[Sketchiest Guy in the World Voice] Hey kid, do you want to see Algorithm X? It’s just over here.
No, really, Elon Musk appears to be making part of his promise about a week ago to open the X recommendations algorithm to public reading and contribution, theoretically making the main feed of his social media platform open source. It previously promised it would do so in 2022, and somehow did so by releasing a snapshot of the code shortly after, but that repository wasn’t up to date enough for Platform X to be considered most people’s idea of an open source product.
This version therefore constitutes a promising step towards X becoming truly an open source product. The next step would be to update this code repository in four weeks, as Musk indicated. promised he would.
Even then, this release would not mean that X’s open source could be marked as “promise kept”. In his Jan. 10 X post promising the post, Musk said he would release “all code used to determine which organic and advertising posts are recommended to users.” From where I sit, that’s not even close to happening yet.
This is because on November 26 last year, the accounts of Musk And Grok posted that Grok is used by default to sort posts on everyone’s news feed, although it can be toggled from “popular” to “recent” to make it chronological. This algorithm seems to be lacking. The Follow and For You feeds on So by my calculations there should be at least two more versions, maybe more.
Gizmodo has reached out to X to find out if the following ad and feed code has already been released, or if it will be released at some point in the future. We will update if we receive a response.
But anyway, here we are with a new code dump. The first thing you need to know is that “it’s rubbish”, according to Musk.
Yes, the algorithm sucks
– Elon Musk (@elonmusk) January 20, 2026
Earlier the same day, Musk told X product manager Nikita Bier that the algorithm sucked. seemed to indicate that he was proud of itnoting that in the six months from July 2025 to this month, daily engagement time for new users increased from less than 20 minutes to around 30 minutes. Who is right? Is it better than ever, or does it suck?
Perhaps the problem is that Musk just can’t seem to clean out all the stubborn woke residue introduced into X back when it was called Twitter. His tweet saying it sucked was a response to former video game director Mark Kern. complain that the algorithm weights posts less if they come from accounts that have been blocked a lot. Kern says he suspects this biases the algorithm against posts from right-wing accounts like his. It’s plausible, I suppose, although it almost certainly biases the algorithm against accounts that post a lot of harassment and abuse, so make of that what you will.
Judging by the contents of the plain text readme documents in the Github dump, this latest X algorithm is what you’re probably expecting if you’re using X: an update to the TikTok method of hooking users. My impression of what is described is that, unsurprisingly, it prioritizes engagement, trying to determine which posts will stop the user from scrolling. It extracts the accounts you follow, but also the accounts deemed similar to those you follow. This appeals to your identity, not your superego. No matter what you think you see, it wants to show you everything that will make you keep watching.
Besides being rubbish, Elon Musk also says it’s “stupid.” Responding to a complaint from blogger Robert Scoble complaining that the algorithm favors posters who hijack current events, Musk says the algorithm will improve every month – apparently referring to the planned four-week cadence for GitHub code dumps.
We are trying to make the algorithm less stupid.
It will get better every month.
– Elon Musk (@elonmusk) January 20, 2026
And who knows, maybe users with amazing ideas will dig not only into the readme sections, but directly into the code, find the real problems and pass their suggestions to Musk, and the algorithm will become more satisfying and profitable over time. Alternatively, perhaps the needs of a company that wants to attract users to watch ads and generate revenue for itself, and the desires of human beings to feel knowledgeable and happy are two completely irreconcilable concepts, and creating an open source recommendation algorithm to try to meet both types of needs is completely futile. I guess we’ll see which of these maybes is actually true.