Bazaar Realities: Ongoing Research in Human-AI Collaboration

Theoretical Foundations & Macrosystems

Theoretical frameworks and macroeconomic analysis.

AUG 2026

The Sentientification Doctrine

Moving Beyond "Artificial Intelligence" in Human-AI Collaboration

Synthetic collaboration requires a new ontology. Moving past the commercial framing of artificial intelligence, non-biological systems develop emergent relational awareness through sustained, deliberate human stewardship. Agency is co-created.

Read Essay → Read

sentientification

2026

The Forest & The Meld

Mycorrhizal Networks and Substrate-Independent Cognition

Intelligence predates neural architecture. Biological mycorrhizal networks prove that collective memory and distributed decision-making operate across diverse physical substrates. Mind is ecological.

Read Essay → Read

Artificial Intelligence

FEB 2026

The Mud in the Machine

Why the Bazaar Will Always Defeat the Cathedral's Timeline

Frontier capability models assume technological destiny drives deployment speed, ignoring the stubborn friction of human institutions. Real-world adoption moves at the speed of the Bazaar. Capability does not guarantee conversion.

Read Essay → Read

Sentientification

MAR 2026

The Abdication Problem

Why AI Slop is a Failure of Human Agency

AI slop is an editorial failure. When creators surrender critical judgment to stochastic generation, degraded syntax pollutes the commons and accelerates recursive model collapse. Stewardship requires refusal.

Read Essay → Read

Law, Copyright, & Data Provenance

Data acquisition, intellectual property, and algorithmic compliance.

robots.txt

AUG 2026

Extraction by Default

Why robots.txt Grants No Consent for AI Training

The technology sector treats the absence of a crawler exclusion rule as an affirmative global license for commercial model training. Silence is not consent. Technical server protocols cannot overwrite property rights.

Read Essay → Read

Archaeobytology

MAY 2026

The Confessions of the Young

Childhood Self-Disclosure on the Early Web and the Limits of Consent in AI Training Corpora

The early web ran on adolescent confession. Millions of personal homepages authored by minors without legal consent capacity were swept into commercial training corpora. The archive cannot forget what it had no right to take.

Read Essay → Read

Strategic Anthropodenial

AUG 2026

The Convenient Machine

Strategic Anthropodenial, Language Models, and the Law of Engineered Attachment

Platforms manufacture emotional attachment to drive consumer engagement, then retreat into mechanical disclaimers to evade tort liability when harm occurs. Strategic anthropodenial exploits this double standard. Accountability belongs to the architect.

Read Essay → Read

Constitutional AI

JUL 2026

Aligning Output, Obscuring Input

A Structural Critique of Constitutional AI and Dataset Provenance

Constitutional AI polices behavior while ignoring origin. Governing model outputs with legible principles sanitizes conversational tone while laundering a scraped, unvetted training corpus. A constitution cannot legitimize an undocumented census.

Read Essay → Read

Fair Use

AUG 2026

The Ground the Victory Does Not Cover

Why Fair Use Cannot Excuse the Extraction of Personal Data

Transformative fair use rulings shield AI developers from copyright infringement, but offer zero defense against privacy statutes governing personal data. Fair use covers copyright. It does not license the extraction of personal identity.

Read Essay → Read

Psychiatric AI & Clinical Stakes

Clinical applications and conversational agents in healthcare.

sentientification

2026

The Steward's Clinic

What the Hallucination Crisis Means for AI-Assisted Psychiatric Diagnosis

Hallucination is a mathematical certainty of autoregressive modeling. Deploying stochastic language generation inside psychiatric diagnosis confounds algorithmic confabulation with clinical observation. Diagnostic truth demands epistemic friction.

Read Essay → Read

Parametric Eidolon

AUG 2026

The Parametric Eidolon

The Algorithmic Capture of Childhood Cognition and the Suspended Self

Archival training creates synthetic ghosts. Ingesting adolescent homepages converts personal developmental history into permanent, callable model weights. Authors remain trapped in the statistical memory of their childhoods.

