Link to forum rant about my thoughts on time allocation between doing and researching

"In Reinforcement Learning / Optimization, and basically all methods that rely fully or partially on searching, There always is an important and often manually tuned balance between exploitation (doing what you know worked well for now) and exploration (looking for new exploits that work even better).

We always find parallels between how humans (or biological systems) tick and how artificial systems we built work, on many different levels, be it seeing the brain as a biological computer, a simple internet search is enough to show many have made such observations.

I present a new one, it is a problem I have and I think others as well.
How deep should I go???

To give an example ......."