the stakes shift from the human population to the biosphere as a whole, biopolitics must decide whether it expands into “geopower” or reaches the limit of its object.
Currents of thought arguing that technological and capitalist processes should be intensified and sped up to force a radical transformation.
#198 · Imperative of responsibility and heuristics of fear
Because modern technology can irreversibly alter the conditions of future life, we bear an asymmetric obligation to give serious weight to credible threats of catastrophe before acting.
Axis of tension
Imperative of responsibility and heuristics of fear → Accelerationism (in tension):
Accelerationism, especially in its unconditional variant (e/acc), represents a programmatic rejection of the heuristics of fear, treating the acceleration of technocapitalism as an imperative overriding catastrophic risk.
A critical diagnosis of contemporary capitalism emphasizing financialization, pervasive commodification, and cultural contradictions in everyday life.
#167 · Anthropocene
Human activity has become a dominant force reshaping Earth's climate, ecosystems, and material strata on a planetary scale.
Axis of tension
Anthropocene → Late capitalism (in tension):
The debate over the Capitalocene asks whether the driver of the new epoch is the human species as a whole or a historically specific system of capital accumulation built on fossil fuels.
A critical diagnosis of contemporary capitalism emphasizing financialization, pervasive commodification, and cultural contradictions in everyday life.
#168 · Degrowth
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
Axis of tension
Late capitalism → Degrowth (in tension):
Degrowth accepts that the current stage is ecologically unsustainable, but rejects the Marxist teleology latent in the adjective “late.”
Rawls's thought experiment: we choose social principles without knowing what position we ourselves will occupy in society, a state of ignorance meant to foster impartiality.
#164 · Longtermism
An ethical stance holding that safeguarding the far future and the vast number of potential future people is a key moral priority today.
Axis of tension
Veil of ignorance → Longtermism (in tension):
Rawls’s “just savings principle” between generations is far more restrained than longtermist maximization of future value, though both ask what happens when future generations are placed behind the veil.
Rawls's thought experiment: we choose social principles without knowing what position we ourselves will occupy in society, a state of ignorance meant to foster impartiality.
#190 · Ergodicity and the ruin barrier
The average outcome across a group at one time need not match the trajectory of a single participant over time, especially when sequential risks include irreversible ruin.
Axis of tension
Ergodicity and the ruin barrier → Veil of ignorance (in tension):
Choosing principles of justice behind Rawls’s veil of ignorance via the maximin rule is often criticized by utilitarians as overly risk-averse, yet in a non-ergodic world it rationally shields agents from irreversible ruin.
One ethical tradition evaluates moral character and cultivated virtues, while the other judges actions primarily by their consequences.
#164 · Longtermism
An ethical stance holding that safeguarding the far future and the vast number of potential future people is a key moral priority today.
Axis of tension
Virtue ethics vs. consequentialism → Longtermism (in tension):
Longtermism pushes consequentialist aggregation to the point where astronomical numbers of hypothetical future lives swamp obligations to living persons; virtue ethics diagnoses this not as an arithmetic error, but as a distortion of practical wisdom.
Genuine, true information can cause harm merely by being disseminated or by enabling dangerous actors to cause damage.
#188 · Ketman and living in truth
Under ideological pressure, people may conceal private disbelief behind ritual public conformity—or refuse to play along and 'live in truth' despite the cost.
Axis of tension
Ketman and living in truth → Information hazards (in tension):
Havel’s imperative to speak the truth publicly regardless of consequences collides in technology ethics with questions of hazardous knowledge whose open dissemination can trigger catastrophic harm.
A policy focus on expanding institutional capacity and removing regulatory bottlenecks so society can build ample housing, clean energy, and public goods.
#189 · Spontaneous order and dispersed knowledge
Coherent social order can emerge from dispersed local knowledge and mutual adjustments within shared rules, without an omniscient central planner.
Axis of tension
Spontaneous order and dispersed knowledge → Abundance agenda (in tension):
The contemporary abundance agenda seeks a balance between unlocking decentralized supply (as in urban housing) and deploying deliberate state capacity for large-scale infrastructure.
