Read by dependency
Begin with claim discipline and formal representation; later essays assume those distinctions rather than restating them.
Complete research collection
Human–Machine Intelligence Convergence
A single, dependency-ordered route from the mathematics of inference to the politics of cognitive sovereignty. The collection keeps competence, agency, selfhood, consciousness and sentience distinct, then asks where they genuinely meet. Each argument ends in a decision, an intervention or a test.
Read by problem
Search titles and filter by meaningful topics when a live research or architecture decision matters more than sequence.
Read against the claim
Every essay distinguishes observation, mechanism, rival explanation and the decision the evidence can actually support.
Browse the complete research routeResearch home
Five words people confuse: intelligence, consciousness, sentience, selfhood and agency
A five-axis method for separating claims about intelligence, consciousness, sentience, selfhood and agency, with an executable headline-claim classifier.
Consciousness scienceCausality and interventionEvidence and provenanceInquiry and epistemic disciplineThree ledgers: function, mechanism and experience
A practical method for stopping claims about what an AI system does from being mistaken for claims about how it works or what it may experience.
Causality and interventionConsciousness scienceEvidence and provenanceInquiry and epistemic disciplineFrom observation to explanation: correlation, causation and constitution
A first-principles guide to separating association, causal intervention and constitutive explanation, with an executable depth-matched intervention audit.
Causality and interventionResearch methodsRisk and regulationInquiry and epistemic disciplineWhat counts as evidence for another mind?
Why testimony gains or loses evidential force through causal coupling, calibration, elicitation, provenance, source independence and tests against rival generators.
Causality and interventionEvidence and provenanceConsciousness scienceInquiry and epistemic disciplineThought experiments as instruments, not oracles
A practical method for using zombies, Mary, teletransporters and Chinese rooms to expose hidden premises, bridge principles and countermodels without treating intuition as a verdict.
Causality and interventionResearch methodsConsciousness scienceInquiry and epistemic disciplineLevels, boundaries and the grain of explanation
A practical method for choosing whether an explanation belongs at the level of a component, organism, loop, dyad, team or institution.
Causality and interventionResearch methodsReliability and recoveryInquiry and epistemic disciplineFirst-person, interpersonal and third-person evidence
A practical method for making reports, elicitation, behaviour, physiology and mechanism constrain one another without treating them as interchangeable evidence.
Research methodsEvidence and provenanceCausality and interventionInquiry and epistemic disciplineHow to read a frontier claim
A practical method for dissecting frontier research claims through four separate fields: source, assumptions, measurement bridge and decision consequence.
Research methodsCausality and interventionEvidence and provenanceMathematical language and representationNumber, measurement and representation
A first-principles guide to what changes when a property becomes a quantity, coordinate, code or probability, with a practical decision-invariance audit.
RepresentationResearch methodsRisk and regulationMathematical language and representationFunctions and compositional behaviour
A first-principles explanation of why modern models are best understood as ordered, routed compositions of transformations rather than inventories of parameters, with an executable function-composition lab.
Transformer mechanismsCausality and interventionRepresentationMathematical language and representationSets, logic and proof
A first-principles guide to how membership, relations, quantifiers, validity, necessity, proof and counterexamples constrain claims in mathematics, research and intelligent systems.
Reasoning and planningFinancial servicesResearch libraryMathematical language and representationCounting and combinatorial explosion
Why small independent choices create search spaces that defeat enumeration, and how pruning, bounds, symmetry, state merging and approximation make useful search possible.
Reasoning and planningPlatform economicsReliability and recoveryMathematical language and representationVectors, dot products and representation spaces
A first-principles guide to vectors, bases, dot products, cosine similarity and projection, with a practical test for deciding when geometry supports a representation claim.
RepresentationCausality and interventionResearch methodsMathematical language and representationMatrices, transformations and eigenstructure
How repeated linear transformation selects invariant modes, why eigenvectors reveal persistent behaviour, and when non-normal dynamics make the spectrum unsafe to trust alone.
RepresentationCausality and interventionProduction systemsMathematical language and representationLow rank, singular value decomposition and adaptation
A first-principles guide to when singular value decomposition, low-rank approximation and low-rank adapters preserve the work a transformation actually performs.
RepresentationLearning systemsProduction systemsMathematical language and representationPosition from rotation: complex phases, Fourier structure and RoPE
A first-principles explanation of how rotary position embedding turns absolute token indices into relative attention geometry, why its frequency bank matters, and how to audit it before extending context length.
RepresentationTransformer mechanismsCausality and interventionInference, dynamics and causalityCalculus, chain rules and credit assignment
How the chain rule turns local sensitivities into distant learning signals, why gradients vanish or mislead, and how to audit credit before acting on it.
Learning systemsCausality and interventionFinancial servicesInference, dynamics and causalityOptimisation, approximation and guarantees
Why convergence to a chosen objective is not yet a useful solution, and how to compose optimisation, estimation, approximation and deployment bounds into a decision-relevant guarantee.
Learning systemsResearch methodsPlatform economicsInference, dynamics and causalityProbability and Bayesian model comparison
A first-principles guide to Bayesian model comparison, Bayes factors, marginal likelihood, predictive checking and a rival-first evidence contract with an executable notebook.
Evidence and provenanceCausality and interventionResearch methodsInference, dynamics and causalityEstimation, Fisher information and calibration
Why an estimate deserves confidence only when the experiment separates relevant alternatives and the uncertainty procedure remains calibrated where it is used.
Research methodsProduction systemsCausality and interventionInference, dynamics and causalityInformation, entropy, maximum entropy and compression
A first-principles guide to what entropy, mutual information, maximum entropy, synergy and rate-distortion measure, and why none alone establishes meaning or experience.
Causality and interventionResearch libraryReliability and recoveryInference, dynamics and causalityVariational inference and the ELBO
How the evidence lower bound turns intractable Bayesian posterior inference into optimisation, why the projection creates approximation debt, and how to test whether that debt changes a decision.
Learning systemsRepresentationCausality and interventionInference, dynamics and causalityDynamical systems, attractors and bifurcations
A first-principles guide to deciding when trajectories, perturbation recovery and bifurcations explain cognition better than symbolic labels alone.
Causality and interventionReliability and recoveryResearch methodsInference, dynamics and causalityCausality, intervention and counterfactual worlds
Why prediction fails under action, how interventions replace causal mechanisms, and what stronger assumptions counterfactual histories require.
Causality and interventionResearch methodsRepresentationComputation, control and emergenceWhat is computation?
A first-principles test for deciding when a physical process implements a computation rather than merely matching one observed trace.
Causality and interventionResearch methodsSecurity and containmentComputation, control and emergenceState, memory and the anatomy of an algorithm
A first-principles account of what must persist for a process to have history, how minimal state compresses the past, and how to test state with causal interventions.
Causality and interventionMemory systemsReliability and recoveryComputation, control and emergenceMultiple realisation without implementation magic
A practical account of when different physical systems run the same computation, based on counterfactual transitions, causal interventions, explicit boundaries and a realiser-substitution lab.
