This is a bit of a shorter post, but it's just some observations on how the internet has felt lonelier and darker - with a drop of optimism around what comes after this current phase.
As social platforms continue to be muddied with AI-generated content and hate-bait engagement, we’re in for an interesting shift in how we interact online. I believe we’ll all find solace on more concentrated platforms, with a renaissance of internet forums and more sophisticated tooling for vetted group chats.
But for now, we have one last curtain call for large social media platforms.
It seems that The Dead Internet Theory is becoming increasingly correct in predicting the end of mass-scale human interaction on the internet – but really, it’s just the end of the writerly purpose of large social networks. The core idea behind the theory is that the internet has become dominated by bots and that there are few organic interactions between real people. It supposedly originated in the late 2010s but was amplified to more mainstream audiences after Kaitlyn Tiffany from The Atlantic covered it.
Although her article recognizes the difficulties in applying the theory to the fullest extent, every year has brought us a bit closer to bots winning. All content is increasingly becoming AI-generated and posted to serve some end, such as selling a product or generating hate with manipulated engagement. Very rarely do you find conversations between actual humans online, and when it’s "important", it’s two humans in an arena of bots.
The theory’s timing was quite impressive, as it seems the internet was on a collision course towards this end. Changes in OpenAI and social media platform incentive have truly accelerated it. Websites such as X now fully incentivize any form of engagement, leading toward countless Tweets and endless amounts of replies that aren’t even related but serve to generate views and harvest likes. Facebook has become a cesspool of AI-generated content geared towards baby boomers, and we even now have AI influencers taking the place of actual people.
Given the sophistication of AI at the current time, replies on social platforms could just as easily be generated to sound as if you’re engaging with a real person. But people are slowly figuring out that some of the most incendiary responses are just being generated for the exact purpose of engagement:
We’ve arrived here by incentivizing the wrong kinds of behavior across social platforms, favoring raw engagement rather than meaningful engagement. Likes, comments, and views increasingly became measures of “successful” content, so naturally, controversial posts became kindling for such a fire. There was a misguided attempt to ‘return’ incentives to creators, but it ultimately led to bots continuing to drive political discourse and views to generate returns.
So, while a few of us are still trying to communicate online for reasons other than generating engagement, we’re all likely to retreat elsewhere for meaningful discussions.
The question is, who's left, where are they currently, and where are they going?
Besides being buried in tweet replies, the real people you want to speak to are in your group chats, emails, one-on-one text conversations, and niche internet forums. People who wish to have meaningful experiences on the internet and participate in a writerly web will slowly entrench themselves deeper in these mediums. Large social media platforms will widely become grounds for spectators of brands, news, and sports, rather than generating any meaningful discussion.
More human interactions require assurances about who you're talking to. I originally described the concept here, but essentially, levels of assurance represent the trust cost based on the stakes of what you’re doing—in this case, the belief that you’re talking to a human being on the internet. There are two ways we typically have high assurance about who we’re interacting with on the internet: meeting them in person or using peers to validate that someone is who they claim to be.
There are very few venues in which we can have guarantees about talking to authentic people and having truly meaningful interactions on the internet, and that’s why we’re likely to retreat to curated group chats where members are selectively invited. The market opportunity here involves building around curated group chats and what you can do with them. Whether that's deeper forms of engagement within the context of a conversation (threading discussions, interactive activities within a chat, etc.), or just different ways of imagining group chats as something beyond just quick replies.
Internet forums function similarly as well, encouraging long-form, hypermedia-rich discussion, but run into the aforementioned issues unless they're gated. Although limited incentives exist for someone to post on a forum other than genuine interest in a subject, the cost of attacking a forum is quite low. It would be cheap to eclipse someone entirely in a forum where they think they’re speaking to real people to sell them something or try to influence their thinking. Again, in these scenarios where it's meaningful interactions, we'll likely have to curate groups and have some high assurances around group membership.
All of this is to say that we're in for an extreme shift in traditional social media towards a more readerly experience with limited human interaction. I'm not saying that this is novel in any way - people have been having deeper discussions and sharing ideas in these kinds of concentrated groups for years - but I think it'll accelerate a renaissance of features around those mediums.
It's fine to accept the fact that we're likely going to be using existing social media platforms to continue to scroll endlessly, but bots will likely drive human interactions and groundbreaking new ideas to smaller groups instead of 'the public square.' Eventually, these 'public square' platforms will be around to release information to the world, but further discussion and the advancement of ideas will go back to group chats and forums.
The internet is dying, but it'll bring us closer together.
