Twitter Inc. Chief Executive Elon Musk recently co-signed an open letter along with hundreds of others calling for a six-month pause on the development of artificial intelligence technology. But while doing so, he has secretly been investing “tens of millions” of dollars into Twitter’s own AI projects.

That’s according to a report by Insider, which cites people familiar with the matter as saying that Twitter in recent months has purchased about 10,000 graphics processing units, which are used to train and run AI models, while also hiring a number of AI experts from companies including Alphabet Inc.’s DeepMind.

Insider says Twitter is believed to be working on a large language model project, which suggests it’s building something similar to OpenAI LP’s ChatGPT. One source reportedly said the AI project is still in its early stages, though with Musk (pictured) purchasing such a significant amount of computing power, he’s clearly also extremely dedicated to it.

It’s not clear exactly what kind of hardware Twitter has acquired, but it’s known that OpenAI used thousands of Nvidia Corp.’s A100 GPUs to train ChatGPT. Nvidia has since launched a more powerful GPU known as the H100 that’s said to be several times faster than the A100, so it makes sense that Twitter would invest in the most advanced hardware. Whatever it has acquired, it seems likely that the new chips will be deployed in one of Twitter’s two remaining data centers, with Atlanta said to be the most likely destination.

The reported GPU purchase comes several months after Twitter hired former DeepMind AI engineers Igor Babuschkin and Manuel Kroiss, Business Insider said. Musk apparently is hoping to recruit additional AI talents too.

The billionaire helped to co-found OpenAI back in 2015, and three years later made an attempt to take full control of the company after reportedly disagreeing with its other co-founder Sam Altman. When that takeover attempt failed, Musk quit the company instead. OpenAI is clearly at the forefront of AI developments now while Musk is playing catch-up, so it could well be that the Twitter boss had ulterior motives when calling for research in the area to be temporarily paused.

Charles King of Pund-IT Inc. told SiliconANGLE that Musk’s past shenanigans make it clear that he doesn’t care much what others think of him and that he appears to be immune to the notion he may have erred or acted hypocritically.

“Some commentators have asked whether Musk would have signed the “six month pause” letter if he had succeeded in taking leadership of OpenAI,” King said. “If the Business Insider story is correct, the answer to that question seems clear. But Musk’s interest is hardly surprising. Many in the industry consider AI to be the tech industry’s next great leap forward. As a sponsor, funder and promoter of commercially successful technologies, it would be more surprising if Musk didn’t have plans for developing some sort of differentiating AI platform and solutions.”

As to what Twitter’s AI will actually be used for, Insider says it could improve its search functionality. Another possible use case is beefing up Twitter’s advertising capabilities. That especially makes a lot of sense, as Musk’s antics have notably discouraged a number of advertisers since he took control of Twitter in October.

Photo: Wikimedia Commons