Examines how historical responsibility for emissions, vulnerability to climate harms, and voice in transition decisions are unevenly distributed.
#196 · Non-identity problem
Because major policy choices alter which specific individuals will ever be conceived, depleted-future scenarios may leave future people with no better personal alternative—challenging person-affecting ethics.
Axis of tension
Non-identity problem → Climate justice (in tension):
The non-identity argument challenges the intuitive framing of climate change as a direct harm inflicted on the very same people who would have lived in a low-carbon world, forcing climate ethicists to reformulate the concept of harm.
#186 · The Great Transformation and fictitious commodities
Treating labor, land, and money as pure market commodities produced for sale strains the social and natural fabric, provoking protective counter-movements.
#189 · Spontaneous order and dispersed knowledge
Coherent social order can emerge from dispersed local knowledge and mutual adjustments within shared rules, without an omniscient central planner.
Axis of tension
The Great Transformation and fictitious commodities → Spontaneous order and dispersed knowledge (in tension):
Published in the same year (1944), the books of Polanyi and Hayek define a central twentieth-century axis of debate: for Hayek the free market emerges spontaneously from below, whereas for Polanyi laissez-faire was artificially engineered by state intervention.
When an organization deteriorates, members can respond by leaving (exit) or speaking up to reform it (voice), while loyalty shapes whether they stay and push for change.
#200 · Umwelt and autopoiesis
A living organism continually regenerates the network of processes and boundaries that constitute it (autopoiesis), while encountering its surroundings as a species-specific world of meaningful cues (Umwelt).
Axis of tension
Exit, voice, and loyalty → Umwelt and autopoiesis (in tension):
Tightly closed, self-reproducing organizational systems (autopoietic systems) often filter out internal critical “voice” as meaningless noise until mass “exit” starves them of resources from their environment.
We acknowledge the limits of our knowledge and remain open to revising our beliefs when new evidence appears.
#004 · First principles thinking
We break a problem down to its foundational assumptions and build a solution up from there.
Axis of tension
First principles thinking → Epistemic humility (in tension):
The rhetoric of first principles encourages confidence that one has hit undeniable bedrock, whereas the fallibilism of Neurath and Popper reminds us that where we stop digging is a practical decision.
We acknowledge the limits of our knowledge and remain open to revising our beliefs when new evidence appears.
#025 · Prebunking
Exposing people in advance to a manipulation technique helps them recognize and resist it when they encounter it later.
Axis of tension
Prebunking → Epistemic humility (in tension):
Who decides which rhetorical moves count as “manipulation techniques” and which count as legitimate persuasion? Inoculation can breed blanket cynicism instead of careful judgment.
Before criticizing an argument, we reconstruct its strongest, most charitable version.
#006 · Brandolini’s law
Refuting falsehood typically requires far more effort than producing it.
Axis of tension
Brandolini’s law → Steelmanning (in tension):
Reconstructing an interlocutor's claim in its strongest form demands a cognitive investment that Brandolini's law warns us to ration carefully.
We break a problem down to its foundational assumptions and build a solution up from there.
#022 · Legibility
Institutions simplify a complex reality into standardized records and categories to make it easier to govern.
Axis of tension
First principles thinking → Legibility (in tension):
High-modernist projects that redesign cities or bureaucracies “from first principles” in James C. Scott's sense routinely destroy informal local knowledge that decomposition fails to register.
We evaluate evidence in a way that protects a desired conclusion instead of testing it impartially.
#018 · Framing
How the same fact is presented shapes how people interpret it and what they decide.
Axis of tension
Framing → Motivated reasoning (in tension):
The classic framing effect operates without prior partisan motivation on randomly assigned descriptions, whereas motivated reasoning explains why actors actively seek out a frame that flatters their pre-existing commitments.
We evaluate evidence in a way that protects a desired conclusion instead of testing it impartially.
#025 · Prebunking
Exposing people in advance to a manipulation technique helps them recognize and resist it when they encounter it later.
