The live-coding community values human struggle and failure over AI-generated output

A recurring position in live-coding culture holds that the point of live coding is the human process — learning through failure, improvisation, and incremental discovery — not the quality of the final sonic output. This stance means that using an LLM to generate Strudel patterns short-circuits the very experience the practice is designed to produce. The community consistently answers newcomers asking ‘how can I use ChatGPT to make it easier?’ with ‘why would you want to?’ The argument is not that AI tools are useless generally, but that they remove the productive friction that creates skill and identity in this specific context. This positions live coding as an authenticity practice: the performer’s visible, fallible process is part of what makes the output meaningful to audiences and players alike.

Examples

A new Strudel user asks on Discord how to use ChatGPT to generate patterns. Experienced community members redirect: the value is in figuring out the commands yourself, making sounds that don’t work, and discovering what does.

Assessment

Explain in your own words why a live-coding practitioner might reject AI code generation even if it would produce better-sounding output. Then identify one domain where this argument would not hold and explain the difference.