Yes, I definitely threw out a very polarized perspective. It's important not to shut down an entire option of churning out something specialized from scratch, because there is a lot of merit to this option.
It is definitely not the best option to write a feature from scratch in every case, and not all libraries are bloated. The decision to write something or use a pre-existing library is more often than not a very difficult, nuanced, and important choice. Pigeonholing into one side like a zealot will in the end be an inefficient use of time. Instead, decisions like these should be taken case by case with some care.
Firstly you're assuming ... snip
Assumptions like these are made just to show your statement "There's really no need to invent the wheel for the umpteenth time over." is wrong. The idea is: if a programmer goes and takes a look at a library's source code, if the assumptions you refer to are indeed true, they can then make an intelligent decision.
As opposed to a dogmatic decision of never having a need to reinvent the wheel.
Trying to do everything by yourself is generally a sign of a inexperienced programmer.
Ironically, trying to use libraries for everything to avoid reinventing any wheels is also a sign of an inexperienced programmer that stopped learning and growing.
At least if a programmer is diligently writing things themselves they are definitely learning and growing.