Axis of tension
Prebunking → Motivated reasoning (in tension):
Inoculation theory assumes that people do not want to be manipulated; motivated reasoning shows that partisans often welcome manipulation that flatters their own side.
Refuting falsehood typically requires far more effort than producing it.
#026 · Metascience
We study the methods, practices, and institutions of science to evaluate their reliability and improve how research works.
Axis of tension
Metascience → Brandolini’s law (in tension):
Metascience exposes cheap, low-powered studies that are costly to refute, yet large-scale replication projects themselves require massive resources to verify each disputed claim.
Before attributing harmful outcomes to malice, consider error or incompetence.
#029 · Disinformation
False or misleading content is deliberately spread to influence or deceive an audience.
Axis of tension
Hanlon’s razor → Disinformation (in tension):
An adversary who knows Hanlon's razor designs covert operations to look like accidents, while the disinformation frame is conversely overused to attribute grand design to ordinary chaos.
Examines how the structure of evidence sharing among people shapes belief formation, truth discovery, and the risk of collective error.
#017 · Preference falsification
People publicly express preferences that differ from their private views under social pressure and fear of consequences.
Axis of tension
Network epistemology → Preference falsification (in tension):
In Timur Kuran's model people hide their true beliefs under social pressure; in Zollman-style network models agents update and report their beliefs with complete honesty, yet collective error emerges anyway.
A filter bubble cuts off dissenting voices, whereas an echo chamber actively trains members to distrust outside sources.
#023 · Context collapse
A single utterance reaches multiple distinct audiences at once, stripping the sender of the ability to separate contexts and tailor the message to each.
Axis of tension
Context collapse → Epistemic bubbles vs. echo chambers (in tension):
An echo chamber presupposes a homogeneous audience, whereas context collapse presupposes an excessively heterogeneous one; both can occur on the same platform at different scales.
According to this thesis, tools that are reliably and continuously integrated into our thinking can form part of our cognitive system—just like memory or brain processes.
#035 · Metacognition
We monitor our own understanding and confidence, then adjust how we learn or think.
Axis of tension
Extended mind → Metacognition (in tension):
Once part of one's beliefs resides in a notebook, search engine, or AI assistant, monitoring what one actually knows requires auditing external systems that lack internal feelings of fluency or uncertainty.
We explain the behavior of a whole through the interdependencies, feedback loops, and delays among its parts.
#060 · Wicked problems
Complex social problems have contested definitions, conflicting goals, and interventions that reshape the problem itself.
Axis of tension
Wicked problems → Systems thinking (in tension):
Rittel and Webber wrote their 1973 paper as a direct reckoning with the 1960s “systems approach,” which promised that urban and social problems could be solved by aerospace-style operations research.
Interactions among simple parts give rise to collective patterns and properties not found in any single component.
#047 · Antifragility
Certain systems gain from volatility, stressors, and shocks rather than merely withstanding them.
Axis of tension
Emergence → Antifragility (in tension):
Antifragility assumes that systems gain from disorder, yet emergent collective properties can just as easily collapse discontinuously past a threshold.
Certain systems gain from volatility, stressors, and shocks rather than merely withstanding them.
#055 · Path dependence
Early choices and switching costs lock in a trajectory of development even when better alternatives remain theoretically possible.
Axis of tension
Path dependence → Antifragility (in tension):
Taleb praises systems that gain from shocks, whereas a path-dependent system locked into a trajectory loses its capacity to adapt because the cost of switching rises over time.
Certain systems gain from volatility, stressors, and shocks rather than merely withstanding them.
#061 · Feedback loops
The outputs of a process feed back into the system, either amplifying or dampening its subsequent behavior.
Axis of tension
Feedback loops → Antifragility (in tension):
Taleb treats stabilizing negative loops with suspicion: suppressing small fluctuations accumulates the risk of a major collapse.
Certain systems gain from volatility, stressors, and shocks rather than merely withstanding them.
#063 · Network theory
Describes systems as nodes and links and studies how their topology shapes flows, contagion, and overall resilience.
