The vibes are off with social media
While it’s hard to put a pin on exactly when things started to shift, it is somewhere between when Meta introduced the metaverse, and when Twitter became X. The last two years have seen a rapid increase in turbulence, with users moving from service to service trying to find a fresh start. Perhaps it is the effort that this involves, or the reflection on what matters in a social network in the first place, but the vibes have shifted with social media as an entire product category. Executives have pivoted from racing to deliver features that users want, to trying to convince a userbase that they need the new features being announced. Services have begun to mimic one another in an effort to be the one app that people can go for everything; it isn’t working, and what was once shiny and new is beginning to tarnish.
The technology behind many social platforms has been solved for a while. It isn’t as hard as it once was to operate an app that shares text, images, and videos at scale [1]. Keeping users engaged and on the platform is a far greater challenge. The 2010 decade saw the Surveillance Capitalism model solidify and gave rise to techno-giants, however companies were always worried about the day that they would run out of new users. For Facebook, that moment seemed like a possibility in February 2022 when it lost daily active users for the first time in the United States, and its share price fell as a result. Meta shares total users within its “Family Daily Active People (DAP)” metric, which is a combination of Facebook, Instagram, Messenger, and WhatsApp usage and makes it hard to gauge how Facebook and Instagram are doing independently (no doubt intentional). That number is still on the rise, crescendoing to 3.29 billion as of the last earnings report, however momentum has slowed. The earnings report at the end of this month will show if that number keeps going up, however improbable that may seem.
That number, however, doesn’t account for the vibes. Hear me out. In conversations that you have with friends, family and colleagues, how many people are happy using those products? More and more people that I talk to use Facebook only to avoid the hassle of working out how to keep Messenger. Instagram is increasingly just for the reels. Many are souring on the idea of supporting the mainstream tech industry in general. Large migrations from one service to another may make headlines, but it is the slow loss of the “it” factor that kills platforms that are hyper-dependent on growth.
Imagine being at a party five years ago, and someone mentions that they have 100,000 followers on Instagram or Twitter. Most people would probably react with a mix of shock and curiosity, and might ask questions about how they did it or if they had any tips to replicate it. Now, imagine the same statement is made about Instagram or X. The reaction may not be hostile, but it might be met with a shrug. Is it cool to be contributing to the bottom line of companies, and in turn individuals, that already have a net worth larger than many countries’ GDP? Is it inspiring to post where users fly by in a second because they are so desensitized to content that scrolling has to be almost constant to make things interesting? Small as it may be, that difference in reaction does have a real impact. One of the intended goals for many while using social media is the validation that comes from followers or likes, and if likes or followers cease to translate as well to that feeling, the product suffers.
While writing this, I stumbled upon an article centering around millennial complaints. There is certainly sampling bias in an article focusing on complaints, however I felt the following quote from one of the people interviewed for the article sums up this feeling quite well:
Everyone having their own social media and everyone being their own content influencers, it just created a lot more noise. - Tristan T.A Hill, as quoted
What is the next big thing in tech?
After the surveillance capitalism model had matured, tech giants began to look to what was next. Meta chose mixed reality, Twitter flirted with Spaces for a bit (remember when Clubhouse seemed like it might be the next big thing?), before focus turned towards the acquisition. Instagram has tried to transition to more video. None of it has resulted in monumental shifts in direction or profitability, and the metaverse pivot in particular was largely seen as a disaster. All of the experimentation doesn’t matter, so long as there is runway left. Ultimately, what matters most is growth in profits, not growth in users.
Once the runway of user growth is exhausted, it becomes a game of getting more money out of existing users. I don’t know enough about the ad market to know if demand is rising enough to effect prices, but there is a limit to how many ads you can display to users before ruining the service entirely, and companies do have caps on their ad spend that do eventually materialize. Once the ad to content ratio is maxed out, the only lever left to pull is time spent on the app.
Enter AI, which is both a blessing and a curse for social media. Acting in favour of big tech, AI can produce content ad infinitum with very little marginal cost [2]. Furthermore, that content can be tailored to individual users in a way which we haven’t seen yet with algorithmic matching of human-made content. This is especially true for companionship or romantic AI bots, in which one user seemed to report a willingness to pay of $1000 USD per month for access to a model that delivered on the desired features. Companies may view the race to a model that generates that much value as winner-take-all; if someone can spend time with the best model with no limits, they may not want to spend any time with competing products. Contrast this with how many switch between Netflix, HBO, Apple TV+ etc, because no one platform has all the content we want. Viewed through that lens, investment in vast amounts of compute for AI seems worth it.
Acting against big tech is the slop that many users are currently generating in hopes of increasing their own engagement metrics. While this increased engagement is still benefiting tech companies, I don’t think that it is the end game that they want. Instead, they want to be the ones generating the content—this reduces unpredictability in the user experience. With reams of additional data collected over decades on their users, the platform can produce content that has a much lower probability to offend or disgust, and instead engage the user even more [3]. This has already started to occur with Meta generating AI profiles and Instagram generating AI images featuring the user.
The end game is a feed that is the perfect split between AI generated content to keep you engaged, and AI generated ads that show what you would look like in the newest sports car. The dependency on users has been broken, because everyone exists in their own silos. If the company can make a feed that is far more engaging than the others, competitors go out of business and absolute monopolies are formed. Currently, an engaging feed is a combination of having a good algorithm, and having enough users for the algorithm to efficiently source content. In the future, all that will matter is the model.
An escape hatch
There is one remaining requirement that big tech needs to meet in order to achieve that vision; social norms need to shift enough such that it is seen as normal, or even desirable, to exist in such a state. The crux is that I don’t think they’ve figured out how to drive that shift. While iPhone sales did grow after the announcement of Apple Intelligence, they have slumped over the holiday season. Key features, such as the notification summary for news, are being rolled back. Meta rolled back the AI profiles shortly after they were introduced. AI products have moved from people being on wait lists, to being integrated into mainstream products along with price increases. In short, the vibes aren’t there.
The readership of this blog is highly skewed towards users of the fediverse. There is always talk of the fediverse becoming the next big thing. I don’t see that happening, but I also think that is perfectly okay. Rather, I believe that the overall decline of hyperscale social media, to a state in which we return to smaller communities is the next step. Those small communities are where the vibes are best. After years of living in an ecosystem obsessed with follower growth, it is refreshing to finally step off the treadmill. We shouldn’t be looking to start it back up again hoping that it will be different.
In finishing this piece, I’m thinking about the possibility that my view of the whole ecosystem is bias from not being in it. Dinner conversations among friends have given me some confidence that this idea isn’t just a product of who I interact with online, but time will tell. I do feel like the suggestion of quitting mainstream social media no longer comes across as radical, rather possible, even sensible. Even if most are not leaving everything right away, leaving is at least something that many are beginning to think about.
As a good example of this, Pixelfed, a new alternative to Instagram has been developed largely by a single individual. Costs will add up for various instances with more users, but absent that, the service could probably scale to millions of users in much the same fashion that it scaled to its first 100,000. ↩︎
AI queries and content are computationally expensive to generate, however compared to the ad revenue that is obtained through showing the content, it would be difficult for the companies not to make a profit once they get over the hurdle of initial training costs. ↩︎
While outrage may drive engagement when people know there is a human with the opposite opinion on the other end, I don’t think the same thing would hold if the user knew they were arguing with a bot. ↩︎