I suspect there are some wins to be had by thinking on how to automate or otherwise delegate well-understood repeated behaviors, while still collecting the proceeds from them.

This is something that could interest a smart and curious person for some time, and actually benefit from the intellect and curiosity expended on it. The person can also reframe the "repetitious monotonous task" as "learning the best way to execute and then automate the task". But learning with any confidence takes a lot sampling for any statistical significance!