Causality and interventionProduction systemsReliability and recoveryComputation, control and emergenceFeedback, control and cybernetics
A first-principles account of how goals, observation, error, action, disturbance and delay form a closed loop, with a runnable controller laboratory.
Reliability and recoveryCausality and interventionProduction systemsComputation, control and emergenceNetworks, topology and collective dynamics
Why identical nodes can produce different collective behaviour when only their connections change, with an executable rewiring and synchronisation laboratory.
Agent architectureCausality and interventionReliability and recoveryComputation, control and emergenceThe thermodynamics of prediction and computation
Why memory, erasure, prediction, finite-time control and error correction have different thermodynamic costs, with a practical information-discard audit and executable calculator.
Production systemsReliability and recoveryCausality and interventionComputation, control and emergenceSelf-organisation, autopoiesis and operational closure
A first-principles test for distinguishing systems that regenerate their own organisation from systems whose apparent autonomy is maintained by an outside caretaker.
Production systemsCausality and interventionReliability and recoveryComputation, control and emergenceEmergence and the ruler that creates it
A first-principles method for deciding whether an apparent capability jump belongs to a system, its behavioural distribution, or the measurement ruler used to report it.
Causality and interventionRisk and regulationResearch methodsLearning systems from first principlesLearning, generalisation and the shape of a task
Why a model can pass a random test split yet fail when the world changes, and how to design a train-test-shift contract that exposes the rule it learned.
Causality and interventionLearning systemsProduction systemsLearning systems from first principlesThe artificial neuron, backpropagation and credit assignment
A first-principles explanation of how backpropagation assigns path-conditioned credit, why gradients fail across depth, and how to audit a hand-built neural network.
Learning systemsCausality and interventionFinancial servicesLearning systems from first principlesRegularisation, normalisation and architectural bias
Why constraints on scale, capacity and connectivity change what a neural network can learn, with a controlled inductive-bias experiment and a practical constraint passport.
Causality and interventionPlatform economicsLearning systemsLearning systems from first principlesEmbeddings and latent geometry
A first-principles guide to when useful latent structure emerges, why probe accuracy can overstate mechanism, and how to test causal use with selective interventions.
RepresentationCausality and interventionReliability and recoveryLearning systems from first principlesSelf-supervision, generative learning and compression
Why self-supervised prediction learns some reusable structure, misses other structure, and how to audit the shortcuts created by masks, horizons, augmentations and decoders.
Causality and interventionLearning systemsRepresentationLearning systems from first principlesReinforcement learning, preference and control
A first-principles account of how reward, policy, exploration and preference learning produce controlled behaviour without, by themselves, establishing desire or felt value.
Learning systemsCausality and interventionResearch methodsLearning systems from first principlesUncertainty, calibration and the right to abstain
Why confidence is not permission, how calibrated abstention should be tied to decision loss and fallback capacity, and how to build a gate that fails safely under shift.
Reliability and recoveryProduction systemsPlatform economicsLearning systems from first principlesTraining provenance and evaluation validity
Why benchmark scores are evidence about configured training and evaluation lineages, and how to audit the capability claims they can support.
Evidence and provenanceResearch methodsCausality and interventionTransformer anatomyAttention from first principles
A first-principles explanation of query, key and value attention, why square-root scaling matters, how temperature controls routing, and how to test the mechanism yourself.
Transformer mechanismsCausality and interventionRepresentationTransformer anatomyMany heads, many relations
A causal method for distinguishing specialised, redundant and distributed attention-head computation using routing, writing, replacement and coalition tests.
Causality and interventionTransformer mechanismsReliability and recoveryTransformer anatomyThe block and the residual stream
A first-principles account of how transformer attention, feed-forward transformation, normalisation and skip paths compose computation through a shared residual state, with a causal audit and executable lab.
Transformer mechanismsCausality and interventionRepresentationTransformer anatomyPosition, rotation and relative order
A first-principles account of absolute, relative and rotary position encodings, the sequence symmetries they preserve, why length extrapolation fails, and how to test a positional contract before deployment.
RepresentationCausality and interventionTransformer mechanismsTransformer anatomyFeed-forward networks as conditional memory
A first-principles account of when transformer feed-forward sublayers deserve to be called conditional memory, with causal tests, boundary conditions and an executable edit lab.
Causality and interventionTransformer mechanismsMemory systemsTransformer anatomyTokenisation and the worlds it makes visible
How token choice redistributes context, statistical evidence and computational difficulty before a transformer begins to reason, with an executable tokenizer stress suite.
Transformer mechanismsRepresentationCausality and interventionTransformer anatomyTraining a language model end to end
A first-principles account of how data, objective, optimiser, scale and measurement jointly create a language model, with a reproducible miniature training lab.
Learning systemsCausality and interventionPlatform economicsTransformer anatomyLong context, efficient attention and persistent state
A first-principles guide to choosing attention, retrieval, recurrent compression and explicit state by the future questions a system must preserve.
Transformer mechanismsMemory systemsReliability and recoveryFrontier model behaviour and interpretabilityScaling laws, test-time compute and apparent emergence
A first-principles account of why capability cliffs can be created by metrics, multi-step tasks and test-time search, with an executable audit for locating the real source of a jump.
Causality and interventionProduction systemsLearning systemsFrontier model behaviour and interpretabilityDecoding: how sampling governs expressed cognition
Why temperature, truncation and search causally reshape what a language model expresses, with an executable decoding laboratory and a practical policy contract.
Transformer mechanismsCausality and interventionProduction systemsFrontier model behaviour and interpretabilityIn-context learning and temporary adaptation
A first-principles account of how prompts create temporary task states, and how interventions can distinguish retrieval, task inference and implicit optimisation.
Learning systemsCausality and interventionTransformer mechanismsFrontier model behaviour and interpretabilityReasoning traces, chain of thought and hidden computation
A causal method for deciding when a model's articulated rationale tracks the computation that produced its answer, with an executable TRACE audit.
Causality and interventionEvidence and provenanceResearch methodsFrontier model behaviour and interpretabilityMultimodality, embodiment and grounding by proxy
A first-principles account of what perception, action and feedback add beyond symbol correlations, with a causal grounding test and an executable intervention lab.
Causality and interventionEvidence and provenancePlatform economicsFrontier model behaviour and interpretabilityCircuits, causal tracing and intervention science
A first-principles guide to when activation patterns, causal traces and circuit diagrams warrant a genuine mechanism claim, with an executable intervention lab and evidence passport.
Causality and interventionTransformer mechanismsResearch methodsFrontier model behaviour and interpretabilitySuperposition, sparse features and the limits of interpretation
A first-principles guide to deciding when sparse autoencoder latents are merely useful coordinates and when causal, stability and coverage tests justify treating them as model features.
