Tag
The author explores whether language models can create art through an iterative painting process rather than one-shot generation, building an app that uses a vision-language model to apply strokes one at a time. The experiment highlights the fragility of LLM-generated artefacts and reflects on artistic sincerity.