What Separates a Winning Public-Good Project From a Worthy One

HERO optimizes ambulance routes, but rather than just calling a maps API it carries a separate model predicting how much time siren privilege actually saves, and it builds the screen a control room would use. VitaSort reorders an outpatient queue by urgency but deliberately refuses to let a generative model make the call, using the deterministic NEWS2 score so behavior stays predictable. The Dallas waste system goes as far as dynamically rerouting collection based on fill predictions.

The common thread is that each one specifies who looks at which screen, when, and what they decide. Toddle AI offers gait analysis but states plainly that it is a research prototype, not a diagnostic. The OpenMRS consultation stack builds in the assumption that a clinician reviews the output before it is relied on. In domains where lives are at stake, designing so that the AI's output is not blindly trusted is what earns credit.

Takeaway: if you pick a social problem, cut the time you spend describing how serious it is and spend it on operational detail instead. Who uses this? What happens when it is wrong? Where does responsibility stop? A project that answers those three separates itself sharply from others on the same subject. A pitch that ends at 'this problem matters' is the pattern judges see most often.