Causality and interventionRepresentationReliability and recoveryFrontier model behaviour and interpretabilityBenchmarks, capabilities and construct validity
A first-principles method for deciding whether an AI benchmark score measures the intended capability, a shortcut, or the interface to the test.
Causality and interventionResearch methodsIdentity and authorityAgents, world models and multi-agent systemsWorld models that survive intervention
Why predictive accuracy is not enough for a world model, and how to test whether learned dynamics preserve causal structure when actions, mechanisms, sensors and environments change.
Causality and interventionProduction systemsLearning systemsAgents, world models and multi-agent systemsFrom model to system: scaffolding, memory, tools and agents
A first-principles method for deciding whether an apparent agent capability belongs to model weights, runtime, context, orchestration, tools or the wider environment.
Agent architectureMemory systemsIdentity and authorityAgents, world models and multi-agent systemsReasoning, planning and tool use under uncertainty
A causal test for deciding when generated reasoning genuinely guides tool-using action, when it merely narrates a choice, and how to instrument the difference.
Causality and interventionReasoning and planningEvidence and provenanceAgents, world models and multi-agent systemsWorld state is not memory
Why reliable agents must keep checkpoints, episodes, knowledge, live world state and evidence ledgers distinct, with a typed state inventory and executable authority gate.
Memory systemsTemporal systemsEvidence and provenanceAgents, world models and multi-agent systemsActive inference as a design hypothesis for agents
A first-principles test of when expected free energy improves artificial agents, how it differs from reward-based control, and where the biological analogy stops.
Causality and interventionResearch methodsAgent architectureAgents, world models and multi-agent systemsWhen multi-agent systems are justified
A first-principles test for deciding when several agents create real value, when one agent with tools is enough, and how to measure the coordination break-even point.
Agent architectureIdentity and authorityEvidence and provenanceAgents, world models and multi-agent systemsChoosing agent topology from task structure
A practical method for compiling hierarchy, market, debate, swarm and pipeline agent topologies from task dependencies and authority boundaries.
Agent architectureIdentity and authoritySecurity and containmentAgents, world models and multi-agent systemsAgent plus operational twin
A first-principles architecture for letting an AI agent rehearse actions in a stateful operational twin while authoritative systems retain control, evidence and recovery.
Production systemsEvidence and provenanceIdentity and authorityBrain, body and biological intelligenceThe neuron from membrane to computation
A first-principles account of how membrane conductances, voltage-dependent state, dendritic geometry and spike trajectories differ from a scalar artificial activation, with an executable Hodgkin-Huxley conductance lab.
Causality and interventionResearch methodsSecurity and containmentBrain, body and biological intelligenceNeural codes, synapses and plasticity
A first-principles account of why neural codes only become explanatory when activity, downstream readout and plasticity are tested together.
Causality and interventionResearch methodsLearning systemsBrain, body and biological intelligenceBrain architecture as a changing network
Why the brain's functional architecture changes without rewiring: a first-principles account of hierarchy, recurrence, oscillatory timing, synchrony and neuromodulatory gain, with an executable network lab.
Causality and interventionResearch methodsEvidence and provenanceBrain, body and biological intelligenceReward prediction error and dopamine
A first-principles account of when dopamine signals support temporal-difference reward prediction error, and how to separate value error from movement, salience and measurement effects.
Causality and interventionResearch methodsLearning systemsBrain, body and biological intelligenceAttention under biological scarcity
A first-principles account of how biased competition, precision, arousal and neuromodulatory gain allocate finite biological processing, with an executable causal test.
Causality and interventionTransformer mechanismsReliability and recoveryBrain, body and biological intelligencePerception as inference, action and correction
Why perception is a closed-loop hypothesis test in which expectations propose causes, precision-weighted evidence corrects them, and action determines what can be learned next.
Causality and interventionResearch methodsRepresentationBrain, body and biological intelligenceMemory from synapse to autobiography
How working, episodic, semantic, procedural and autobiographical memory repeatedly re-encode one another, creating durable meaning, skilled action and a coherent but revisable life story.
Memory systemsCausality and interventionResearch methodsBrain, body and biological intelligenceWhy biological intelligence has stakes
A first-principles account of how embodiment, viability control, interoception and valence make outcomes matter to biological intelligence, with an executable control lab.
Consciousness scienceCausality and interventionLearning systemsHuman intelligence and its facultiesThe human intelligence stack
Human intelligence is a task-specific feasible set created by coupled bodily, cognitive, social, linguistic and cultural constraints.
Research methodsCausality and interventionReasoning and planningHuman intelligence and its facultiesMind before language
A causal account of thought without fluent speech, using evidence from infants, animals, imagery and aphasia, plus an executable non-verbal task battery.
Causality and interventionResearch methodsRisk and regulationHuman intelligence and its facultiesLearning and development, twice
Why equal test scores do not show equal learning: a causal comparison of play, attachment, curriculum, scaffolding and machine adaptation.
Causality and interventionLearning systemsResearch methodsHuman intelligence and its facultiesThe social origins of mind
A first-principles account of how joint attention, imitation, norms, language and institutions make cognition publicly corrigible, with an executable social-learning lab and developmental protocol.
Research methodsCausality and interventionLearning systemsHuman intelligence and its facultiesSelf-models, symbols and editable identity
When self-models improve control, how symbols make identity editable, and why recursive self-description can become confabulation.
Identity and authorityCausality and interventionSecurity and containmentHuman intelligence and its facultiesMetacognition, confabulation and disciplined self-report
A first-principles account of why calibrated confidence can coexist with unreliable explanations of action, with a dual-ledger protocol and an executable metacognitive calibration lab.
Causality and interventionResearch methodsEvidence and provenanceHuman intelligence and its facultiesMotivation, goals and veto architecture
Why action depends on a timed contest between drives, habits, goals and vetoes, and how to redesign that contest before willpower is needed.
Causality and interventionReliability and recoveryIdentity and authorityHuman intelligence and its facultiesOffline cognition, dreaming and consolidation
A first-principles account of how selective replay, recombination and governed slow updates let minds and machines learn when action stops.
Causality and interventionLearning systemsProduction systemsWestern and Islamic foundations of mindAristotle for artificial intelligence
A first-principles account of how matter, form, efficient cause, final cause, potentiality and actuality expose what model-only explanations miss in AI systems.
Causality and interventionProduction systemsEvidence and provenanceWestern and Islamic foundations of mindIslamic cognitive science and the internal senses
Avicenna's internal senses reveal how perception becomes unified, retained, recombined and appraised before explicit reasoning, while the flying person isolates a different boundary: primitive self-awareness.
Causality and interventionEvidence and provenanceRisk and regulationWestern and Islamic foundations of mindFrom Descartes to Kant: the constructed subject
A first-principles argument showing how the modern subject shifted from thinking substance to the synthesis that makes representations one ordered experience, with an executable assay for artificial systems.
Causality and interventionIdentity and authorityEvidence and provenanceWestern and Islamic foundations of mindThe mind-body problem in one serious sitting
A first-principles guide to the mind-body problem that compares rival theories through redescription, intervention and subject-continuity tests, with an executable bridge-contract workbook.
