Twitter source code explained

Twitter has officially posted parts of its source code in two main repositories on the code-sharing platform GitHub. The question is, what exactly is being revealed in the cache, and will the move end up being a win for both users and the company? Let's find out.

After months of speculation, tech professionals, as well as the average tech enthusiast, have been given the chance to examine the internal workings of the popular social platform and suggest modifications to its algorithm.

CEO Elon Musk followed through on his pledge to open parts of Twitter source code, which he had been hinting at for almost a year.

Calling themselves the “town square of the internet,“ the company stated, “At Twitter 2.0, we believe that we have a responsibility…to make our platform transparent. So today we are taking the first step in a new era of transparency and opening much of our source code to the global community.”

Much of the excitement has been focused around the reveal of Twitter’s sought-after recommendations algorithm – the controlling factor for which tweets users see on their timeline.

Twitter has been under the microscope in recent years over how their algorithms actually work.

“The real magic of Twitter is in our recommendations algorithm, which powers the hit Tweets you see in your For You timeline. We broke down how it all works here,” engineers tweeted.

An engineer blog post released, in conjunction with the source code, explained why the recommendation algorithm – which distills up to 500 million Tweets posted daily into only a handful of tweets that users ultimately see – is so important to the workings of the app.

“The foundation of Twitter’s recommendations is a set of core models and features that extract latent information from Tweet, user, and engagement data,” the blog said.

“Acid test is that independent third parties should be able to determine, with reasonable accuracy, what will probably be shown to users” Musk said to his Twitter followers on Friday.

“No doubt, many embarrassing issues will be discovered, but we will fix them fast!, he tweeted.

The company also posted a blog Friday, detailing exactly what its engineers made available in the two repositories, named the ‘main repo’ and ‘ml repo.’

“We invite the community to submit GitHub issues and pull requests for suggestions on improving the recommendations algorithm,” Twitter stated.

The company said its currently developing a tool that can handle the suggestions and “sync changes to our internal repository.”

Musk, who held a session on Twitter’s audio chat feature Spaces after the launch, and said more code will be revealed in time.

During the session Musk asked users to present questions and recommendations about how the platform's code works.

Musk also said Twitter will update its recommendation algorithm based on user suggestions every 24 to 48 hours.

Twitter excluded “any code that would compromise user safety and privacy or the ability to protect our platform from bad actors, including undermining our efforts at combating child sexual exploitation and manipulation.”

The company said they also excluded the code that controls Twitter ad recommendations.

To guarantee the safety and privacy of users, Twitter said it would “not release training data or model weights associated with the algorithm at this point.”

For those curious what to do if they run across any security concerns? All concerns are still going through the HackerOne bug bounty program, Twitter said.