The business has made the source code for Twitter’s recommendation system available on GitHub almost a year after Elon Musk originally suggested doing so.
Musk expressed optimism that consumers would find potential issues in the code and help with its improvement during a Twitter Space conversation.
People will uncover many errors in the so-called algorithm when we first release it, which will be pretty embarrassing, but we’ll swiftly address them.
It’s crucial to remember that the code that was released on Friday only affects how tweets are displayed in Twitter’s “For You” stream. The underlying code for Twitter’s search algorithm or how material is shown in other areas of the platform was not made public.
Musk did confirm that the company would eventually open-source the search algorithm, though.
Created to Benefit Musk
The corporation has described the procedures in its recommendation system, including filtering and rating, in a blog post. Intriguing elements in the code, such as how Twitter’s algorithm identifies Elon Musk as the tweet’s author, have already been found by users.
This discovery might offer a further explanation for why his tweets frequently appear in Twitter feeds. As noted by Jane Manchun Wong, the algorithm also includes labels indicating the author’s status as a “power user” and whether they are a Republican or Democrat.
A discussion about the classification of users by Twitter’s algorithm as Republicans, Democrats, Power users, and Elon Musk took place at a Twitter Space. Musk disagreed with the classification, citing the way that it categorises people according to their political beliefs as “not making sense.”
Later, a Twitter engineer emphasised that the categories were utilised for statistics tracking rather than being connected to the algorithm. The engineer did not explain why Musk had his own category, which astonished Musk. Soon after, the Twitter Space came to an end.