Are these people serious? Or is it one if those reddit groups like the one where they pretend birds are all government drones?
Still, I'm with you on this one: there are at least some real people who seriously feel like that (TFA) about LLMs and it's pretty bonkers.
100-ish upvotes is nothing compared to groups with ideas like flat earth or sovcit.
Human connection can be difficult and dangerous, and a lot of people are deeply antisocial or fearful of interaction. Many peoples' primary means of communication was already through the abstraction of the internet before AI came along and a whole generation lost years of potential social development time to lockdowns.
It shouldn't surprise anyone that people connect with AI more deeply than they do to real people. AI is almost perfectly designed to encourage parasocial relationships.
It does however remind me of all the numbers of novice chess players who would rather play against "AI" or chess bots than face up against humans, for reasons of fear or shame in playing against (anonymous) human opponents, and risk looking "stupid" or being silently judged by someone they'll never meet or interact with.