Causality and interventionPhilosophy of mindConsciousness scienceWestern and Islamic foundations of mindFunctionalism and the limits of multiple realisability
Why matching outputs is too weak for a functional theory of mind, and how intervention-closed role tests make multiple realisability a disciplined claim.
Causality and interventionConsciousness sciencePhilosophy of mindWestern and Islamic foundations of mindSyntax, meaning and the grounding problem
A first-principles account of when symbols become about the world, why correlation and multimodality are insufficient, and how to test grounding through counterfactual action and correction.
Causality and interventionEvidence and provenanceReliability and recoveryWestern and Islamic foundations of mindSubjectivity as data
A first-principles account of what first-person character contributes to consciousness science, with an executable Perspective Preservation Audit.
Consciousness scienceResearch methodsCausality and interventionWestern and Islamic foundations of mindAccess, phenomenality and the report boundary
Why report is a causal measurement transformation rather than a transparent window onto experience, and how to test the boundary with matched report, no-report and channel-substitution conditions.
Consciousness scienceResearch methodsCausality and interventionPhenomenology, process and enactionHusserl I: intentionality and directed experience
A first-principles account of how act, object, profile, horizon and fulfilment make an experience directed towards something, with an executable intentional-analysis protocol.
Consciousness scienceCausality and interventionIdentity and authorityPhenomenology, process and enactionHusserl II: retention, the living present and protention
Why a melody is heard as one unfolding temporal object, how retention and protention thicken the present, and how to test temporal horizon dependence without confusing function with experience.
Causality and interventionConsciousness scienceResearch methodsPhenomenology, process and enactionHusserl III: phenomenological method without private certainty
A first-principles account of how epoché, systematic variation and intersubjective criticism can make first-person evidence useful without treating introspection as infallible.
Consciousness scienceResearch methodsCausality and interventionPhenomenology, process and enactionHeidegger: tools, breakdown and situated intelligence
Why smooth tool use hides the structure of intelligence, and how breakdown-first testing reveals whether a human or machine can reconstruct a practical world.
Causality and interventionReliability and recoveryEvidence and provenancePhenomenology, process and enactionMerleau-Ponty and the body-subject
Why bodily orientation is a continuously recalibrated field of possible action, and how to test it through an embodiment-substitution experiment.
Causality and interventionConsciousness scienceResearch methodsPhenomenology, process and enactionWilliam James and the stream of consciousness
Why thought looks like a chain of discrete contents, how James's fringe and transitive parts change the causal model, and a transition-sensitive experience-sampling protocol for studying the stream without freezing it.
Consciousness scienceResearch methodsCausality and interventionPhenomenology, process and enactionBergson and Whitehead: duration, events and process
A first-principles test for deciding when a mind must be modelled as a history-constituted process rather than a sequence of snapshots.
Causality and interventionEvidence and provenanceConsciousness sciencePhenomenology, process and enactionVarela: enaction and neurophenomenology
Why first-person descriptions, organismic dynamics and environmental interaction become scientifically useful only when they are designed to revise one another under intervention.
Research methodsConsciousness scienceCausality and interventionIndian systems of mind and awarenessPramāṇa and many ways of knowing
A first-principles account of how perception, inference, testimony, introspection and contemplative inquiry can constrain one another without being collapsed into one hierarchy of evidence.
Comparative traditionsResearch methodsConsciousness scienceIndian systems of mind and awarenessSāṅkhya and cognition without a conscious agent
A first-principles account of how Sāṅkhya separates discrimination, memory, action, ownership and witnessing, with an executable function–witness audit for artificial agents.
Consciousness scienceCausality and interventionAgent architectureIndian systems of mind and awarenessYoga I: vṛttis, knowledge and error
A first-principles account of Pātañjala Yoga’s five vṛttis as a dual-axis audit of cognitive route and practice effect, with an executable claim-and-practice concordance.
Evidence and provenanceCausality and interventionRisk and regulationIndian systems of mind and awarenessYoga II: attention training and the kleśas
A first-principles account of how Patañjali's eight limbs address recurrent attentional capture, with a safe capture-rebound experiment and executable simulation.
Causality and interventionLearning systemsResearch methodsIndian systems of mind and awarenessAdvaita Vedānta and the observer problem
Why recursive self-models can establish access and control without establishing the non-objective presence in which anything appears, and how to audit the difference.
Causality and interventionConsciousness scienceResearch methodsIndian systems of mind and awarenessKashmir Shaivism I: recognition and reflexivity
A first-principles account of prakāśa, vimarśa and recognition, with a causal instrument for testing machine self-access without mistaking self-report for awareness.
Consciousness scienceCausality and interventionResearch methodsIndian systems of mind and awarenessKashmir Shaivism II: contraction and the thirty-six tattvas
A first-principles account of māyā, the five kañcukas and the thirty-six tattvas, with an executable instrument for testing where machine comparisons hold and break.
Causality and interventionProduction systemsTransformer mechanismsIndian systems of mind and awarenessVāk and the depth of speech
A first-principles inquiry into whether parā, paśyantī, madhyamā and vaikharī describe representational stages, ontological manifestation, or both, with an executable translation audit for human and machine comparisons.
Causality and interventionResearch methodsIdentity and authorityBuddhist and East Asian models of mindEarly Buddhism: aggregates, dependent arising and no-self
A first-principles account of how early Buddhism explains continuity, agency and responsibility without positing a permanent owner behind experience.
Causality and interventionConsciousness scienceResearch methodsBuddhist and East Asian models of mindAbhidhamma I: moments and cognitive process
A first-principles account of the Theravāda citta-vīthi as an ordered, gated event grammar, with an executable temporal-process lab and explicit limits on modern comparison.
Causality and interventionConsciousness scienceEvidence and provenanceBuddhist and East Asian models of mindAbhidhamma II: mental factors and co-arising structure
A construct-valid method for using Theravāda Abhidhamma mental factors to generate machine-interpretability hypotheses without projecting Buddhist moral categories onto model features.
Causality and interventionIdentity and authorityResearch methodsBuddhist and East Asian models of mindYogācāra: storehouse, seeds and transformation
A first-principles account of how Yogācāra explains continuity, disposition and learning without a permanent self, with an executable trace-and-disposition model.
Causality and interventionConsciousness scienceEvidence and provenanceBuddhist and East Asian models of mindMadhyamaka and relational semantics
A first-principles account of why Madhyamaka does not reduce reality to relations, and a practical four-dependence audit for testing semantic labels under change.
Risk and regulationRepresentationCausality and interventionBuddhist and East Asian models of mindZen and Mahāmudrā: what resists method
A first-principles investigation of why Zen kōan work and Mahāmudrā direct instruction sometimes proceed by changing how representations are held rather than by adding a final description.
