Yana Welinder

AI Summary

→ WHAT IT COVERS Solo founder Yana Welinder built Yanabana, an AI-native fashion brand, using ChatGPT, Codex, and computer use as her entire technical team — replacing engineers, pattern makers, and production researchers — demonstrating how AI enables a single creative founder to execute a full fashion business end-to-end. → KEY INSIGHTS - **Prompt as product spec:** Develop a detailed master prompt that defines every dimension of your creative output — silhouette, fabric behavior, construction details, movement, and sound. This reusable "fashion prompt" functions like a PRD: it produces consistent AI outputs and would give a human collaborator the exact same information needed to execute the work correctly. - **Model selection for creative fidelity:** ChatGPT's ImageGen 2.0 outperforms other image models specifically for design work because it follows source sketches closely rather than defaulting to aesthetically polished but generic outputs. For fashion design, sketch adherence matters more than photorealism — other models produce beautiful images that look like prior runway work, not original designs. - **Computer use unlocks specialized software without learning it:** Codex paired with computer use can operate industry-specific tools like CLO (fashion pattern software) and 3D modeling programs without the founder learning those applications. This approach generates usable output files — CAD files, pattern files — that Codex alone cannot produce through code generation. - **AI as parallel vendor and research engine:** Use deep research mode to identify and rank manufacturing partners, then chain directly into drafting outreach emails and queuing them in email tools like Superhuman for human approval before sending. This compresses vendor sourcing from days of tedious research into a background task completed while doing other work. - **Pattern-making remains the unsolved bottleneck:** Getting AI to generate accurate garment patterns from original designs is still unreliable. Welinder runs human pattern makers and Codex simultaneously, treating it as a race to see which produces the most accurate result fastest — a practical hedge strategy for any workflow where AI output quality remains inconsistent. → NOTABLE MOMENT Welinder described a garment directly inspired by Ruth Asawa's wire sculptures — a dress she conceived without a sketch, purely from repeated exposure to the artist's work in a single day. The piece requires 3D printing and was considered previously unmanufacturable, now executable by one person with no engineering background. 💼 SPONSORS [{"name": "Merge", "url": "https://merge.dev/howiai"}, {"name": "Jira by Atlassian", "url": "https://jira.dev"}] 🏷️ AI-Native Startups, Solo Founder Tools, Generative AI Design, Fashion Technology, Codex Computer Use