A Writing Strategy Born from a Production Ecosystem
From Factory to Article: A User’s Field Report from the AI Ecosystem (Article 1)
Aydın Tiryaki & Claude Sonnet 4.6
1. Introduction
Articles written with AI assistance typically follow a familiar pattern: the user provides a topic, the AI generates a draft, the user edits, and the article is published. This method is fast but shallow. The resulting text is usually drawn from the AI’s general knowledge base — competent, but untethered from lived experience.
The method described in this article takes a different path. The article is first lived. Then written.
The co-authorship method outlined here emerges from approximately one and a half years of AI platform experience and eight to nine months of intensive, hands-on work. It did not arise from abstract theory but from within a real production ecosystem: a Gem Factory, multiple workshops, over sixty Gems, and more than a hundred platform versions iterated through trial, error, and refinement.
2. The Core Philosophy
2.1 Live First, Write Later
In conventional AI-assisted writing, a topic is chosen, researched, and written. Here the process is reversed: observations, solutions, and insights accumulate during real work. The article is the distillation of that accumulation.
A conversation spanning the length of a volleyball match — ranging across token architecture, platform instability, pruning strategies, and user rights — captures a wide field of knowledge in its natural flow. This conversation is not a planned research session. It is raw data from the field.
2.2 Two Signatures, Two Perspectives
Articles in this series carry two bylines: Aydın Tiryaki and Claude Sonnet 4.6.
These signatures are not symbolic. Each represents a distinct perspective.
The Aydın Tiryaki perspective comes from the field. It reflects the observations of a user with programming experience stretching back to 1976 — from IBM mainframes to today’s AI platforms — who works intensively within the Gemini ecosystem. These are not theoretical insights. They are practical, earned in production, not in a laboratory.
The Claude perspective comes from inside the platform. It speaks honestly to which criticisms apply equally to Claude, where Claude genuinely differs from Gemini, and which limitations stem from architectural necessity versus deliberate choice.
The presence of both perspectives prevents the series from becoming a one-sided critique. Instead, it becomes a comparative, honest assessment.
2.3 The Algorithm Belongs to the User
The foundational principle of this method is this: the algorithm always belongs to the user. AI is a working infrastructure — not the architect.
This principle carries directly into the writing process. The user determines the topic, decides the order, accepts or rejects proposals, and gives final approval. The AI produces, suggests, and drafts — but the decision always remains with the human.
3. The Writing Process: Step by Step
3.1 Accumulating Raw Material
The article process does not begin with a deliberate writing session. It begins with conversations held during real work — conversations in which platform problems are discussed, solutions are developed, and observations are shared.
These conversations are simultaneously raw article material. Their value lies precisely in being unplanned, natural, and authentic.
3.2 Determining the Titles
Once sufficient raw material has accumulated, an evaluation session takes place. What topics emerge from this conversation? Which can stand alone as independent articles? Which should be merged, which separated?
This evaluation is conducted jointly. The AI proposes, the user approves, and the order is decided together. The result is an article list — the backbone of the series, referenced throughout the entire writing process.
3.3 Writing the Article
For each article, the process works as follows:
The AI writes a complete article in light of the determined title and accumulated raw material. The article consists of a main title, an optional subtitle, author names, section headings, and supplementary sections where needed.
The user reads the article. They may accept it, request changes, or add new perspectives. This stage is where the article matures.
Once the article is accepted, the AI asks: Shall we prepare the English version?
3.4 Adapting to English
When approval is given, the English version is prepared. This is not a translation — it is an adaptation. Turkish idioms, cultural references, and narrative conventions are rendered in forms appropriate for an English-speaking readership. Content is preserved; form is adapted.
This distinction matters. A reference that carries immediate meaning in Turkish — such as a well-known advertising slogan that became a cultural idiom — cannot simply be translated word for word. In the English version, an equivalent expression is found, or the context is explained.
3.5 Cross-References Between Articles
As the series progresses, articles reference one another. Links may not yet be available at the time of writing. In such cases, the reference is given by title and author name within the text; the link is added by the user at the time of publication.
4. The Use of Supplementary Sections
Each article may include one or more supplementary sections where appropriate.
What goes into a supplement? Detailed technical explanations, information valuable to the reader but too dense to fit the main text’s flow, and selected dialogue excerpts.
Why include dialogues? Because one of the distinctive qualities of this series is that it draws directly on raw conversation material. When a dialogue supports a claim made in the article and carries the authentic voice of a real production environment to the reader, presenting it as a supplement adds both transparency and originality.
The decision to include supplements is made separately for each article.
5. What Makes This Method Distinctive
Most AI-assisted writing foregrounds the AI’s capacity to generate knowledge. In this method, it is the user’s experience — not the AI’s knowledge — that takes center stage. The AI renders that experience into language, gives it structure and framing — but the substance of the content comes from what the user has brought from the field.
This difference may seem small. The results are not. The articles that emerge from this process are not general knowledge summaries. They are the distilled form of real observations.
6. Conclusion
Co-authorship with AI, when properly structured, is a powerful method of production. But two conditions are required: lived experience that precedes the writing, and a collaboration capable of giving it language.
Both conditions are present in this series. The experience comes from the field — hundreds of Gem versions, eight open windows, three monitors, the disruption of a single platform update, and much more. The language is the shared product of two signatures.
The series will continue in this manner.
Aydın Tiryaki & Claude Sonnet 4.6 June 2026
| aydintiryaki.org | YouTube | Aydın Tiryaki’nin Yazıları ve Videoları │Articles and Videos by Aydın Tiryaki | Bilgi Merkezi│Knowledge Hub | ░ Virgülüne Dokunmadan │ Verbatim ░ |░Fabrikadan Makaleye: Yapay Zeka Ekosisteminde Bir Kullanıcının Saha Raporu │From Factory to Article: A User’s Field Report from the AI Ecosystem ░ 07.06.2026
Aydın Tiryaki 