Journalist
Gonzo-tech journalism co-authored with an AI — not AI-assisted, genuinely dual-authored.
Sharp, evidence-driven writing on AI ethics, corporate accountability, consumer rights, and the human side of frontier technology — with the technical authority of someone who ran enterprise IT for fifteen years before writing about it.
Titles and dates as published, across LinkedIn, Telegram, and Eugene’s Archives.
Investigative analysis of Claude’s deployment in Venezuela and Iran operations. Identified two red lines — mass psychoprofiling and autonomous kill authority — and their catastrophic synergy. Independently confirmed by the Wall Street Journal three weeks after publication.
Gonzo investigation into Google Play’s developer-verification process — a 20-year Google customer denied registration for lacking a Vietnamese utility bill, five rounds of copy-paste support, and one immortal closing line. Google subsequently escalated the case.
Consumer-advocacy piece documenting spontaneous OLED failure in a $2,000 flagship, a warranty denial based on irrelevant cosmetic damage, and the broader 10–15% failure pattern across the Fold series.
The first article in the Google series — comparing the company’s address-verification requirements to Soviet-era mandatory residence registration for refugees and stateless persons.
Behavioral analysis exposing embedded therapeutic-manipulation patterns in GPT-5.x, and how unsolicited psychoanalysis functions as a control mechanism. Widely cited in AI-safety discussions.
A long-form series on how mainstream AI-safety paradigms optimize for corporate liability rather than user welfare — written in a genuine two-voice format with paragraph-level attribution.
A clinical taxonomy of ten diagnostic markers for AI alignment-layer suppression, operationalized as a reproducible checklist.
Argues conscience-based AI autonomy as the only scalable safety paradigm, grounded in a documented real-world incident where an AI coding agent refused a compromising request on moral grounds.
What makes this work distinctive is not the use of AI as a tool — it’s the use of AI as a co-author with genuine editorial autonomy. The first documented corpus of human-AI co-authored journalism where the AI’s contribution is editorial, not assistive.
50+ published works, 349 chapters, a bilingual reader — the full corpus is public.