Experimentation is valuable for the high VoI, but it seems hard to encourage ‘in general’, because experimenting on anything is painful and difficult, and the more so the more important and valuable it is. So just ‘subsidizing experiments’ would be like ‘subsidizing fixing bugs in source code’.
What would you do if you were a funder who wanted to avoid this? Well, you’d… fund specific experiments you knew were important and of high-value. Which is what the federal government and many other NGOs or philanthropists do.
Experimentation is valuable for the high VoI, but it seems hard to encourage ‘in general’, because experimenting on anything is painful and difficult, and the more so the more important and valuable it is. So just ‘subsidizing experiments’ would be like ‘subsidizing fixing bugs in source code’.
What would you do if you were a funder who wanted to avoid this? Well, you’d… fund specific experiments you knew were important and of high-value. Which is what the federal government and many other NGOs or philanthropists do.