After Gutting Twitter’s Labor Force, Elon Musk Appears To Be Eyeing Reduced Dependency on Google Cloud’s AI Training Platform

Rohail Saleem
Elon Musk Twitter AI
Image Source: MedPage Today

This is not investment advice. The author has no position in any of the stocks mentioned. Wccftech.com has a disclosure and ethics policy.

Elon Musk these days is a content writer’s dream come true, providing the choicest morsels of juicy tidbits at an admirably regular pace. While much of the media continues to examine the Tesla CEO’s Twitter space comments of last night, our attention remains glued on Musk’s recent massive purchase of GPUs.

https://twitter.com/chancery_daily/status/1646017125637578759

Related Story CXMT Debuts Its Domestically-Produced DDR5 Memory: 8000 MT/s DDR5 & 10667 MT/s LP5X For Servers, Desktops, & Laptops

It would not be a quintessential Musk interaction if it did not provide the usual dose of frivolity. In yesterday’s Twitter Space meetup hosted by the BBC and later characterized by the host as simply “bizarre,” Elon Musk said that his dog Floki was, in fact, the actual CEO of Twitter. Bear in mind that Twitter had temporarily changed its logo to one that depicted Dogecoin’s Shiba Inu a few days back.

On a more serious note, Elon Musk informed the audience that Twitter was “roughly breakeven” right now and that the company just might become cash flow positive in the current quarter.

However, when looking under the hood, the vast majority of this financial improvement has come as a result of unprecedented spending curbs. At the Twitter Space meetup, Elon Musk noted that Twitter has now lost 81 percent of its employees relative to the headcount in October 2021. The company now retains a threadbare headcount of just around 1,500 employees vs. the 7,800 workers that were on its payroll back in October. As for the good news, the CEO of Twitter commented that almost all of the advertisers that had boycotted the social media platform in the early days of his takeover have since then returned.

Elon Musk also tried to reassure investors that his sale of Tesla stock did not indicate a “lack of faith in Tesla.”

Musk did spend quite a lot of time debating with BBC on Twitter’s content moderation policies, arguing that problematic content should be countered by providing the correct information and context instead of issuing sweeping bans.

While the interaction did provide some important information regarding Twitter’s future trajectory, our attention has been snagged by a concurrent development that has not been adequately analyzed by the broader media amid the general pandemonium over yesterday’s Twitter Space meetup.

Elon Musk Has Purchased 10,000 GPUs For a New AI Project That Should Pave the Way For Reduced Dependency on Google Cloud’s AI Training Platform

As per the reporting by Business Insider, Elon Musk has hired engineers from Alphabet’s subsidiary, DeepMind, and already purchased around 10,000 GPUs to run a new AI project that involves a Large Language Model (LLM). While the report does not specify, the GPUs are likely to be NVIDIA’s A100 or H100 units.

Do note that back in February 2021, Twitter entered into a multi-year partnership with Google Cloud’s AI training Platform (GCP). Under the terms of the deal, the social media giant would extract critical insights from the trillions of events that are ingested by its data platform every day by leveraging Google’s tools such as BigQuery, Dataflow, and Cloud Bigtable.

However, given Elon Musk’s aggressive austerity drive at Twitter, coupled with the fact that he is now poaching talent from a subsidiary of Google’s parent Alphabet, it stands to reason that the global town square's partnership with GCP remains at risk.

Elon Musk is a fan of vertical integration, as is evident by examining Tesla’s modus operandi. It seems that the CEO of Tesla is now trying to introduce the same level of vertical integration at Twitter by sanctioning an ambitious LLM project, whose success would pave the way for the social media platform to reduce its reliance on the GCP and bring an ever-greater number of data-related exigencies in-house.

Follow Wccftech on Google to get more of our news coverage in your feeds.

Button