Elon Musk delivers on promise as Twitter releases part of its source code, including algorithm for recommending Tweets

In a move towards greater transparency, Twitter CEO Elon Musk has delivered on his promise to make a portion of the platform's source code public, including the algorithm that recommends tweets to users' timelines.

“The goal with open sourcing the Twitter algorithm is to build trust and transparency with users. Build trust through transparency, I should say. In that, I don’t think we should trust any social media algorithm that is a black box. You don’t know what’s going on in there,” said Musk on Twitter Spaces, adding that he wants Twitter “to be the most trusted place on the internet, where you’re talking to real people, and why you know things are happening on Twitter, and that it’s not, or the least, game-able system on the Internet … is our goal."

Twitter has published two repositories on GitHub, which include code for many elements that power the social network, such as the mechanism behind the For You timeline. This move is considered a "first step to be[ing] more transparent," while simultaneously minimizing risks to Twitter and its users.

During a Twitter Spaces session, Musk acknowledged that the initial release might contain errors but emphasized that they would be quickly addressed. He compared the company's transparency effort to the open-source operating system Linux, suggesting that the community could help identify and fix potential issues.

“Our initial release of the so-called algorithm is going to be quite embarrassing, and people are going to find a lot of mistakes, but we’re going to fix them very quickly,” Musk said. “Even if you don’t agree with something, at least you’ll know why it’s there, and that you’re not being secretly manipulated … The analog, here, that we’re aspiring to is the great example of Linux as an open source operating system … One can, in theory, discover many exploits for Linux. In reality, what happens is the community identifies and fixes those exploits.”

The released code does not include Twitter's ad recommendations or data used to train the recommendation algorithm. Furthermore, it provides limited instructions for inspecting or using the code, indicating that these releases are primarily aimed at developers.

Twitter emphasized that it excluded any code that could compromise user safety and privacy or facilitate malicious activity on the platform, including undermining its efforts against child sexual exploitation and manipulation.

This assurance comes just weeks after Twitter dismissed much of its trust and safety staff, responsible for content moderation and other user security-related tasks. Nonetheless, the company insists that it has taken steps to protect user safety and privacy with the code release.

Twitter is currently developing tools to manage code suggestions from the community and synchronize changes to its internal repository. The company is open to feedback on potential improvements and envisions an evolving process.

“We’re going to look for suggestions, not just on bugs but also on how the algorithm should work,” Musk said on the Spaces session. “It’s going to be an evolving process. I wouldn’t expect it to be a nonstop upward movement… but we’re very open to what would improve the user experience.”

As detailed by TechCrunch, "the algorithm is fairly complex — but not necessarily surprising in any way from a technical standpoint.”

The publication reports:

"It’s made up of multiple models, including a model for detecting 'not safe for work' or abusive content, determining the likelihood of a Twitter user interacting with another user and calculating a Twitter user’s 'reputation.' (It’s unclear what 'reputation' refers to, exactly; the high-level documentation isn’t clear on that.) Several neural networks are responsible for ranking the tweets and recommending accounts to follow, while a filtering component hides tweets to — forgive the jargon — 'support legal compliance, improve product quality, increase user trust, protect revenue through the use of hard-filtering, visible product treatments and coarse-grained downranking.'"

Twitter's engineering blog post provides additional information about the recommendation pipeline, which reportedly runs approximately five billion times per day. It explains that the For You timeline typically comprises 50% tweets from people a user doesn't follow and 50% tweets from those they do follow. This ratio, however, may vary between users.

Ian Miles Cheong

Contributor

Ian Miles Cheong is a freelance writer, graphic designer, journalist and videographer. He’s kind of a big deal on Twitter.

https://twitter.com/stillgray

COMMENTS

Be the first to comment

Please check your e-mail for a link to activate your account.