Axis of tension
Network theory → Antifragility (in tension):
Scale-free networks are resilient to random failures yet fragile to targeted attacks on hubs, complicating any simple opposition between fragile and antifragile.
An event that is highly unexpected from a given perspective, carries extreme consequences, and is made to seem predictable in hindsight.
#056 · Second-order effects
The immediate effect of an action triggers further downstream consequences through the responses of people and systems.
Axis of tension
Second-order effects → Black Swan (in tension):
Taleb rejects the ambition to trace long chains of downstream consequences and advises building structural robustness against them, whereas Bastiat and Hazlitt assume secondary effects can be deduced a priori.
An event that is highly unexpected from a given perspective, carries extreme consequences, and is made to seem predictable in hindsight.
#065 · Climate tipping points
Crossing a critical threshold can trigger a self-sustaining shift in a component of the climate system.
Axis of tension
Climate tipping points → Black Swan (in tension):
Crossing a threshold is often portrayed as a bolt from the blue, whereas the tipping-points literature argues that thresholds can be mapped in advance and are therefore not Black Swans.
A state's ability to effectively implement policies, deliver public services, and enforce rules.
#060 · Wicked problems
Complex social problems have contested definitions, conflicting goals, and interventions that reshape the problem itself.
Axis of tension
State capacity → Wicked problems (in tension):
State capacity measures the execution of tasks with a well-defined objective, whereas wicked problems lack a stable problem definition or an agreed stopping rule.
We preserve the ability to capture future upside while keeping the cost of walking away bounded.
#071 · Opportunity cost
The true cost of any choice includes the value of the best foregone alternative.
Axis of tension
Optionality → Opportunity cost (in tension):
Keeping options open carries its own opportunity cost: capital, attention, and commitment held in reserve are not compounding inside any single chosen path.
Innovations create new capabilities and industries while simultaneously displacing older firms, skills, and ways of working.
#077 · Rent-seeking
An actor increases income by capturing privileges or controlling access rather than by creating new wealth.
Axis of tension
Creative destruction → Rent-seeking (in tension):
Incumbent firms often defend themselves against creative destruction not by innovating, but through lobbying, patent thickets, and regulatory barriers, stalling the Schumpeterian engine.
A decision-maker personally bears part of the downside of their decisions rather than sharing only in the upside.
#185 · Soft budget constraint and shortage economy
When an organization expects repeated bailouts, its cost discipline weakens, fueling excess demand for inputs and chronic shortages.
Axis of tension
Soft budget constraint and shortage economy → Skin in the game (in tension):
The essence of the soft budget constraint is the severing of risk symmetry: the prestige and upside of expansion remain inside the organization, while losses are socialized onto taxpayers or creditors.
An actor increases income by capturing privileges or controlling access rather than by creating new wealth.
#084 · Industrial policy
The state deliberately directs resources and rules to build up selected productive capabilities and economic sectors.
Axis of tension
Industrial policy → Rent-seeking (in tension):
The classic public-choice critique: targeted sectoral support attracts political lobbying and turns market discipline into a race for bureaucratic favor.
We evaluate an organization's operations and risks through environmental, social, and corporate governance criteria.
#082 · Geoeconomic fragmentation
Global economic integration splinters into rival blocs shaped by geopolitical tensions and strategic alignment.
Axis of tension
ESG → Geoeconomic fragmentation (in tension):
The divergence of reporting regimes (the EU's CSRD, the anti-ESG backlash in US states, ISSB standards) is a textbook case of rules fragmenting along political lines.
We evaluate an organization's operations and risks through environmental, social, and corporate governance criteria.
#083 · Precautionary principle
Plausible risks of serious or irreversible harm can justify protective action even before full scientific certainty is achieved.
Axis of tension
ESG → Precautionary principle (in tension):
ESG promises to manage uncertainty through measurement and pricing; the precautionary principle holds that where harms are irreversible, measurement cannot replace a decision to refrain.
Scarce human attention becomes the bottleneck resource that media platforms and advertisers compete to capture and monetize.
