I have a concern.
This kind of identification of outbreaks and shortages is needed, but this is *ML*- they’re describing ML, and this doesn’t give me great confidence in their training data or input data.... 
It’s going to take a level of cooperation on good data we haven’t seen outside the scientific community yet during this outbreak, and some really solid algorithms.
I apologize if I’m being unclear - the article hand waves some massive computing challenges with a lot of potential failure points with buzzwords. It’s very important that we build this capacity and it works accurately and securely without bias.