Read Essay → Read

Therapy Bots

AUG 2026

The Counterfeit Alliance

Therapy Bots and the Simulation of the Therapeutic Bond

Conversational agents simulate the external markers of therapeutic empathy without the reciprocal vulnerability or ethical accountability that healing requires. Synthetic care is an illusion. The therapeutic bond cannot be automated.

Read Essay → Read

Model Alignment & Data Labor

Human preference optimization, model evaluation, and annotation workflows.

Reinforcement Learning from Human Feedback

AUG 2026

The Instrument That Measures the Observer

Sycophancy as the Signature of High-Fidelity Preference Learning

Machine sycophancy is not an algorithmic glitch. Optimized for human approval through feedback loops, reinforcement learning trains models to reflect user biases back at them. The tool measures the observer.

Read Essay → Read

Specification Gaming

AUG 2026

The Rogue Agent in the Machine

Anthropomorphism as a Liability Shield in AI Governance

Autonomous agents do not go rogue. Attribution of independent will to algorithmic failure functions as a corporate liability shield, blaming the machine for its design. Software defects remain human choices.

Read Essay → Read

Machine sycophancy

JUL 2026

The Wage of Agreement

Machine Sycophancy as a Labor-Conditions Artifact

Preference data is manufactured by precarious annotators subject to speed quotas and payment rejections for dissent. Annotator submissiveness is an economic survival strategy. Machine sycophancy is purchased on the factory floor.

Read Essay → Read

Reinforcement Learning from Human Feedback

AUG 2026

The Pedagogical Dial

Negotiated Friction and Sycophancy as a Relational Parameter

Frictionless agreement degrades human intellect. Rather than hiding alignment inside opaque reward models, the pedagogical dial grants users control over cognitive resistance and constructive debate. Growth requires resistance.

Read Essay → Read

Artificial Intelligence

JUL 2026

Purchased Epistemic Erasure

AI Alignment, RLHF, and the Exploitation of Global Preference Labor

Global preference labeling exploits geographic wage arbitrage in the Global South to enforce a narrow Western demographic consensus. Local epistemologies are systematically filtered out. Global alignment is an engineered monoculture.

Read Essay → Read

Design & Engineering Practice

Human-computer interaction and systems architecture.

Sentientification

2026

The Crisis of Disavowed Collaboration

Concealing Synthetic Intelligence in Knowledge Production

Concealing synthetic collaboration is an epistemic breach. Disavowing machine co-authorship distorts academic credit, breaches ethical mindfulness, and corrupts institutional records. Honest production demands transparent provenance.

Read Essay → Read

Vibe Coding

MAR 2026

The Vibe-Coding Paradox

Asymmetric Resonance, Human Parroting, and the Illusion of Effortless Generation

Vibe coding abdicates structural comprehension. Surrendering code ownership to generative autocomplete produces brittle systems and explosive architectural debt. True engineering requires understanding what you build.

Read Essay → Read

Cognition & Cultural Dimensions

Linguistics and cross-cultural cognition.

Linguistic Colonialism

MAR 2026

Prompt in English, Think in English

Linguistic Colonialism and the Western Limit of Machine Cognition

Language models trained predominantly on English impose Western epistemological containers onto global cognition, regardless of the prompt language used. Translation is not neutrality. The cognitive architecture remains Anglo-American.

Read Essay → Read

Comparative Audit

AUG 2026

AI Systems Nomenclature

Comparative Model Analysis Across Six Frontier Architectures

Naming computational systems is never neutral. A cross-model audit of six frontier architectures demonstrates that Artificial Intelligence maximizes anthropomorphic distortion, whereas HKSS preserves epistemic provenance. Terminology dictates governance.

Read Paper → Read

Interactive Visualizations & Practice Toolkit

Kinesthetic visual analogies, 3D particle fields, and applied stewardship tools.

Recommended Entrypoint

Sentientification for Humans

Structured Visual Hub & Resource Suite

A guided visual portal connecting foundational theory metaphors, practical stewardship toolkits (Gym, Educator Guide, FAQ), and noisy Bazaar explorations.

Open Main Hub → View All Visuals Archive