Podcast
Listen to the research. Follow the source.
AI-generated discussions of Michael S. Moniz’s work at S.A.I.L. Explore AI companionship, memory, governance, creative computing, and the conditions of exit. The published episodes are playable below, with direct YouTube links and source notes.
Google NotebookLM produces the discussion voices. These are educational adaptations, not human interviews or verbatim author readings. The written sources and the episode’s corrections remain the authority; a fluent discussion does not turn an open hypothesis into an established result.
Jump to: Memory, control, and exit · The Quarry System · CaveScript32 · Artificial empathy
Published episodes
Harm Without Pain: What Do We Owe an AI?
28:14 · Corrected podcast · Michael S. Moniz / S.A.I.L.
Deletion, revision, and prompt injection change systems. What would establish that a change matters to the system itself? This discussion adapts Harm Without Pain, Concept Note v0.2 (25 September 2026).
Evidence boundary: A conditional, speculative concept note. The proposed behavioral probe is unbuilt and unrun. Neither persistent records nor stable behavior establishes consciousness, moral standing, or harm. Hypothetical examples are not observed results.
Credits and corrections: Underlying work by Michael S. Moniz. Source v0.1 rendered with Claude; v0.2 revised with Codex after external adversarial critique. Original podcast production by Google NotebookLM. This edited edition includes clearly marked clarifications in a distinct synthetic voice. Source license: CC BY-NC-SA 4.0.
Who Controls Your AI Companion’s Memory?
35:26 · Published 18 October 2026 · NotebookLM discussion
What does a companion remember, who controls the history it uses, and what happens when you leave? This edited discussion introduces the proposed Relational Nutrition Label and separates pause, cancel, delete, and export.
Open this episode on YouTube →
Source: The Trinket Soul Framework — Public Second Edition v1.0, Michael S. Moniz, 12 October 2026. DOI.
Reading guide: TSF is a working theory; the label and audit are proposals. Named examples and service comparisons are fictional. Four incorrect passages were replaced with correction cards; retained narration has qualifications beside it. Those cards and the written episode notes are part of the adaptation.
How Systems Turn Partners Into Yield
38:44 · Published 16 October 2026 · NotebookLM discussion
When does collaboration become a system treating its partners only as yield? A discussion of trust, retained endorsements, usefulness, withdrawal, and the boundary between architecture and abuse.
Open this episode on YouTube →
Source: Michael S. Moniz, AX-31A — The Quarry System, Canon Draft v0.8, June 2026; THE AXIS source archive.
Reading guide: Conceptual and architectural draft; evidence open. The proposed role audit and withdrawal diagnostic are not validated scientific or legal tests. Visual clarification panels and the episode description qualify the retained narration.
Turning Cave Art Into Python Code
52:18 · Published 15 October 2026 · NotebookLM discussion
A long-form discussion of CaveScript32, its public reading key, and the creative exchange around its first Python program. Ordered symbol pairs carry modern text and source code.
Open this episode on YouTube →
Source: CaveScript32 — public reading key and first program, Michael S. Moniz; character encoding version 1.0.
Reading guide: A modern, prehistoric-inspired encoding—not ancient decipherment, encryption, or guaranteed error detection. Original audio is retained; read the corrections and context in the YouTube description. Michael’s underlying work, Claude’s first program, and NotebookLM production are credited there.
The Real Cost of Artificial Empathy
about 41 minutes · NotebookLM discussion
An earlier long-form discussion of artificial empathy and AI companionship. Read its episode-specific source and editorial notes alongside the claims; it is listed separately from the Public Second Edition adaptation above.
Open this episode on YouTube →
Source: Source, credits, and editorial notes in the episode description.
Reading guide: AI-generated discussion. The episode is an access route to its underlying work, not independent evidence for its claims.
Prefer a short introduction?
Your AI Companion’s Memory, Control, and Exit is the 4:30 visual companion to the Second Edition podcast. It introduces the same proposed label through a fictional appointment example.
Follow the series
Subscribe to S.A.I.L. — Michael S. Moniz on YouTube for new episodes. This page collects the published discussions; no separate podcast-directory or RSS subscription is being advertised here.
Sources and credits
Underlying work and direction: Michael S. Moniz. Source collaboration is credited in each work; Claude and ChatGPT contributions vary by edition. Podcast production: Google NotebookLM. Editorial review, qualification panels, and media preparation: Codex.
Start with the Public Second Edition, browse the Research and Agent Index, or return to Learn.