The company is using LLMs for algorithm personalization.
Meta
If you've ever scrolled through your Facebook feed and wondered why you're suddenly flooded with weird AI-generated celeb bait and posts from cast iron pan enthusiasts (just me?), Meta has a new tool that might help. The company is rolling out a new algorithm personalization feature that aims to make it easier for users to explain exactly what they do and don't want to see in their recommendations.
With the update, you can dive into specific feed posts and Reels to access a new "your algorithm" menu. From there, you can add specific topics to your interests and to your "not interested in" list. There's also a box where Meta prompts you to describe in your own words what you want to see in your feed.
The update mirrors a similar change that landed on Instagram earlier this year. The update is possible, Meta says, because it now uses large language models in its ranking systems. Instead of simply "inferring what you're into from what you engage with on Facebook," the company explains, "LLMs can incorporate the interests, context, and intent behind natural language, unlocking an algorithm you can actually talk to."
The change has yet to roll out on my app, so I haven't been able to try it yet. The Instagram version of the feature had some limitations, though. It didn't allow me to add "posts from people I follow" as one of my interests or request fewer ads. Given Meta's previous efforts to clean up the amount of spam and engagement bait in the Facebook feed, I am curious whether users will be able to request more holistic changes, like "less AI-generated content" or "no more Reddit copypasta." At the very least, I am looking forward to finally telling Facebook that I actually have no special interest in cast iron pans.