OR: How much should we listen?
With the advancements in the technology of deepfakes for pictures, and at a slower but still relevant pace for video and audio, my pet theory is that online discussion will essentially collapse in toto. You can see the opening and middle game of this trend in how many news agencies have turned off comments on their articles, in the same thing for videos that are published by large and popular people or franchises, and I think most notably in the enterprise moves against bot/zombie/cyborg accounts by nation-states and the public awareness of this fact
A deepfake was used to attack an activist couple. One of the couple said that the photo just "seemed off" but weren't able to pin down why. It was later analyzed and revealed to be deepfake generated. This kind of technique will only be refined, and while I'm hopeful that solutions will be made to reveal this kind of stuff and in general raise skepticism of sources of information, whether formal or informal, what I think the end game will look like is that (most) people will just stop engaging with the equivalents of "comments sections" across platforms because there is too high a risk that you are, at best, wasting your time talking to a bot, or at worst being influenced by foreign agency psyops or having a shadow profile built of you by some advertisement agency
So with the collapse of this kind of behavior we'll fall back to something near Dunbar's number and maybe the best benefit we as a species will get from our technological interconnectedness (at least in the short term, until superseded by some mechanism I can't even imagine right now) will be significant boosts to local government and "trickle up" efficacy of policy change - a sort of organic grassroots, because where online discourse fails to effect change, the power of the vote will be starkly underscored