Biohacker Feeds a Decade of MRIs to Claude. The Model Plays Doctor. It Is Not One.
Bryan Hayes, 62, of St Paul, Minnesota, spent nearly a decade accumulating blood work, MRI results, and lung scans scattered across separate patient portals. He used Anthropic's Claude to build a personal website themed after The Hitchhiker's Guide to the Galaxy, where he uploads and stores all his results so AI can analyze them collectively. The 2024 data shows 29% of American adults now use AI for similar health-related purposes.
This illustrates the principle of data aggregation amplification: siloed medical records have limited diagnostic value, but unified datasets enable pattern recognition across modalities. The mechanism is straightforward. When an AI model can cross-reference blood markers with imaging results simultaneously, it surfaces correlations no single portal would reveal. The lesson is that data interoperability, not raw data quantity, creates analytical leverage.
Bryan Hayes, a 62-year-old St Paul resident and DIY longevity enthusiast, built his health data platform using Anthropic's Claude. He represents a growing cohort: 29% of American adults now report using AI for health advice.
- Download your last blood panel and any imaging report summaries from your patient portals as PDF files. Most portals have a download or export button. Expected outcome: two or three files on your desktop.
- Upload them to Claude.ai or ChatGPT in a single conversation and ask it to identify any patterns or flag values outside normal ranges. Expected outcome: a summary highlighting which markers are out of range and potential connections between them.
- Ask the AI to generate a list of specific questions to bring to your next doctor appointment based on the combined results. Expected outcome: a printed list of informed questions that make your ten-minute appointment actually productive.