Stanford Health Care data scientist: technical performance alone doesn't yield clinical value
jonc101x · x · 2026-11-10
Nigam Shah's Stanford colloquium traces an arc in healthcare AI: longitudinal clinical data generating real-world evidence, through predictive and generative AI, to agentic systems capable of acting across electronic health records, payer portals, and other operational infrastructure.
The central finding across every stage: technical performance alone does not produce clinical value. Benefit depends on the workflow, decisions, and interventions surrounding the AI system — on the capacity to act on its output, and on the fairness, safety, reliability, and cost of doing so.
More from Companies & People
- jana_iris joins AI coding startup exe as founding marketer — davidcrawshaw · 2026-11-10
- Nvidia investors publicly defend Jensen Huang despite private disappointment over timing — GaryMarcus · 2026-11-10
- Allen AI opens intern applications with November 15 first-round deadline — allen_ai · 2026-11-10
- Fireworks, top-valued inference provider, hit by hack exposing customer data; clients eye exit — steph_palazzolo · 2026-11-10
- NYT: Commerce Dept Chip-Smuggling Probe Halted After Nvidia Intervention; Huang Under Internal Probe — asusarla · 2026-11-10
- iamtrask: labs deprioritize cost-efficient algorithms due to short moats and cheap cash — iamtrask · 2026-11-10