Feeding personal symptoms/complaints into an LLM (e.g., ChatGPT) to get a probabilistic differential diagnosis and a list of exams that should be run, then using that output to push a provider to actually perform those exams. Framed by the 'Don't Die' team (Bryan Johnson, Dr. Mike Min) as a practical self-advocacy tool for navigating an overburdened, reactive medical system — see Preventable Disease & the Reactive-Medicine Framing.
Method: describe symptoms in detail (a long, thorough intake — e.g., ~100 questions) to an LLM, ask what workup/exams should be happening for that presentation, then advocate for those specific exams with your actual provider.
Real example cited: Bryan Johnson used a ChatGPT intake to correctly flag his mother's eye condition as inflammation/a corneal scratch, later confirmed by an optometrist and resolved with steroid drops.
Caveat: LLMs reportedly outperform physicians on standardized medical tests, but 'life is rarely standard' — this limits how far the self-advocacy approach generalizes beyond textbook-shaped presentations, and it's positioned as a supplement to advocacy with a real provider, not a replacement for one.
None noted against existing material — this is a new concept for this vault.
Из тем: Proactive Health & Self-Advocacy