Causality and interventionResearch methodsComparative traditionsBuddhist and East Asian models of mindDzogchen and the limits of architectural comparison
A causal method for comparing AI architectures with the Dzogchen distinction between sems and rigpa without mistaking resemblance for identity.
Causality and interventionComparative traditionsConsciousness scienceBuddhist and East Asian models of mindOne architecture, many maps?
A first-principles method for comparing Indian, Buddhist, Chinese and phenomenological maps of mind without forcing equivalence, with an executable consistency matrix.
Causality and interventionConsciousness scienceResearch methodsScientific theories of consciousnessThe hard problem and the price of every exit
A first-principles method for comparing physicalism, illusionism, biological naturalism, dualism, panpsychism, neutral monism, idealism and process views by the explanatory obligations each position transfers rather than removes.
Consciousness scienceCausality and interventionPhilosophy of mindScientific theories of consciousnessDoes the substrate matter, or the causal organisation?
A first-principles replacement protocol for distinguishing software functionalism, structure-sensitive computation and biological naturalism without mistaking behavioural continuity for phenomenal proof.
Causality and interventionResearch methodsConsciousness scienceScientific theories of consciousnessGlobal workspace and conscious access
A causal test of global workspace theory that separates broad neural availability from the leverage a broadcast has across memory, reasoning, action and report.
Consciousness scienceCausality and interventionResearch methodsScientific theories of consciousnessIntegrated information and the causal structure of artificial systems
Why integrated information theory evaluates a physical cause-effect structure rather than an algorithm, and how to build a bounded causal dossier for artificial systems.
Causality and interventionConsciousness scienceMemory systemsScientific theories of consciousnessHigher-order theories and metarepresentation
A first-principles test of whether higher-order representation contributes causally to conscious awareness, distinct from perception, confidence and report bias.
Causality and interventionResearch methodsRepresentationScientific theories of consciousnessRecurrent processing and the unfolding challenge
A first-principles account of when recurrent processing differs scientifically from a behaviourally matched feed-forward realisation, with an executable intervention-fingerprint lab.
Causality and interventionResearch methodsConsciousness scienceScientific theories of consciousnessPredictive processing, free energy and consciousness
A first-principles account of what predictive processing and variational free energy explain, where consciousness claims require an independent bridge, and how to test that bridge.
Consciousness scienceCausality and interventionResearch methodsScientific theories of consciousnessBuilding an attention schema without mistaking report for experience
A causal test for distinguishing an attention-control model from a fluent awareness-report mimic, and for stating exactly what the result can and cannot establish about artificial experience.
Causality and interventionConsciousness scienceTransformer mechanismsMeasurement and the atlas of possible mindsStates, contents and neural correlates
Why neural correlates of consciousness become interpretable only when global state, specific content and the evidence bridge are experimentally unbundled.
Consciousness scienceCausality and interventionResearch methodsMeasurement and the atlas of possible mindsPerturb and listen: complexity without report
Why perturbational complexity can reveal latent causal capacity in human brains, why the scalar cannot travel alone, and how to audit any proposed transfer to another system.
Causality and interventionResearch methodsConsciousness scienceMeasurement and the atlas of possible mindsThe report problem and no-report paradigms
Why removing overt reports does not automatically isolate consciousness, and how to validate no-report proxies with causal controls, delayed post-cues and rival-generator tests.
Causality and interventionResearch methodsConsciousness scienceMeasurement and the atlas of possible mindsResearch without theory victory
A practical protocol for advancing rival theories through prediction deltas, versioned commitment packets and revision receipts rather than winner-takes-all verdicts.
Evidence and provenanceResearch methodsCausality and interventionMeasurement and the atlas of possible mindsAnimal minds through the umwelt
A practical method for translating cognitive tests across species while preserving their causal demand and measuring the interface costs that distort results.
Causality and interventionResearch methodsReliability and recoveryMeasurement and the atlas of possible mindsInfants, organoids and split brains
A three-bridge method for studying consciousness when report, embodiment or subject unity can no longer be assumed.
Causality and interventionConsciousness scienceResearch methodsMeasurement and the atlas of possible mindsA comparative atlas of possible minds
A first-principles framework for comparing biological, artificial and collective candidates without turning missing evidence or interface burdens into false rankings.
Causality and interventionConsciousness scienceResearch methodsMeasurement and the atlas of possible mindsBasal cognition and the boundary of intelligence
A first-principles investigation of how bioelectric state, tissue-level memory and counterfactual control can reveal minimal intelligence without neurons.
Causality and interventionResearch methodsConsciousness scienceHow to study a possible artificial mindBefore experience: the minimal constraints
A practical method for defining the configured causal identity of an artificial system before testing any claim about consciousness, subjectivity or experience.
Research methodsCausality and interventionConsciousness scienceHow to study a possible artificial mindBinding and the candidate-subject boundary
A first-principles method for locating candidate subject boundaries in distributed artificial systems through binding-specific causal interventions.
Causality and interventionConsciousness scienceRepresentationHow to study a possible artificial mindIndicators, not verdicts
A theory-relative, dependency-aware method for assessing indicators of possible artificial consciousness without manufacturing a consciousness score.
Causality and interventionResearch methodsConsciousness scienceHow to study a possible artificial mindAn evidence grammar for artificial minds
A first-principles method for keeping behavioural, report, mechanistic, provenance and replication evidence distinct while composing them into defensible claims about possible artificial minds.
Evidence and provenanceCausality and interventionConsciousness scienceHow to study a possible artificial mindRival generators and false friends
A first-principles method for testing claims about artificial minds by building matched rival generators and separating them with causal interventions.
Causality and interventionResearch methodsEvidence and provenanceHow to study a possible artificial mindThe transfer problem and four homologies
A first-principles method for deciding when a biological measure can support claims about an artificial system, using separate tests for quantity, mechanism, phenomenon and inference.
Causality and interventionResearch methodsConsciousness scienceHow to study a possible artificial mindMachine introspection: evidence and contamination
A first-principles test for whether a language model's self-report comes from privileged access to its hidden state or from prompt, output, semantic and measurement shortcuts.
Causality and interventionResearch methodsReliability and recoveryHow to study a possible artificial mindInterventions on artificial minds
A first-principles guide to when ablations, patches, state swaps and counterfactuals support a mechanistic explanation of an artificial system, with an executable causal signature assay.
Causality and interventionReliability and recoveryResearch methodsArtificial subjects, identity, valence and welfareSynthetic phenomenology as a research programme
A practical framework for modelling experiential structure with artificial systems while keeping structural evidence separate from claims of machine experience.
Causality and interventionConsciousness scienceResearch methodsArtificial subjects, identity, valence and welfareValence, synthetic interoception and functional suffering
A first-principles account of why negative reward, homeostatic error, aversive control and felt suffering must be separated, with an executable intervention assay for synthetic interoception.
Causality and interventionConsciousness scienceLearning systemsArtificial subjects, identity, valence and welfareCopies, forks, resets and software identity
A first-principles account of software identity under duplication, rollback and merge, with a configured-lineage manifest and an executable decision tool.
