Late Stage Capitalism
Earlier this week, Zelenskyy provided a brief reprieve of the shadow abyss that is online social media. Through Ukraine’s efforts using drone strikes, Yandex, “the google of Russia”, have been taken offline for the first time, in a long while. The net effect of which has been palpable, at least on social media. For the first time in so far as anyone can particularly remember, the number of trolls and bad faith actors who are attempt to push specific views, or specific points of agenda, have dramatically been reduced. There is much less fighting, division of opinion, and identity politics going on, and, dare I say it, people are agreeing with one another more frequently, or at least, leaving those with differing opinions alone.
While this window of opportunity will only last as long as it takes to stand everything back up, it’s important to understand why it is that a nation state would devote entire datacenters as a means to disrupt or even reduce the central cohesion of a population. After all, who in their right mind would make it a point to sew division, and to what end?
What Russia (and certainly other countries) understand, is that people are in effect easier to control and manipulate others when they feel isolated. As many people online submit ideas or thoughts on a particular topic, if they are berated brigated for having shared an opinion, chances are, they are less likely to share their opinions in the future. After all, why share anything if the reward is punishment?
Bad faith actors understand the value of shaping the cultural zeitgeist, and they understand that this is far easier to do if you can control who it is that is likely to share information, and who it is that is likely to feel empowered to voice their opinion, if that opinion happens to agree with a certain platform or view that buys them certain freedoms. It is in this way, that specific organizations throughout the world manufactures and releases information that is designed to distract and divide people on key issues. When people feel isolated, it is easier to sell particular ideas and mindsets to others.
Get L.O.’s stories in your inbox
Join Medium for free to get updates from this writer.
Similarly, this is why some social media companies make it a point to put individuals into very specific buckets in order to drive engagement. As it turns out, if you collect enough information about a person, you can feed or suppress any kind of natural biases someone may have. There isn’t a lot in the way of regulations or controls around this either, and many people are reporting that their own family members have been co-opted by echo chambers that are not of their own choosing. If an AI can be used to drive engagement, and information can be garnered about someone who engages with a platform, then all that would seem to remain is to turn around and sell access to a given person based on their existing points of view and biases about the world. In our geopolitical environment, especially right now around election time, people’s attentions drive a premium, and the platform reaps the benefits of those that they would entrap with identity politics and manufactured rage. As with any “product” that people use, if you do not pay for the product yourself, you are the product.
Those who understand the value of their own data operate online in somewhat paranoid capacities. Almost every service online drives engagement with enough breadcrumbs to piece together various bits of information about who you are. When services assume that an email address belongs to one person, and track that piece of information across every website, they do so with the tacit understanding that every company who records that data can then turn around and sell it. In a world where TV manufacturers record what you watch and say when the TV is on, search index companies record what it is you search for, licence plate readers record where you go, and your phone’s location data is sold to various applications in order to target you with ads, its no wonder why people feel that everything they do is gamed and sold to the highest bidder — this only enforces the data conscious person to become more paranoid.
And now we have AI companies, who are all vying for ways to take ownership over domains of intellectual property. For many years, the open source community has developed best practices in way of security in order to help other human beings develop code. Entire languages and libraries are provided “for free”, that can be strung together in order to complete any number of projects that could be possibly imagined. Then, over the last few years, those same projects were used to train AI models in order to make approximiations of those same solutions. Mathematicians around the world use these same AI solutions to postulate solutions to long standing problems in the field, only to find, days later, that an AI model has “solved” the problem that otherwise had sat around unsolved for hundreds of years.
Just like with social media, if you are not paying for the solution, your engagement becomes part of the product. Every conversation that a user has with an AI is saved away as part of future training data. Every disagreement a user has with a found solution is later encoded as training to make new data available. This is why companies like Anthropic, OpenAI, Google, Facebook and Grok are all vying for users to interact “for free” with the latest and greatest version of AI — certainly not out of the goodness of their hearts, but because every interaction helps them to refine new ways to seperate tangible ideas away from people who otherwise engage in an interactive conversation with their LLM.
Once individuals understand exactly how this model works, it will unfortunately be too late. You all know that colleague you have that uses AI for everything — to “consult the oracle” of information, even if the takeaway from that impression is negative, that data is still valuable to the company itself. These same quasi-sentient models end up contributing to a moat of information all held behind a paywall. As these companies shoulder the burden for storage of this information, they can then charge access for that information back in the future, when those primary sources are no longer available.
Is this a tower of babel moment for humanity? If knowledge can be consumed and then removed, we may be in for a new dark age, or at least an age in which access to information, which provides agency, is gate kept by a small few.
