Hao Hu
I build things. Mostly in health. Mostly where the infrastructure isn’t there yet.
Seattle Area
I’ve spent sixteen years in data modeling and AI, most of it producing the analysis that other people’s decisions rested on. At the Gates Foundation I was Deputy Director for Data and Insights, and Principal Officer for AI and Health Delivery, building and leading the data, modeling and decision-support function for one of its global divisions, and helping scope several of the Foundation’s big bets on AI in health delivery.
Before that, in the early years of the Institute for Disease Modeling, which joined the Foundation in 2020, I founded and led its epidemiology team, one of several modeling groups there. Doctorate in biophysics and informatics, Indiana University.
I still build. Where a question can be settled faster by a prototype than by a document, I write the prototype, so most of what I do arrives with something working attached.
Now
Independent technical advisory on AI investment in low- and middle-income health systems, through Glassbox. Vendor-neutral by contract.
Writing
- Intelligence Was Rarely the Bottleneck OpenAI resolved ten open problems in mathematics for roughly $2,000 in tokens, then proposed a resolution of Navier-Stokes. Cheap reasoning relaxes one constraint. The question is which one was binding. 13 min
- Fear Is Not a Probability Researchers inside frontier AI labs are frightened, and they may be right to be. But sincerity is not calibration. What I learned modeling epidemics from sparse data is that the number matters far less than the mechanism that generates it. 32 min