Twitter Makes Some of its Source Code Public

Twitter app logo is seen in this illustration taken, August 22, 2022. REUTERS/Dado Ruvic/Illustration/File Photo
Twitter app logo is seen in this illustration taken, August 22, 2022. REUTERS/Dado Ruvic/Illustration/File Photo
TT

Twitter Makes Some of its Source Code Public

Twitter app logo is seen in this illustration taken, August 22, 2022. REUTERS/Dado Ruvic/Illustration/File Photo
Twitter app logo is seen in this illustration taken, August 22, 2022. REUTERS/Dado Ruvic/Illustration/File Photo

Twitter on Friday made public parts of the computer code that decides how the social media site recommends content, with its owner Elon Musk adding that the entirety of the code will be available in the next few weeks.

The announcement will allow users and programmers a peek into its workings and the ability to suggest modifications to the algorithm.

"In the coming weeks, we will open source literally everything that contributes to showing a tweet," Musk said in a tweet on Saturday.

The company said in a blog post it had uploaded the code in two repositories on code-sharing platform Github. They include the source code for many parts of Twitter, including the recommendations algorithm which controls the tweets that users see on their timeline.

The move comes at the behest of Musk, its billionaire owner, who has said code transparency would lead to higher trust among users and rapid improvements to the product.

It also serves to address common concerns among users and lawmakers, who are increasingly scrutinizing social media platforms over how algorithms select the content that users see.

Musk tweeted on Friday that third parties should be able to analyze the open-sourced code and "determine, with reasonable accuracy, what will probably be shown to users."

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

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

On Friday, Musk and some Twitter employees held a session on Spaces, Twitter's audio chat feature, asking users to bring recommendations and questions about how the platform's code works.

One person questioned why Twitter's code appeared to classify users as Republicans or Democrats. A Twitter employee responded that it was an old feature that was not important to the platform's recommendation system, and the company was looking to remove it.

The repositories on Github do not include the code that powers Twitter's ad recommendations, the company said.

It also said it excluded code that would compromise user safety or privacy, as well as details that would undermine efforts to prevent child sexual abuse material on the platform.

The news also comes after parts of Twitter's source code were leaked on Github, which took down the code last week at Twitter's request.

Twitter asked the US District Court for the Northern District of California to order Github to produce "all identifying information" associated with the Github account that had posted the leaked code, according to a legal filing.



AI Cloud Provider SMC Plans Global Rollout

People attend a media tour of Sustainable Metal Cloud's Sustainable AI Factory in Singapore July 25, 2024. REUTERS/Caroline Chia/File Photo Purchase Licensing Rights
People attend a media tour of Sustainable Metal Cloud's Sustainable AI Factory in Singapore July 25, 2024. REUTERS/Caroline Chia/File Photo Purchase Licensing Rights
TT

AI Cloud Provider SMC Plans Global Rollout

People attend a media tour of Sustainable Metal Cloud's Sustainable AI Factory in Singapore July 25, 2024. REUTERS/Caroline Chia/File Photo Purchase Licensing Rights
People attend a media tour of Sustainable Metal Cloud's Sustainable AI Factory in Singapore July 25, 2024. REUTERS/Caroline Chia/File Photo Purchase Licensing Rights

Singapore-headquartered AI cloud provider Sustainable Metal Cloud (SMC) is planning to expand globally as its sees fast-growing demand for its energy saving technology, its CEO said on Thursday.

"Due to client demand, we’re looking to expand in EMEA (Europe Middle East and Africa) and North America," CEO and co-founder Tim Rosenfield said, Reuters reported.

The startup, a partner of AI chip giant Nvidia, already operates what it calls "sustainable AI factories" in Australia and Singapore and is set to launch in India and Thailand.

Its clients in Singapore, where it operates over 1,200 of Nvidia's high-end H100 AI chips, include Facebook owner Meta who uses SMC's cloud to run its Llama 2 AI model.

While most data centres depend on air cooling technology, SMC uses immersion technology, submerging servers from Dell fitted with GPUs (graphics processing units) from Nvidia in a synthetic oil called polyalphaolefin to draw heat away faster.

The technology behind the approach reduces energy consumption by up to 50% compared to traditional air cooling, according to the CEO.

Demand for AI is expected to increase 10-fold compared with 2023, according to the International Energy Agency (IEA).

The electricity consumption of data centres globally is expected to top 1,000 terawatt-hours in 2026, roughly equivalent to Japan's total annual consumption, the IEA said in March.

SMC is currently raising $400 million in equity and $550 million in debt according to a source with direct knowledge of the matter.

The company declined to comment. The fundraising was first reported by Bloomberg.