Meta’s AI development isn’t going quite as well as billionaire CEO Mark Zuckerberg had hoped. Once a frontrunner in the AI race, a year’s worth of technical failures and setbacks to his company’s “state-of-the-art” Llama 4 Behemoth have spurred Zuckerberg to pull out all the stops.
Now, the second richest man on the planet is said to be handpicking a team of some 50 AI researchers for a shadowy “superintelligence group.” To do so, he’s set aside some $15 billion for a 49 percent stake in Scale.ai — the despotic AI startup that’s been accused of using wage-slave labor to train chatbots, commiting systemic wage theft, and is contracted to run the Pentagon’s “flagship” weapons automation program.
Zuckerberg’s goal, according to the New York Times, is to be the first to develop the first AI system that exceeds the abilities of the human brain — a dream, it’s worth noting, which the majority of AI researchers still think is “very unlikely.”
Scale.ai is probably the biggest player in the data labeling game, hiring third world laborers to do the laborious task of writing, scraping, and cataloguing data for companies to train their AI systems. It’ll be in good company at Zuckerberg’s operation; Meta was recently caught scraping over seven million copyrighted books to feed Llama, leading to a major class action lawsuit.
“This is like the AI equivalent of the Marshall Plan,” wrote tech critic Ed Zitron, referring to the post-WWII scheme that made the US the global dynasty it is today. “[Meta’s] running out of training data (if they haven’t already) and all they have left is one company that generates it using sweat shop labor. Gotta wonder why this much money too.”
Per Bloomberg, the “superintelligence” initiative is Meta’s largest external investment to date. As part of the investment, Scale.ai co-founder Alexandr Wang is set to join Zuckerberg’s secret society of AI developers. The Meta CEO even rearranged the desks so that his new team could “sit near him,” according to reporting.
As part of his shadow cabal recruitment drive, Zuckerberg highlighted Meta’s ad-based revenue stream, suggesting that AI developers would be free from investor scrutiny if they came to work for him, according to Bloomberg. The CEO also stressed that he had enough cash tucked away to erect a “multi-gigawatt data center,” which if ever realized be among the most powerful in the world. (For context, it currently costs about $30 billion for a one-gigawatt data center, which is about the amount it takes to power 750,000 homes.)
It’s not currently known who else will be joining Meta’s AI scrum — though the company is reportedly offering potential recruits anywhere from seven to nine figures to join. If Wang is any indication, Zuckerberg’s desk is sure to be flanked by a league of headline-grabbing tech personalities.
Breathlessly described as the “world’s youngest self-made billionaire” — as if any billionaire is self-made — Wang has become a darling of the US defense industry as of late. A notorious China hawk, Wang’s cozy deals with the Pentagon come alongside drastic escalations in anti-China rhetoric, declaring an all-out “AI war” with the nation.
He’ll be a major asset for Zuckerberg, whose own collaboration with the US military apparatus are well known, and ramping up in recent months. The tech billionaire has previously given the federal government carte blanche to spy on US citizens through data scraped from Facebook. He eventually went on to become an outspoken champion of free speech, even as he helped build a “clean” global internet closed off to Chinese nationals.
Even if his “superintelligence” scheme doesn’t work out — remember the metaverse, a scheme he was so obsessed with that he renamed the entire company? — Zuckerberg should have more than enough dough serving up brainrot to the elderly. There’s always money in the banana stand.
More on Meta: Meta Says It’s Okay to Feed Copyrighted Books Into Its AI Model Because They Have No “Economic Value”