Evidence and provenanceIdentity and authorityProduction systemsArtificial subjects, identity, valence and welfareRights, forks and interrupted minds
A first-principles governance framework for pause, deletion, copying, rollback and conflicting branch preferences in candidate digital minds.
Evidence and provenanceIdentity and authorityConsciousness scienceArtificial subjects, identity, valence and welfareMoral patients, moral agents and the basis of status
Why the capacity to be harmed and the capacity to be held responsible require separate evidence, thresholds and governance decisions.
Causality and interventionConsciousness scienceAgent architectureArtificial subjects, identity, valence and welfarePrecaution without proof
A first-principles framework for deciding when low-credence AI welfare risks justify proportionate, reversible safeguards under scale and uncertainty.
Evidence and provenanceCausality and interventionReliability and recoveryArtificial subjects, identity, valence and welfareResponsible research and communication ethics
A first-principles account of why elicitation, framing and publication are causal interventions in research on possible artificial minds, with an executable TRACE release protocol.
Evidence and provenanceResearch methodsCausality and interventionArtificial subjects, identity, valence and welfareA consciousness assurance case
A practical framework for governing AI systems when consciousness remains unresolved, using versioned claims, typed evidence, defeaters and conditional welfare obligations.
Evidence and provenanceEvaluation and assuranceConsciousness scienceAttention, contemplative science and altered statesThe plain phenomenology of attention
A first-principles guide to distinguishing consciousness, attention, intention, concentration and awareness through controlled observation of lived episodes.
Consciousness scienceCausality and interventionTransformer mechanismsAttention, contemplative science and altered statesTraining attention across Yoga and Buddhist practice
A first-principles account of why withdrawal, concentration, mindfulness, absorption and open monitoring belong in a practice state-space rather than on one ladder, with an executable trajectory and safety workbench.
Causality and interventionResearch methodsTransformer mechanismsAttention, contemplative science and altered statesAttention turning upon itself
A first-principles comparison of witness practice, self-inquiry, apophatic method, Mahāmudrā and Dzogchen, with an executable claim-concordance instrument.
Comparative traditionsResearch methodsCausality and interventionAttention, contemplative science and altered statesNeurophenomenology and disciplined first-person data
How disciplined first-person reports can become constraint-bearing experimental evidence without being treated as anecdotes, transparent truth or metaphysical proof.
Reliability and recoveryConsciousness scienceCausality and interventionAttention, contemplative science and altered statesComputational neurophenomenology
A first-principles account of when a generative model genuinely bridges experiential transitions, neural or physiological dynamics, and behaviour under intervention.
Causality and interventionResearch methodsConsciousness scienceAttention, contemplative science and altered statesContemplative neuroscience, expertise and safety
Meditation is a configured exposure, not a single neural ingredient. A first-principles guide to expertise, expectancy, adverse effects and safer contemplative research.
Causality and interventionResearch methodsComparative traditionsAttention, contemplative science and altered statesAltered states as precision experiments
A causal method for comparing psychedelics, anaesthesia, sleep, hypnosis and absorption through precision, self-model, connectedness and report fingerprints.
Research methodsCausality and interventionConsciousness scienceAttention, contemplative science and altered statesBelief engineering without epistemic collapse
A first-principles method for using expectation, ritual and belief as causal inputs without allowing them to corrupt the evidence used to decide what is true.
Causality and interventionResearch methodsRisk and regulationFormal bridges and comparison disciplineThe analogy audit: from resemblance to preserved structure
A first-principles method for separating metaphor, correspondence, structural homology, formal mapping and proven equivalence, with an executable mapping-status dossier.
Causality and interventionRisk and regulationResearch methodsFormal bridges and comparison disciplineCategories, morphisms and composition
A first-principles guide to categories, morphisms and composition, with worked examples, failure cases and an executable finite-category validator.
Evidence and provenanceIdentity and authorityCausality and interventionFormal bridges and comparison disciplineFunctors and natural transformations
A first-principles guide to structural translation: why functors must preserve identity and composition, why natural transformations demand path-independent change, and how to audit a mapping in code.
Production systemsRepresentationRisk and regulationFormal bridges and comparison disciplineYoneda and knowledge through relations
A first-principles account of the Yoneda lemma, why composition-preserving probes can characterise an object, and how restricted probes create false equivalence in real systems.
RepresentationEvidence and provenanceRisk and regulationFormal bridges and comparison disciplineString diagrams and process theories
A first-principles guide to when boxes, wires, deformations and rewrite rules constitute a proof of process equality rather than a suggestive illustration.
Identity and authorityEvidence and provenanceProduction systemsFormal bridges and comparison disciplineSheaves: local truth and global consistency
A first-principles guide to sheaf gluing, cycle obstructions, ambiguity and stability, with an executable consistency lab for local-to-global systems.
Research methodsReliability and recoveryEvidence and provenanceFormal bridges and comparison disciplineHigher structure and internal logics
A first-principles account of context-indexed truth, higher coherence and a practical scoped logic model for deciding when claims may travel across levels.
Evidence and provenanceReliability and recoveryProduction systemsFormal bridges and comparison disciplineWhat a structural equivalence proof requires
A first-principles guide to proving structural equivalence through typed structures, preservation, reflection, reconstruction, interventions and explicit break conditions, with an executable formal-mapping gate.
Causality and interventionIdentity and authorityReliability and recoveryConvergence hypotheses and executable testsThe bounded convergence thesis
A practical method for comparing structures across machines, brains, lived experience and contemplative systems without smuggling in a shared ontology.
Research methodsCausality and interventionConsciousness scienceConvergence hypotheses and executable testsMachine attention, biological selection and intentionality
A first-principles comparison of transformer attention, biological selection and intentionality, with an executable landmark-and-break protocol for testing what the comparison actually preserves.
Transformer mechanismsCausality and interventionConsciousness scienceConvergence hypotheses and executable testsRetention, recurrence and machine state
A first-principles account of how caches, recurrent states and state-space dynamics preserve the past, with an executable temporal-preservation benchmark.
Causality and interventionTransformer mechanismsMemory systemsConvergence hypotheses and executable testsResidual computation and global workspace
A first-principles test of whether transformer residual computation satisfies competition, capacity, broadcast and broad-use commitments of a global workspace.
Causality and interventionTransformer mechanismsRepresentationConvergence hypotheses and executable testsTemperature, precision and gain
A first-principles test of when softmax temperature, probabilistic precision and neural gain share a mathematical form, and when their causal roles force them apart.
Causality and interventionRisk and regulationRepresentationConvergence hypotheses and executable testsConstraint through the kañcukas
A first-principles investigation of how the five kañcukas can become a disciplined intervention grammar for action, knowledge, preference, temporal horizon and causal order in finite artificial agents.