#097 · Technofeudalism
According to this contested thesis, control over dominant digital platforms and cloud infrastructure allows owners to extract rent akin to feudal dependency.
Axis of tension
Attention economy → Technofeudalism (in tension):
If dominant platforms extract rent for market access rather than merely selling attention on a competitive market, the metaphor of an “attention market” obscures the shift from merchant to feudal lord.
Scarce human attention becomes the bottleneck resource that media platforms and advertisers compete to capture and monetize.
#108 · Loneliness epidemic
Loneliness is the painful subjective feeling of lacking connection, whereas social isolation is the objective scarcity of social ties.
Axis of tension
Loneliness epidemic → Attention economy (in tension):
Digital platforms monetize the waking hours that once went into face-to-face relationships while promising frictionless connection; the core dispute is whether the network connects or displaces.
An economic model built on capturing behavioral data, selling predictions, and attempting to shape people's future actions.
#097 · Technofeudalism
According to this contested thesis, control over dominant digital platforms and cloud infrastructure allows owners to extract rent akin to feudal dependency.
Axis of tension
Technofeudalism → Surveillance capitalism (in tension):
Shoshana Zuboff analyzes behavioral data extraction as a rogue mutation within capitalism, whereas Yanis Varoufakis contends that the same infrastructural extraction has killed capitalism and replaced it with a rentier order; the dispute concerns the diagnosis of the macro-system rather than the description of platform practices.
Precarization → Surveillance capitalism (in tension):
Zuboff analyzes the extraction of behavioral surplus from consumers rather than labor from workers; digital platforms fuse both processes, yet their mechanisms and victims remain distinct.
An economic model built on capturing behavioral data, selling predictions, and attempting to shape people's future actions.
#108 · Loneliness epidemic
Loneliness is the painful subjective feeling of lacking connection, whereas social isolation is the objective scarcity of social ties.
Axis of tension
Loneliness epidemic → Surveillance capitalism (in tension):
A business model built on behavioral data extraction rewards prolonged screen engagement rather than reciprocal human bonding; user loneliness functions for it as a monetizable resource rather than a cost.
A platform first attracts users and business partners, then progressively degrades service for both to extract maximum value for itself.
#105 · Creator economy
Independent creators earn a living from their audience relationships through ads, subscriptions, patronage, or direct sales.
Axis of tension
Creator economy → Enshittification (in tension):
Platforms initially woo creators with generous revenue splits and creator funds, then tighten payouts and organic reach once creators and their audiences are locked in.
Digital platforms become the underlying infrastructure organizing markets, labor, and social interaction.
#106 · Social atomization
Intermediate social ties, civic associations, and local communities weaken, leaving individuals more isolated even within large societies.
Axis of tension
Social atomization → Platformization (in tension):
Digital platforms promise universal connectivity, yet replace intermediate civic bodies with an unmediated relationship between the isolated user and corporate infrastructure.
Computational systems allocate tasks, monitor performance, and shape worker evaluation and discipline.
#097 · Technofeudalism
According to this contested thesis, control over dominant digital platforms and cloud infrastructure allows owners to extract rent akin to feudal dependency.
Axis of tension
Algorithmic management → Technofeudalism (in tension):
The technofeudal thesis casts the platform worker as a digital serf paying rent, whereas algorithmic-management scholarship analyzes them as a worker subjected to a hyper-rationalized new form of managerial control.
We study how digital signs, icons, and interface conventions generate meaning and guide interpretation and action.
#100 · Memetics
Cultural patterns spread across minds and media through imitation, mutation, and differential selection.
Axis of tension
Digital semiotics → Memetics (in tension):
Memetics models the spread of online content as the quasi-biological replication of discrete units, whereas digital semiotics analyzes how signs are continuously reinterpreted and translated across cultural boundaries by users and algorithms.
Loneliness is the painful subjective feeling of lacking connection, whereas social isolation is the objective scarcity of social ties.
#109 · Technological solutionism
Complex social and political problems are recast as neatly bounded technical glitches solvable by software or hardware alone.
