No really, Elon Musk seems to be making good on his promise a week ago to open up the X recommendation algorithm for public scrutiny and input, apparently making it the main feed on his open source social media platform. He previously promised to do this by 2022, and he did so by publishing a single code snippet soon after, but that cart hasn’t been kept up-to-date enough to make the X platform worthy of the crowd’s vision of an open source product.
This release, then, is a promising step in the direction of X being a truly open product. The next step will be to update this final code in four weeks, as Musk promised to do.
However, this release does not mean that an open release of X can be marked as “promise kept.” In his January 10 X letter promising this release, Musk said he would release “all codes used to determine which live posts and ads are recommended to users.” From where I’m sitting that’s not even close to happening.
That’s because on November 26 of last year, Musk and Grok’s accounts posted that Grok is being used to automatically filter posts in everyone’s feed, though it can be switched from “popular” to “recent” to do it in chronological order. That algorithm doesn’t seem to exist. The following and For You feeds on X also have ads, which Musk revealed are served by an algorithm that he said he will make public. So by my count there should be at least two more releases, maybe more.
Gizmodo has reached out to X for information about whether or not the advertising and tracking feed code has been released, or will be released in the future. We will update when we hear back.
But anyway, here we have a new dump code. The first thing you need to know is that it “pulls,” according to Musk.
Yes, the algorithm is simple
– Elon Musk (@elonmusk) January 20, 2026
Earlier that day Musk said that the algorithm is absorbing, the head of X product Nikita Bier seems to show that he is proud of it, noting that in the six months from July of 2025 to this month, the time of daily engagement from new users went from less than 20 minutes to somewhere in the mid-30s. Who is right? Is it better than before, or worse?
The problem may be that Musk can’t seem to clean up all of the stubborn wake-up residue that was put on X back when it was called on Twitter. His tweet claiming to be a sucker was a response to former video game executive Mark Kern complaining that the algorithm weights posts too slowly when they come from heavily blocked accounts. Kern says he suspects this biased the algorithm against posts from right-wing accounts like his. It’s obvious I think, although it almost biases the algorithm against accounts that post a lot of abuse and abuse, so make of it what you will.
Judging by what’s in the plain text reading documentation in the Github dump, this latest X algorithm is what you might expect if you’re using X: an update on TikTok’s way of hooking users. My opinion of this description is that, surprisingly, it prioritizes engagement, trying to figure out which post will make the user stop scrolling. It pulls from the accounts you follow, but also the accounts you follow. It appeals to your id, not your superego. No matter what you think you’re there to see, it wants to show you anything that will keep you staring.
In addition to breastfeeding, Elon Musk also says it’s “dumb.” In response to a complaint from blogger Robert Scoble complaining that the algorithm favors posters who steal news events, Musk says the algorithm will improve every month—apparently referring to the expected four-week cadence of GitHub code dumps.
We try to make the algorithm not dumb.
It will improve every month.
– Elon Musk (@elonmusk) January 20, 2026
And who knows, maybe users with amazing ideas will not just grab the readme sections, but directly into the code, find real problems, and pass on suggestions to Musk, and the algorithm will get more satisfaction and benefit in the long run. Also, maybe the needs of a company that wants to engage users to make them watch ads and make money for themselves, and the desires of people who want to feel informed and happy are two completely irreconcilable concepts, and making an open source algorithm recommendation to try and provide both of those types of needs is futile. I guess we’ll see which of these is probably true.
