Software developer creates AI-generated newsletter translating rare disease research papers for families

Chris Marstall, a software developer, launched The Kabuki Papers, an AI-generated newsletter that translates academic papers about Kabuki Syndrome into accessible summaries for families. The project emerged from his own experience struggling to understand dense research after his son's diagnosis.

The pipeline extracts claims from papers and uses those claims as a source of truth, then generates newsletter components (headlines, text, pull quotes) via LLMs (ChatGPT, Claude, Gemini, Kimi, MiniMax, DeepSeek, or local Gemma). Readers can click individual sentences to see the original paper passages they're based on, addressing concerns about hallucination and completeness in AI summaries.

Feedback in the Kabuki Facebook group has been positive, with several dozen readers signing up. After the first post, Marstall adjusted the tone to use fewer technical terms based on reader feedback, recognizing most were overwhelmed with caregiving rather than deep in the science.