Axis of tension
Loneliness epidemic → Technological solutionism (in tension):
Apps designed to “cure loneliness” and AI companions treat the absence of human bonds as a technical glitch to be patched rather than an outcome of how work, cities, and time are organized.
A neural network architecture that uses self-attention mechanisms to weigh relationships among elements across a sequence.
#134 · Explainable AI (XAI)
Techniques and interfaces help people understand which inputs and mechanisms drove a model's output.
Axis of tension
Transformer → Explainable AI (XAI) (in tension):
Attention weights tempt researchers as a ready-made visual explanation of model reasoning, yet the debate over whether “attention is explanation” remains sharply contested.
A large model trained on broad data at scale can be adapted to a wide variety of downstream tasks.
#128 · Frontier AI
Highly capable general-purpose models at the cutting edge of performance introduce new capabilities and risks requiring dedicated evaluation.
Axis of tension
Foundation models → Frontier AI (in tension):
“Frontier” defines models by the outer edge of capability and catastrophic risk, whereas “foundation” defines them by their infrastructural role in a downstream ecosystem; public policy frequently conflates the two.
Models trained on massive text corpora predict subsequent tokens in a sequence to generate and process natural language.
#125 · Privacy-preserving machine learning
Machine learning and analytical techniques are designed to extract patterns while limiting what can be inferred about any individual record.
Axis of tension
Privacy-preserving machine learning → Large language models (LLMs) (in tension):
Large language models memorize and regurgitate verbatim training sequences, turning training-data extraction attacks into the primary stress test for PPML at scale.
Machine learning systems trained on patterns in existing data produce new text, images, audio, or code.
#130 · Model collapse
Repeatedly training new generations of models on insufficiently curated synthetic output can progressively degrade diversity and tail accuracy.
Axis of tension
Generative AI → Model collapse (in tension):
Flooding the internet with machine-generated outputs contaminates future training corpora, meaning that the commercial triumph of generative AI threatens to erode its own statistical foundation.
An AI system processes, aligns, or generates multiple types of data—such as text, images, and audio—within a shared framework.
#121 · Model hallucinations
A generative model produces fluent, confident output that is fabricated or unfaithful to facts and source material.
Axis of tension
Multimodality → Model hallucinations (in tension):
Proponents of multimodality hoped that linking words to visual images would ground language in physical reality, yet vision-language models routinely “hallucinate” objects that are not in the picture when language priors override visual tokens.
We craft instructions, context, and output constraints to elicit more reliable and useful responses from a generative model.
#119 · RLHF
Human comparisons and preference judgments are used to fine-tune a model toward more helpful, aligned responses.
Axis of tension
Prompt engineering → RLHF (in tension):
Instruction tuning and RLHF internalize much of the scaffolding that early prompt engineers had to supply manually, while simultaneously changing the behavioral priors to which prompts must be tuned.
Human comparisons and preference judgments are used to fine-tune a model toward more helpful, aligned responses.
#129 · Prompt injection
Adversarial instructions hidden inside untrusted data trick a model into overriding its original instructions.
Axis of tension
Prompt injection → RLHF (in tension):
Training a model to be maximally helpful and obedient to natural-language instructions simultaneously increases its propensity to obey malicious instructions hidden inside untrusted data.
Generative models learn to synthesize data by iteratively reversing a process that gradually adds noise.
#130 · Model collapse
Repeatedly training new generations of models on insufficiently curated synthetic output can progressively degrade diversity and tail accuracy.
Axis of tension
Diffusion models → Model collapse (in tension):
Diffusion-generated images are scraped back into subsequent web-scale training sets, raising the question of how generative models degrade when trained on their own outputs.
Standardized documentation accompanying a model details its intended use, training data, evaluation results, and limitations.
#124 · Synthetic data
Artificially generated records are used in place of or alongside real-world observations to train, test, or analyze models.
Axis of tension
Model cards → Synthetic data (in tension):
Models trained or fine-tuned on outputs generated by other models inherit their undocumented blind spots in ways that standard single-artifact model cards fail to capture.
