After a diagnosis, use LLMs as agents to dissect pathology, treatments and trials
louisvarge · x · 2026-10-03
Responding to someone who received bad medical news, X user louisvarge argues for going full agentic with AI: hammering a frontier model with prompts about pathology, limits of standard treatment, variability between doctors, underdeveloped drugs, clinical trials and cross-literature comparisons, endlessly asking "but why can't I just do X".
His core claim: medical practice is highly inefficient, so a patient using LLMs aggressively could plausibly improve on what a doctor expects.
More from AGI Musings
- Solo Devs With Agent Armies May Write All the Code, but Products Still Need Teams — bendee983 · 2026-10-03
- Will AI research access be locked behind special permissions? Open-source would fill the void — TinfoilTricorn · 2026-10-03
- e/acc case: cheap energy and affordable healthcare via AI, bans won't get us there — mark_k · 2026-10-03
- When LLMs can predict what you'll say before you say it: prediction, persuasion and lost autonomy — Namnagort · 2026-10-03
- MIT co-authored book charts eight years of tech worker organizing and backlash — nordicinst · 2026-10-03
- Ben Horowitz: Biggest personal agent win so far is canceling a NYT subscription — Yuchenj_UW · 2026-10-03