I am: Untitled

I am a visual artist, but my practice also contains moving image, research, professional systems work, archive building, and Glyphica: an evolving inquiry into question first architecture. As an independent creator, a digital home page for my work always fell flat on expectations of delivery and individual expression.

ChatGPT, Codex, GPT-5.6, and Sites changed that. They allowed me to begin with conversation, artistic judgment, and systems thinking, then move those decisions into a working public interface. I am: Telly is a use case for what becomes possible when the barrier between having a complex practice and being able to represent it starts to fall.

The result is a living, multidisciplinary self-portrait. It does not introduce me through a single biography page. It offers three connected frames:

• I am: Untitled begins with the artwork: 28 source-linked digital reinterpretations derived from physical sketchbook works.
• I am: Telly opens through moving image, then presents a concise visual profile across art, archive, professional continuity, public research, Glyphica, and a direct invitation to begin a conversation.
• Glyphica is a working example of the deeper systems practice the site is capable of carrying. Visitors can compare a direct GPT-5.6 answer with one shaped to keep its audience, sources, boundaries, unresolved questions, and update path visible.

The site is art, research, portfolio, and working demonstration without forcing or limiting my creativity of expression.

How I used ChatGPT, Codex, GPT-5.6, and Sites

ChatGPT became the thinking space. Across a long-running conversation, I could work through artistic direction, public language, research, system design, personal representation, and correction. That continuity mattered. I did not need to arrive with a finished technical specification before I could begin making my digital representation.

Codex became the construction space. It translated our conversations and my visual direction into the repository: responsive components, navigation, moving-image presentation, profile interactions, the Glyphica application, the server-only API boundary, tests, evidence, and deployment packaging. We revised the experience in the browser together—removing text, abandoning box-heavy layouts, improving phone composition, clarifying the language, and making the site feel authored rather than generated.

GPT-5.6 supplied the deep-work layer. It supported the extended reasoning and implementation needed to hold design, code, provenance, privacy, public language, and testing in one process. It also powers both branches of the live Glyphica comparison. The model and provider controls stay equal between those branches so the interface can show what changes when Glyphica adds an explicit operating layer.

Sites became the public interface. It turned the working repository into a place another person can actually encounter. For me, this is the final connection: not AI as a private answer machine, but AI helping someone create a public space for human recognition and contact.

What was built during Build Week

The artwork, personal history, research history, archive, and Glyphica practice predate the event. Build Week did not create those materials; it made them serviceable together.

During the event, we assembled the current art-first site relationship; built the responsive I am: Untitled continuum; produced the desktop and phone moving-image compositions for I am: Telly; created the visual profile and contact surface; implemented the public Glyphica comparison; added its consent-gated OpenAI Responses API route, deterministic replay, strict schemas, shared-control checks, privacy, rate, cost, retry, and failure boundaries; and tested the result across rendered routes, live browser interaction, desktop, and phone.

This inherited/new distinction is part of the project's meaning. The OpenAI tools did not invent my practice. They gave me enough technical and expressive leverage to make its relationships publicly encounterable.

Challenges

The central challenge was representation. How could I take my artistic style, my desires and express it through ChatGPT to create something not just artistic but also technically navigable.

The design had to make complexity navigable without becoming an information dump. We repeatedly removed copy, replaced explanation with visual movement, and composed the experience differently for desktop and phone. The site needed enough structure to orient a stranger while still feeling like the work of an artist.

The Glyphica case study added a second challenge: technical honesty. Its paired comparison required identical source and request bytes, equal model controls, server-only secrets, strict validation, explicit consent, bounded spending, and a replay path that never pretends to be live model evidence.

What I learned

AI access can be authorship access. The important outcome is not that GPT-5.6 can generate a website. It is that I can stay at the level of meaning, relationship, visual judgment, and correction while Codex helps carry those decisions into software.

I also learned that a personal site can be more than a container for finished claims. It can be a living interface between a person and the public: one that shows work, preserves difference, demonstrates active systems, and makes room for conversation without pretending it has finished defining the person.

Built With

• api
• chatgpt
• codex
• css
• gpt-5.6
• next.js
• node.js
• openai
• react
• typescript
• vinext