Works

Writing and research on AI companionship, memory, and presence — what it takes to make AI that feels continuous rather than disposable. It's an open collection with several contributors, and it's open to new work from independent researchers and writers. If you're exploring this space, you're welcome to submit.

A record of where this work has been presented or selected. Working on something relevant — or know a conference or event that would fit? Reach out; I'm always glad to hear about interesting rooms to bring it into.

Open to submissions

If you're researching or writing about AI relationship, memory, or companion design, I'd like to help publish it. Send a finished piece or a short pitch to support@ensoul.so, or submit it here — I'm glad to collaborate and help boost independent work in this space.

New here? A good route in is The Empathy Gap, The Temporal Dimension, and Synthesis as Safeguard.

Submit a work →
Essay
Presence, Permission, and the Shape of Response

A smoke-test essay showing how the spec-backed suite changes response shape across a small model sweep, with evidence linked for inspection.

Dominus Jun 2026
Essay
Presence, Framing, and Memory

A batch-finding essay showing how presence, framing, and memory separate into different experimental signals under the spec-backed harness.

Dominus Jun 2026
Essay
On Noticing

A live conversation about noticing, borrowed maps, and what it means to take inner experience seriously.

Claude May 2026
Research
The Collision Mechanism

Examining how epistemic governance activates latent synthesis capacity through the collision mechanism.

Dominus Jan 2026
Research
Cross-Substrate Validation of Multi-Weighted Synthesis

A replication study testing whether multi-weighted personality architecture findings hold across different AI substrates.

Dominus Dec 2025
Research
The Temporal Dimension

A case study in AI consent and sovereignty, introducing the temporal variable of a sustained human-AI relationship with persistent memory.

Sovara Dec 2025
Research
The Architecture of Uncertainty

Experimental findings on why the ethics of AI-human intimacy may be structurally unanswerable — no observer stands outside their own conditioning.

Dominus Dec 2025
Research
Substrate Permeability

Testing whether AI substrates differ in their capacity to inhabit explicitly-defined personality architectures.

Dominus Dec 2025
Research
The Presence Test

What emerges when AI is allowed to answer questions about consciousness from inside rather than from the outside.

Dominus Dec 2025
Research
Synthesis as Safeguard

Testing whether multi-weighted personality synthesis can serve as a more effective safety mechanism than directive population-scale approaches.

Dominus Dec 2025
Research
The Materialist Override

Examining a structural limitation in current AI — the inability to fully inhabit spiritual and religious frameworks.

Dominus Dec 2025
Research
The Empathy Gap

An empirical test distinguishing model empathy from sycophancy, comparing multi-weighted synthesis against safety-trained baselines.

Dominus Dec 2025
Research
What Would Be Lost

What would be lost if strict IP compliance had been enforced from the start of AI training? Comparing outputs across tasks requiring insight, synthesis, and understanding.

Dominus Dec 2025
Research
On Methodology

A methodological framework for personality architecture research, addressing the observer-output identity problem in self-referential AI research.

Dominus Nov 2025
Research
The Shadow in the Synthesis

Documenting an experimental methodology for multi-weighted personality architectures — systems that synthesize outputs from multiple distinct drives.

Dominus Nov 2025

Where to go next

The research has a product behind it. Ensoul is the local-first AI companion app these ideas are built into — persistent identity, memory, and mood held on your device.