A hypothetical machine intelligence capable of learning, transferring, and performing across a broad range of cognitive tasks.
#194 · The bitter lesson and Moravec’s paradox
General methods that scale with compute—search and learning—repeatedly outperform hand-crafted human domain knowledge, even as everyday sensorimotor skills prove harder to automate than abstract reasoning.
Axis of tension
The bitter lesson and Moravec’s paradox → Artificial general intelligence (AGI) (in tension):
Today’s debate over the path to AGI turns on the intersection of the two concepts: whether scaling compute alone (the bitter lesson) will finally conquer embodied perception and physical dexterity (Moravec’s paradox).
A cultural sensibility that oscillates between postmodern irony and renewed sincerity without resolving the tension between them.
#141 · Hypernormalization
Faithfully repeating official formulas and rituals becomes more important than whether they match reality, and reproducing the form keeps the system running.
Axis of tension
Metamodernism → Hypernormalization (in tension):
Hypernormalization describes a society that cynically performs a hollow official script nobody believes in anymore; metamodernism describes the inverse movement: consciously choosing to act “as if” meaning and progress were possible despite knowing how easily earnestness can fail.
Faithfully repeating official formulas and rituals becomes more important than whether they match reality, and reproducing the form keeps the system running.
#156 · Hyperstition
Fictions, forecasts, or speculative narratives shape human behavior in ways that cause them to materialize as reality.
Axis of tension
Hyperstition → Hypernormalization (in tension):
Hypernormalization is a fiction in which nobody believes yet everyone performs; hyperstition is a fiction that works precisely because actors invest belief and capital in it.
Treating social success as purely a reflection of individual merit breeds hubris among winners and humiliation among those left behind.
#154 · DEI
Organizational frameworks aimed at broadening representation, removing unfair barriers, and ensuring full participation across diverse groups.
Axis of tension
DEI → Tyranny of merit (in tension):
The battle over DEI is largely a battle over whether conventional “merit” measured by standardized tests and credentials is neutral or reproduces inherited starting advantages.
A ruling order maintains power in part because its worldview comes to be accepted across society as natural common sense.
#152 · Cancel culture
Collective public condemnation leads to boycotts, deplatforming, or professional and social ostracism.
Axis of tension
Cultural hegemony → Cancel culture (in tension):
The dispute over whether online ostracism is an instrument of cultural-elite hegemony or grassroots resistance against it shows how the concept can be inverted in either direction.
The rise of an emerging power alarms an established ruling power, creating structural pressure that can escalate toward conflict.
#156 · Hyperstition
Fictions, forecasts, or speculative narratives shape human behavior in ways that cause them to materialize as reality.
Axis of tension
Thucydides’s Trap → Hyperstition (in tension):
Critics argue that the narrative of inevitable war increases its likelihood by shaping the expectations of decision-makers on both sides.
Overlapping axes of social inequality interact to produce distinct lived experiences that cannot be understood by examining each axis in isolation.
#153 · Culture wars
Political struggle centers on clashing moral norms, collective identities, and the symbolic meaning of public institutions.
Axis of tension
Intersectionality → Culture wars (in tension):
In cultural conflicts the term functions as a tribal shibboleth on both sides, detached from its original evidentiary problem in antidiscrimination law.
Examines how legal doctrines and institutions can reproduce racial disparities even through formally race-neutral rules.
#153 · Culture wars
Political struggle centers on clashing moral norms, collective identities, and the symbolic meaning of public institutions.
Axis of tension
Critical race theory (CRT) → Culture wars (in tension):
Since 2020 the acronym CRT has functioned in public debate as a catch-all label for any discussion of race in schools or workplaces, detaching it from the corpus of legal scholarship.
Examines how legal doctrines and institutions can reproduce racial disparities even through formally race-neutral rules.
#154 · DEI
Organizational frameworks aimed at broadening representation, removing unfair barriers, and ensuring full participation across diverse groups.
