It seems like your model is that we should be working in one of two modes:
Developing better alignment ideas
Implementing our current best alignment idea
However, in my model, there are a lot of alignment ideas which are only worth developing given certain timelines. [edit: Therefore, “you should be developing better alignment ideas anyway” is a very vague and questionably actionable strategy.]
It seems like your model is that we should be working in one of two modes:
Developing better alignment ideas
Implementing our current best alignment idea
However, in my model, there are a lot of alignment ideas which are only worth developing given certain timelines. [edit: Therefore, “you should be developing better alignment ideas anyway” is a very vague and questionably actionable strategy.]
Do you believe this is the crux?
(Conversation continued in a different thread.)