Imho, this comes down to hiring. I basically didn’t hire people unless I was personally doing work that someone could take off my plate. I was working on turbine, I could interview for Linux kernel udp experience and hand the task I was doing to the new hire. Within 90 days it would be obvious if they could do it as well as me or not. That’s all that matters in a startup. It makes no difference what hours of the day or geographical location the person is. If you hire someone and give them a research project, there is no way to quickly evaluate how well they are doing because they problem is unbounded. It’s not even clear the problem is important enough to solve in the first place.