Eternal September describes how in 1993 when Usenet forums became added to AOL there was a perceived sudden dumbing-down or reduction of quality of discourse due to the floodgates being opened to a larger, less internet-savvy audience. I think the same is seen regarding so-called ‘dead internet theory’–it’s not that AI or bots have replaced human users, although there is a lot artificial engagement–but rather it’s more like another Eternal September, but on a much grander scale.
I put the year of Eternal September 2.0 at around 2022. It’s especially prevalent in topics pertaining to Math or AI, and less so the humanities. So you either have two extremes: Insightful content, which is becoming increasingly scarce, and on the other extreme, content that hardly passes as fully cognizant. It’s not dead or AI-generated, but more like conversing with someone who is obtuse or not ‘getting the point’. Is it a coincidence the ‘midwit meme’ also became popular around 2020-2022 that describes someone who although fully human, is dim in spite of credentials or an interest in science.
But unlike the original Eternal September, there does not seem to be any obvious catalyst for its return. Maybe it has to do with the increasing polarization of society or the politicization of everything, made worse by Elon’s takeover of Twitter and a highly contentious election cycle (but what election cycle isn’t contentious?). Or the AI boom attracting people who are under the illusion of being smart but aren’t.
Here’s an example, from a related post The Interlocutor’s Advantage:
Eliezer Yudkowsky’s niche is on AI. Presumably you would have to be somewhat smart or at least literate to follow this topic, yet it attracts a certain type of person that appears to be neither. But I have never seen this type of person in those ‘soft and feely’ humanities subjects or topics. This ‘talking past each other’ or ‘talking at the wall’ phenomenon does not occur when it comes to discussing books or philosophies. Debate and disagreement directly addresses the relevant points. The correct context is inferred instead of uncharitable misconstructions.
Worse, if you clap back or block the person, you’re accused of censorship or incivility for overestimating the literacy of the audience, if by ‘overestimate’ apparently anything above a 6th grade level. This is would be fine if conversing with actual elementary school students, but your bio says you work at a tech company or have a GitHub profile, so why is understanding a syllogistic argument suddenly too hard?
Proponents of dead internet theory also overlook there was plenty of spam or artificial behavior in the early 2000s to 2010s–spam comments on YouTube for free iPhones, spam comments on blogs, or camgirl profiles on social media. Yet the internet was not considered dead either. The spam was easily ignored or filtered out. There was still high quality discourse. But fast-forward to today, when encountering such poor quality of discourse, it can certainly feel like the internet is dead when there isn’t much quality discourse to be had at all to counterbalance it. But this has to do with people being lazy or a general dumbing-down effect, not AI.
At least the iPhone spammer or camgirl is an obvious nuisance. Nowadays, however, this epidemic of lazy or bad-faith posting is undeservedly given the benefit of the doubt as if its misconstructions arise from a misunderstanding or ‘just asking questions’ rather than laziness or passive aggressiveness.
In the ‘old’ days of the forum-based communities, such individuals were ridiculed or hazed, and if persistent, banned. But with the decline of forums and the rise of Reddit and other social media, there is no hierarchy or standard in which users are expected to follow. The absence of any sort of user seniority system, as commonly seen on traditional forums, means that the low-effort poster who registered yesterday is on the same level as established members. Social networks use tone policing, community moderation, and sentiment analysis to rank comments and for shadow-banning, so this creates no effective avenue or recourse for attempting to correct this behavior except for blocking the person.
Rightfully calling out the person for being a moron opens the risk of being banned from the community or algorithmically shadow-banned by tripping-up the sentiment filter. The result is a proliferation of discourse that exists in a sort of purgatory of not being blatantly spam or otherwise rule-breaking so as to avoid tripping the sentiment filter, but is still crap, in which there is nothing that can be done about it.