The AI Debate Is Not About Art, It's About Money (sublimeinternet.substack.com)

Jasmine Sun left a product role at Substack to write full-time in part because AI made the subject urgent and interesting. In a wide-ranging conversation she explains how she actually embraces LLMs as tools—using them for background research (summarizing technical papers she can’t parse), “last‑mile” revision (rephrasing, word choice), and occasional low‑stakes drafting—while resisting using them for core idea generation or signature prose. Practically, she trains projects in Claude and ChatGPT on ~15 examples of her work, feeds bullet outlines, and gets outputs that pass plagiarism detectors and can serve as first drafts. She also calls out the hypocrisy of many creatives who secretly use AI while publicly denouncing it and argues for transparency about tool use. For the AI/ML community this conversation crystallizes two big themes: economics and human-in-the-loop workflows. Sun echoes data (e.g., 79% of AI adopters hiring more; founders growing more optimistic) that adoption is financially driven, not just ideological, meaning commercial incentives will accelerate tool integration into creative industries. Technically, her experience highlights strengths (summarization, style-conditioned text, editing prompts) and current limitations (idea generation, canned prose), the arms race with detectors, and the evolving notion of authorship—AI both benchmarks and teaches humans, pushing creative practice and labor models rather than simply replacing them.