A Meta executive reportedly warned staff last year that 'we have a significant gap' when it comes to AI

• A Meta exec warned last year that the company is falling behind in the AI race, Reuters reports.
• Mark Zuckerberg and other execs reportedly had a five-hour discussion on AI months before ChatGPT's launch.

Meta's head of infrastructure, Santosh Janardhan, warned last year that the company was falling behind when it came to adopting artificial intelligence, according to a recent report from Reuters.

"We have a significant gap in our tooling, workflows and processes when it comes to developing for AI. We need to invest heavily here," Janardhan wrote in an internal memo that was posted on the company's message board in September, per Reuters.

The executive said that bringing AI into the company's main business would mean Meta had to "fundamentally shift our physical infrastructure design, our software systems, and our approach to providing a stable platform," Reuters reported.

The publication reported that, at the time, Mark Zuckerberg and several Meta executives had met for a five-hour discussion on how the company could implement internal systems that would allow it to perform "cutting-edge artificial intelligence work."

AI systems require more computing power, as its training requires the system to sift through a large amount of data. An analyst previously told The Information that OpenAI's chatbot could cost as much as $700,000 a day due to its need for computing power and expensive servers.

A Meta spokesperson did not respond to a request for comment from Insider ahead of publication, but told Reuters that the company "has a proven track record in creating and deploying state-of-the-art infrastructure at scale combined with deep expertise in AI research and engineering."

Reuters reported that the company was slow to invest in AI-friendly computer chips until last year, instead choosing to run its AI systems on cheaper, but slower chips. In January 2022, Nvidia announced that Meta was using its chips to build a "massive AI research supercomputer."

The company's pivot toward the metaverse in 2021 also slowed the company's focus on AI, four sources told Reuters.

Before PayPal Mafia member Peter Thiel resigned from Meta's board in February 2022, the Silicon Valley investor told Zuckerberg and other leaders that the company's focus on the metaverse had hobbled its attempts to compete with TikTok, Reuters reported, citing two sources.

Meta is one of several tech companies rushing to join the AI race after ChatGPT took the spotlight in November. At the time, Google's management reportedly issued a "code red" warning over the popularity of the chatbot and redirected some of its teams to focus on AI products. Reuters reported that Meta similarly began focusing on converting its generative AI research into sellable products after the launch of ChatGPT in November.

A Meta spokesperson told Reuters Meta has been working on generative AI products for over a year.

Earlier this month, Meta CTO Andrew Bosworth said that Zuckerberg and other top executives were spending most of their time on generative AI and said the company has been investing in the technology for over a decade. In February, the company released its large language model, "LLaMA," to researchers.

Insider previously reported that some analysts were expressing concern that the company had begun spending too much on AI amid its decision to purchase thousands of pricey AI-friendly chips from Nvidia.