Causality and interventionPlatform economicsIdentity and authorityConvergence hypotheses and executable testsStorehouse, weights and dispositional memory
A first-principles test of which relations among Yogācāra seeds, traces, maturation and transformation survive comparison with machine training, model weights and online adaptation.
Evidence and provenanceLearning systemsCausality and interventionConvergence hypotheses and executable testsMental factors and sparse features: a construct-valid experiment
A first-principles protocol for using Theravāda Abhidhamma mental-factor taxonomies to generate testable sparse-feature hypotheses without equating philosophical categories with model latents.
Research methodsCausality and interventionRepresentationHuman-machine intelligence and cognitive sovereigntyExtended, distributed and coupled intelligence
A causal test for deciding when a tool, team or AI system becomes part of a cognitive mechanism rather than merely assisting it.
Causality and interventionIdentity and authorityReliability and recoveryHuman-machine intelligence and cognitive sovereigntyHuman-AI joint cognitive systems
A first-principles guide to allocating perception, interpretation, proposal, authority and recovery across human–AI systems, with an executable complementarity simulator.
Identity and authorityHuman oversightReliability and recoveryHuman-machine intelligence and cognitive sovereigntyCollective intelligence across people and agents
Why communication sometimes creates knowledge and sometimes creates herding, with a correlation-aware protocol and an executable truth-seeking experiment.
Agent architectureEvidence and provenanceReliability and recoveryHuman-machine intelligence and cognitive sovereigntyCognitive integrity and retained human intelligence
A practical framework for deciding which human capabilities must remain exercised when AI delegation is faster, based on failure detection, diagnosis and recovery responsibility.
Identity and authorityReliability and recoveryProduction systemsHuman-machine intelligence and cognitive sovereigntyEcological interfaces for human-machine teams
A first-principles account of interfaces that expose decision constraints, uncertainty, provenance and reversible action in human-machine teams.
Evidence and provenanceReliability and recoveryInterface designHuman-machine intelligence and cognitive sovereigntyTrust calibration and decision-sufficient communication
A first-principles account of why calibrated human–AI reliance depends on the smallest case-specific message that supports belief update, challenge and safe continuation or recovery.
Reliability and recoveryEvidence and provenanceCausality and interventionHuman-machine intelligence and cognitive sovereigntyCreativity, analogy and human-AI co-creation
A first-principles account of when AI-assisted recombination becomes useful novelty, why analogy must preserve causal relations, and how to retain human judgement in co-creation.
Causality and interventionEvidence and provenanceReliability and recoveryHuman-machine intelligence and cognitive sovereigntyWisdom under mortality, scarcity and irreversible choice
Why finite, non-repeatable lives turn wisdom into the disciplined management of irreversible option loss, and how human-AI systems should support accountable commitment.
Reasoning and planningReliability and recoveryPlatform economicsAuthority, context and temporal truthThe stochastic proposal and deterministic authority kernel
Why capable AI agents should propose probabilistically while a small, replayable authority kernel grants narrow, expiring permission for every consequential effect.
Identity and authorityCausality and interventionEvidence and provenanceAuthority, context and temporal truthDelegation as a compiler from intent to capability
A first-principles design for compiling human objectives into narrow, expiring, revocable and evidence-bearing capabilities for AI agents.
Identity and authorityReliability and recoverySecurity and containmentAuthority, context and temporal truthThe irreversibility gradient and transaction semantics
A practical framework for selecting retry, rollback, compensation, escrow and human review as machine-initiated actions become harder to undo.
Reliability and recoveryProduction systemsIdentity and authorityAuthority, context and temporal truthTrust, but verify with a different path
Why a second AI reviewer reduces risk only when its evidence and reasoning path fails differently on the cases that matter, with an executable dependence audit.
Security and containmentEvidence and provenanceCausality and interventionAuthority, context and temporal truthContext contracts under regulation
Why relevant data is not yet admissible evidence, and how a decision-specific context contract can bind subject, purpose, time, provenance and authority before an AI system receives it.
Evidence and provenanceRisk and regulationContext engineeringAuthority, context and temporal truthTyped memory and factual authorisation
Why an AI system may retrieve a memory yet still lack authority to use it as fact, and how to implement a typed factual-use gate.
Evidence and provenanceSecurity and containmentIdentity and authorityAuthority, context and temporal truthTemporal entitlement for decisions
Why historical decisions must be replayed against the facts, policies and permissions that were valid, knowable and authorised at the original cutoff.
Temporal systemsEvidence and provenanceCausality and interventionAuthority, context and temporal truthRetrieval as inference, sufficiency and entitlement
Why governed retrieval must optimise an authorised evidence set for decision sufficiency, not merely rank semantically similar passages.
Evidence and provenanceRetrieval systemsProduction systemsReliability, evaluation and recoveryA reliability model for compound AI
A first-principles reliability model for compound AI systems that separates route exposure, shared failure causes and containment, with an executable risk calculator.
Reliability and recoverySecurity and containmentAgent architectureReliability, evaluation and recoveryThe right to abstain with anytime-valid evidence
How a long-horizon AI system can monitor every step, stop early when evidence warrants it, and still control the probability of a false abstention.
Evidence and provenanceReliability and recoveryIdentity and authorityReliability, evaluation and recoveryPath envelopes and long-horizon mission control
Why long-horizon autonomous systems need path envelopes that preserve invariants, mission feasibility and recoverable futures, with an executable trajectory scorer.
Evaluation and assuranceReliability and recoveryProduction systemsReliability, evaluation and recoveryFailure injection as the primary evaluation
Why consequential compound AI systems should be released against typed fault propagation, containment, recovery and proof, with an executable failure injector.
Reliability and recoverySecurity and containmentProduction systemsReliability, evaluation and recoveryReproducibility and counterfactual replay
A first-principles guide to preserving compound AI executions as causal specimens and using controlled counterfactual replay to identify whether outcomes depended on model, context, tools, policy, environment or chance.
Causality and interventionProduction systemsEvidence and provenanceReliability, evaluation and recoveryRecoverable execution
A first-principles method for resuming consequential AI and distributed workflows only when intent, authority, state, external effects and receipts form one causally consistent recovery cut.
Reliability and recoveryEvidence and provenanceIdentity and authorityReliability, evaluation and recoveryControl theory for guardrails
Why runtime AI guardrails need measured state, bounded delay, sufficient authority and a verified fallback, with an executable guardrail stability test bench.
Reliability and recoveryCausality and interventionEvidence and provenanceReliability, evaluation and recoveryNormal accidents and graceful recovery in agentic systems
A first-principles account of why agentic failures cascade, how causal coupling makes some accidents unavoidable, and how to design firebreaks, degraded modes and recoverable execution.
Reliability and recoveryAgent architectureCausality and interventionSecurity, provenance and oversightModel-independent containment and property placement
A first-principles account of which agent safety properties can be enforced outside the model, where to place them, and how to test a typed containment boundary.
Security and containmentIdentity and authorityProduction systemsSecurity, provenance and oversightThe confused deputy and the non-human insider
Why individually permitted reads and actions can compose into an unauthorised information path, and how to stop agentic systems from becoming non-human insiders.