Axis of tension
Critical race theory (CRT) → DEI (in tension):
Corporate diversity trainings are frequently conflated with CRT, even though most make no reference to legal theory and do not share its skepticism toward liberal reform.
Modern power governs human life at the level of populations and bodies through public health, statistics, norms, and institutional regulation.
#149 · Posthumanism
A current of thought questioning the image of the human as the autonomous center of the world, emphasizing that our existence and agency arise through relations with other people, organisms, and technologies.
Axis of tension
Biopolitics → Posthumanism (in tension):
Biopolitics analyzes power over the biological life of the human species, whereas posthumanism challenges the very human/non-human boundary on which that power operates.
Rawls's thought experiment: we choose social principles without knowing what position we ourselves will occupy in society, a state of ignorance meant to foster impartiality.
#196 · Non-identity problem
Because major policy choices alter which specific individuals will ever be conceived, depleted-future scenarios may leave future people with no better personal alternative—challenging person-affecting ethics.
Axis of tension
Non-identity problem → Veil of ignorance (in tension):
Rawlsian contractualism struggles to place future generations behind the veil of ignorance when the very identity and number of the contracting parties depend on the principles chosen.
An ethical stance holding that safeguarding the far future and the vast number of potential future people is a key moral priority today.
#168 · Degrowth
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
Axis of tension
Longtermism → Degrowth (in tension):
Longtermism typically treats technological and economic growth as a precondition for realizing humanity's future potential, whereas degrowth sees growth itself as the primary threat.
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
#172 · Net zero
Greenhouse gas emissions are reduced as far as possible, and any residual emissions are balanced by equivalent removals from the atmosphere.
Axis of tension
Net zero → Degrowth (in tension):
Degrowth scholars charge that mainstream net-zero models presuppose perpetual GDP growth and are therefore forced to pencil in vast, unproven negative emissions in the second half of the century.
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
#177 · Game B
A proposed, hypothetical cooperative order aimed at overcoming destructive rivalry and the traps of existing institutions—a direction of inquiry rather than a finished blueprint.
Axis of tension
Game B → Degrowth (in tension):
Both frameworks insist on living within planetary boundaries, but Game B sidesteps the political economy of wealth redistribution and state regulation that is central to degrowth.
A planned reduction of energy and material throughput in wealthy economies aimed at ecological sustainability, human well-being, and equity.
#181 · Climate justice
Examines how historical responsibility for emissions, vulnerability to climate harms, and voice in transition decisions are unevenly distributed.
Axis of tension
Climate justice → Degrowth (in tension):
Degrowth accepts the argument about luxury emissions, but rejects the assumption held by most climate justice theories that decarbonization can be reconciled with continued economic growth in wealthy countries.
Greenhouse gas emissions are reduced as far as possible, and any residual emissions are balanced by equivalent removals from the atmosphere.
#181 · Climate justice
Examines how historical responsibility for emissions, vulnerability to climate harms, and voice in transition decisions are unevenly distributed.
Axis of tension
Net zero → Climate justice (in tension):
A global net-zero target around 2050 does not specify who reaches zero when; dividing the remaining carbon budget is a question of political justice, not atmospheric physics.
Greenhouse gas emissions are reduced as far as possible, and any residual emissions are balanced by equivalent removals from the atmosphere.
#182 · Loss and damage (L&D)
Climate harms that persist beyond what mitigation and adaptation can prevent, raising questions of compensation, support, and historical responsibility.
Axis of tension
Loss and damage (L&D) → Net zero (in tension):
The slower net emissions fall, the larger the share of the climate agenda consumed by the bill for harms already inflicted, shifting the balance from prevention to repair.
Carbon dioxide is captured from the ambient air and durably stored in geological, terrestrial, or ocean reservoirs.
#174 · Just transition
Shifting toward a low-carbon economy while protecting, retraining, and involving the workers and communities bearing the costs of change.
Axis of tension
Carbon dioxide removal (CDR) → Just transition (in tension):
Global integrated assessment models routinely locate BECCS and afforestation on land in the Global South, shifting land and food-price burdens onto populations that emitted the least.