Comment by waterhouse

I guess, in theory, this can eventually be countered by people using LLM browser integrations to tell them whether comments are worth reading (and maybe to summarize long comments). Is anyone currently working on that? It might be interesting to see.