Identity and authoritySecurity and containmentEvidence and provenanceSecurity, provenance and oversightCausal attribution of consequential tool calls
A first-principles method for identifying which instructions, evidence, hidden state and authority conditions caused an AI agent to propose, release and execute a consequential tool call.
Causality and interventionProduction systemsMemory systemsSecurity, provenance and oversightProof of governed execution and the five-chain ledger
A practical first-principles design for proving that a consequential digital action was performed by the right identity, on the right evidence, under valid authority, through the declared execution path, with an independently verified outcome.
Production systemsIdentity and authorityEvidence and provenanceSecurity, provenance and oversightThe end-to-end argument for agent protocols
Why sound agent protocols cannot carry the full responsibility for intent, present authority, real-world effects and recovery, with an executable Protocol Obligation Profile.
Identity and authorityEvidence and provenanceResearch methodsSecurity, provenance and oversightThe Ship of Theseus in production
A first-principles method for deciding when component changes preserve, conditionally preserve, or break the identity of a governed production system.
Production systemsEvidence and provenanceReliability and recoverySecurity, provenance and oversightCausal interpretability as release evidence
A first-principles framework for deciding when causal interpretability findings are strong, stable and scoped enough to affect AI release authority.
Causality and interventionProduction systemsEvaluation and assuranceSecurity, provenance and oversightApproval fatigue and the human factors of oversight
Why universal human approval gates become ceremonial, and how to redesign oversight around scarce discrimination capacity, independent evidence and recoverable control.
Human oversightPlatform economicsReliability and recoveryOriginal architectures and a defensible contributionBuying thought as a portfolio
A decision-theoretic method for allocating inference, tools, independent verification and human judgement by their marginal reduction in decision loss.
Evidence and provenanceApplied AI strategyProduction systemsOriginal architectures and a defensible contributionConstraint through contraction
A first-principles architecture for bounding agentic cognition by coupling restrictions on evidence, action and operational self-knowledge.
Security and containmentEvidence and provenanceIdentity and authorityOriginal architectures and a defensible contributionPrecision-weighted attention as a testable architecture
A first-principles design and intervention protocol for separating semantic relevance from source reliability inside attention.
Causality and interventionTransformer mechanismsEvidence and provenanceOriginal architectures and a defensible contributionHybrid temporal architectures and the thick present
A first-principles architecture and executable benchmark for deciding when attention, recurrence and state-space memory genuinely preserve different temporal scales.
Causality and interventionTransformer mechanismsTemporal systemsOriginal architectures and a defensible contributionReflexive self-models without infinite regress
A first-principles architecture for reflexive control: compress an agent's allocation traces into a bounded causal state, feed that state back into the same controller, and test the loop by intervention rather than self-report.
Reliability and recoveryCausality and interventionEvidence and provenanceOriginal architectures and a defensible contributionThe four gaps as a research programme
A causal research programme that separates embodiment, reflexivity, continuity and non-conceptual-awareness claims into four independently testable interventions.
Causality and interventionResearch methodsEvidence and provenanceOriginal architectures and a defensible contributionBuilding the first research system
A first-principles design for the smallest artificial-mind research system that can distinguish mechanisms, survive replication, expose rival explanations and govern welfare uncertainty.
Causality and interventionResearch methodsConsciousness scienceOriginal architectures and a defensible contributionOpen problems worth a research career
A first-principles method for choosing research problems that separate rival explanations, support a feasible first experiment and change consequential decisions.
Research methodsCausality and interventionReliability and recoveryOriginal architectures and a defensible contributionWriting the first defensible contribution
A practical framework for turning a study into a challengeable research contribution through typed claims, preregistered tests, versioned evidence, executable artefacts, null-result rules and explicit limits.
Causality and interventionEvidence and provenanceResearch methodsOriginal architectures and a defensible contributionThe unified map and the work still ahead
A versioned research map for human and machine intelligence that preserves unknowns, connects claims to interventions and turns disagreement into the next executable study.
Causality and interventionResearch methodsEvidence and provenanceThe research route
Inquiry and epistemic discipline
Sets the claim types, evidence ledgers and stopping rules that keep later comparisons honest.
Mathematical language and representation
Builds the minimum formal vocabulary for structure, change, uncertainty and representation.
Inference, dynamics and causality
Separates correlation from intervention and connects probabilistic inference to changing systems.
Computation, control and emergence
Moves from algorithms to feedback, self-organisation and the limits of computational description.
Learning systems from first principles
Tracks how objectives, data, generalisation and credit assignment shape learned competence.
Transformer anatomy
Opens the model stack from tokens and position to attention, residual streams and in-context adaptation.
Frontier model behaviour and interpretability
Examines scaling, elicitation, uncertainty and the difficult gap between behaviour and mechanism.
Agents, world models and multi-agent systems
Studies planning, memory, tool use and coordination once a model becomes part of an acting system.
Brain, body and biological intelligence
Connects neural computation to embodiment, development, regulation and changing network structure.
Human intelligence and its faculties
Reconstructs perception, memory, language, reasoning and social cognition without treating them as one score.
Western and Islamic foundations of mind
Places contemporary claims beside durable accounts of intellect, soul, perception and internal sense.
Phenomenology, process and enaction
Treats experience as lived time, embodied engagement and process rather than a static inner object.
Indian systems of mind and awareness
Compares layered accounts of cognition, self, attention and awareness across Indian traditions.
Buddhist and East Asian models of mind
Explores dependent arising, no-self, storehouse models and non-dual disciplines of observation.
Scientific theories of consciousness
Compares leading theories by mechanism, prediction, intervention and unresolved discriminations.
Measurement and the atlas of possible minds
Replaces a single intelligence ladder with a multidimensional map of capacities and evidence.
How to study a possible artificial mind
Designs adversarial, theory-declared protocols for claims that behaviour alone cannot settle.
Artificial subjects, identity, valence and welfare
Tests the concepts that would matter if systems developed persistence, interests or morally relevant states.
Attention, contemplative science and altered states
Links trained attention and altered experience to measurement without collapsing first-person and neural evidence.
Formal bridges and comparison discipline
Builds precise translations across vocabularies while preserving differences that should not be erased.
Convergence hypotheses and executable tests
Turns broad synthesis into risky predictions, interventions and programmes that can fail informatively.
Human-machine intelligence and cognitive sovereignty
Asks how joint systems change judgement, dependence, authorship and the right to think independently.
Authority, context and temporal truth
Defines who may act, which context is admissible and how a system knows what is currently true.
Reliability, evaluation and recovery
Moves evaluation from benchmark scores to trajectories, incidents, rollback and operational recovery.
Security, provenance and oversight
Treats evidence, permissions, supply chains and human review as parts of one control surface.
Original architectures and a defensible contribution
Assembles the collection into concrete architectures, comparison tools and a bounded research contribution.