"Natural Interfaces" and "Natural Language Programming"
summary
Claim: democratize programming by letting plain language drive behavior rather than teaching syntax. Shows a working prototype called procession that generates p5.js sketches via the OpenAI API, wires in p5.js + MediaPipe for webcam hand‑tracking and timers, reads serial/RFID devices, and triggers sounds. Surprising result: models reliably generate runnable p5.js code for student projects, but generation is brittle—errors occur and boilerplate globals and careful prompt framing are required to make demos repeatable.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 3:13AI design patterns
- 7:31Live demo
- 13:17p5.js with AI models
- 22:28System architecture
- 31:34Managing prompts
- 39:25Key takeaways
related talks
48mDialogues on Natural Code
35mProgramming What Cannot Be Programmed: Aesthetics and Narrative
12mNavigating the World of AI Development Without Going Insane
24mNatural Language Coding Revolutionising Programming with GenAI
56mBeyond Chat: Bringing Models to the Canvas
15mNatural Language-Guided Programming
34mProgramming Languages for AI Programing Agents
20mCode Mode: Let the Code do the Talking