The type of grammar that gets corrected is usually not bad to the point of causing misunderstandings. Simple things like "could care less", "would of", etc. Every once in a while there's a mistake that's so bad that it makes things ambiguous, but that's not common. And it's not what these annoying bots are replying to in any case.
> I don't see how this is derailing at all
These bots can devolve into linguistics discussion, etc, that's not related to the linked article or self text at all. And in reddit's case, there are "meta" bots that respond if people respond "bad bot" or "good bot" to a bot's response.
> Grammatical rules don't exist (solely) because someone decided to be stuck up for no reason
Grammatical guides existing does not mean advice is always welcome. I say guides because a lot of supposed rules are prescriptive opinions, and there's no arbitrator that can deem a rule to be "correct". English does not have a language council, unlike French or Korean.
I meant simply that changing the topic is a mild "derailment" so to speak simply by redirecting the attention of the readers away from the topic at hand.
Sure, but this is sort of begging the question. Bad grammar can subtly affect the effectiveness of communication in ways that sometimes aren't even explicitly clear to the participants. Grammatical rules don't exist (solely) because someone decided to be stuck up for no reason.
> Another critique might simply be that while it is certainly useful to understand rules of grammar, it isn't worth derailing a conversation over.
Sure, this roughly describes why I don't correct strangers in practice. But note that the "derailment" is just another side effect of the insecurity I describe, not an inherent quality of correcting grammar: when someone corrects my grammar, spelling, or diction, my reaction is "Thanks! <rest of comment back on topic>". I don't see how this is derailing at all.