Turning Data Into Generative Chaos
summary
A prototype combines Azure Cosmos DB, OpenAI LLMs (GPT‑4 via Semantic Kernel), React + p5.js rendering, and a simple API to turn short prompts into parametric Taurus node visuals. The pipeline generates embeddings, uses vector search (top‑5, similarity 0.75) and runs two LLM prompts to produce a 1–3 sentence art description and a strict Taurus node JSON schema (P/Q prime‑number rules). The app saves configs and vectors, renders varied, unpredictable shapes reliably, with occasional duplicate or browser‑heavy failures.
AI-written summary · about
chapters (7)
- 0:00Generative art journey
- 4:38Manual to algorithmic art
- 7:54Parametric vs generative art
- 10:22Project concept and input
- 13:33Technical implementation
- 19:11Live demo tool
- 24:54Q&A
related talks
44mInnovate with Cosmos DB, Semantic Kernel & AI
15mHarnessing the Power of Generative AI for Intelligent Applications
01h 08mIn Prompts We Trust
28mCreating generative art with Javascript
12mBuilding Generative AI Apps with Azure Cosmos DB
01h 00mUnleashGPT: Mastering the Implementation of Large Language Models
43mRoleplAI
56mPicasso's terminal; data science and AI in the visual arts