If one is going to use own implementations, one has to extensively test them in order to assure reliable/hardened code. Even more if one publicly promotes dropping battle-proven libraries. But throwing the highly unstable, hence dangerously unreliable, implementation of shuffle at the audience might rather spread doubts whether the author is that knowledgable. The same applies for throttle and debounce, even worse is memoize and isEqual. Almost none of the proposed alternatives can be safely used.

And because of that the author unfortunately fails proving his claim which I otherwise would almost entirely agree with.