Elon Musk, the CEO of Twitter, has reportedly started a substantial artificial intelligence project within the company despite his calls for a halt to AI training across the sector. According to a Business Insider story, the corporation has purchased about 10,000 GPUs (graphics processing units) and hired AI specialists from DeepMind for the project, which focuses on a large language model (LLM).
According to two unnamed individuals in the story, the CEO of Twitter has purchased close to 10,000 GPUs for use on the website. Due to the significant processing needs of large-scale AI models, GPUs are typically used in their management. Elon Musk declared his aim to use AI for the goal of “detecting and drawing attention to the manipulation of public opinion” on Twitter in a post on March 18, and this choice appears to be in keeping with that message.
Twitter has acquired certain GPUs, although their exact specifications are currently unclear. Nevertheless, despite persistent financial issues that Elon Musk has referred to as a “unstable financial situation,” the corporation is claimed to have spent tens of millions of dollars on purchasing these computing GPUs.
Atlanta is considered to be the most likely location for the GPU deployment in one of Twitter’s two remaining data centres. Strangely, Musk shut down Twitter’s main data centre in Sacramento near the end of December, a decision that definitely would have decreased the company’s computational power.
The engineering team at Twitter is growing.
For its generative AI effort, Twitter is also expanding its engineering team in addition to purchasing GPU hardware. Igor Babuschkin and Manuel Kroiss, two engineers from DeepMind, an Alphabet subsidiary engaged in AI research, have recently joined the company. Since at least February of this year, Elon Musk appears to have been aggressively looking for AI expertise to compete with OpenAI’s ChatGPT.
A few weeks ago, Elon Musk and thousands of other technologists issued an open letter urging a temporary halt to AI development due to the danger it poses to humanity. This new development has taken place in response.
Also in 2017, while speaking to the United States National Governors Association, Musk shared his worries about AI. He urged lawmakers to restrict AI development “before it’s too late” at the occasion.
Large corporations like Twitter frequently buy gear in bulk so they can benefit from discounted prices. For instance, Nvidia’s H100 boards can cost more than $10,000 each when bought separately from merchants like CDW. This shows that Twitter’s AI programme may have required significant hardware expenditures.