What Happened To Algorithms?? how AI is changing the feed

There is a new breed of algorithm that is quickly becoming the norm on every major social and you're probably noticing it on a kind of implicit level because it fundamentally shifted the way that content is ranked in feeds. In short, LLM's have very recently been incorporated into algorithms and this kind of chat GPT-ification of the world is a trend that'll likely continue, I.E. large language models becoming more and more enmeshed in the systems that we use daily. I wanna explain the difference between the algorithms that we've known up until nowish and then the incoming algorithms that will define this unknown era of social media. First, it's accurate to say that all algorithms have used AI since 2012ish when suggested feeds began, but not LLMs, just machine learning, which is basically very sophisticated pattern matching. Old algorithms track things like clicks, likes, accounts that users have engaged with and content that users have engaged with. They observe these patterns across astronomical subsets of data and then make really good predictions about what else you might like. And this is the key part, given your historical behavior, this is referred to as correlation. Now new LLM fueled algorithms can read context and actually comprehend. This means that content is suggested based on topics and relevancy, not just historical engagement. So big followings and track records of virality don't necessarily hold as much weight as they used to. That's scary for people who built a following. But the upside is that younger, smaller accounts can see outsized reach and impressions. I feel like I've experienced this myself. Meta, Pinterest, LinkedIn, and YouTube have all incorporated LLMs. And while nothing has been confirmed by TikTok, my guess is that they were probably the earliest one to it because their algorithm has always been scary good. I think this shift is pretty big news and yet I'm still not seeing a ton of media coverage about it. But I think it's important that we understand that the underlying reward system has changed.