ChatGPT 3D Chess Modeling Tested: First Attempt Flawed, Fixable with Prompts
goodside · x · 2026-08-03
The author shared details of their experiment where ChatGPT modeled a chess set in Blender. The initial turn-one output had flaws, such as poorly modeled knights and a levitating captured queen. However, after several rounds of prompting to fix small details, the AI successfully generated the refined model. A share link with all prompts was included.
Related event: ChatGPT Autonomously Installs Blender to Recreate Classic Chess Match(2 posts)→
More from coding & agent
- Open-Sourcing Comp AI: A Sales Agent That Rebuilds CRM with an Evidence Ledger — aigclink · 2026-08-03
- ChatGPT Autonomously Installs Blender and Renders 4K Menger Sponge — goodside · 2026-08-03
- Implementing Finger Frame Video Effect Using Gemini — sophiamyang · 2026-08-03
- oh-my-mermaid: Auto-Generate Codebase Architecture Diagrams with Claude Code — tom_doerr · 2026-08-03
- Agents Learn to Draw Basic Shapes Using X, Y, and Radial Cues — johnowhitaker · 2026-08-03
- NeurIPS 2026 Workshop Call for Papers: LLM Social Reasoning & Simulation — timrudner · 2026-08-03