Fellow CEO Mark Zuckerberg told investors that the newly-created Metas Super Intelligence Lab is committed to building AI models that can rise up on its own, meaning they can learn from themselves without human opinions. The comments were made on Wednesday’s second-quarter earnings call.
“To some extent, [it’s] “It’s not just about learning from people, because you want to build something smarter than people. It’s a very basic thing and will have a wide impact on how we make products and how we run a company,” Zuckerberg said.
This is one of some comments the CEO has made over the past 12 hours to show how Meta’s new research lab plans can compete with competitors like OpenAI, Google DeepMind and Anthropic.
Early this morning, Zuckerberg posted a letter online and shared a video to the Instagram reel, saying Mehta’s super-intelligence efforts will focus on building “personal super-intelligence,” which will provide people with their own tools to empower and improve the world. “This is different from others in the industry, who think that superintelligence should focus on automating all valuable work and then humans will live in the leeway of their output,” he said.
Another key part of Meta AI strategy is SmartGlasses, Zuckerberg said he believes “basically an ideal form of AI.” Since their debut in 2023, Meta has partnered with Ray-Ban to sell 2 million pairs of smart sunglasses. Last September, the company also showed off a futuristic prototype of augmented reality glasses. (The prototype known as Orion is not expected to be shipped to consumers, but Meta says it plans to build new equipment in the next few years according to Orion research.)
For now, Meta seems to make some difference between AI, which helps monetize its core products such as Instagram and WhatsApp and super smart AI that can one day help the future of PARE PARE PARE PARE PARE PARE PARE PARE PARE PARE PARE PARE PARE SHATE.
Zuckerberg’s comments about superintelligence followed closely behind an expected earnings report. The company lags behind competitors in the AI competition, and it has spent billions of dollars to build a small, powerful super-intelligent lab that will focus on building Frontier AI models.
Since the lab was announced publicly last month, Meta has been actively recruiting frenzy, introducing Silicon Valley agents such as Alexandr Wang, Nat Friedman and Daniel Gross, and providing AI researchers with pay of up to nine numbers to attract them into the lab. Wang is part of a “acquisition” agreement reached by Scale AI, a startup he co-created AI data-tagged. Former Openai researcher Shengjia Zhao has been appointed as the head of Meta’s new research laboratory.
Meta revised its expectations for capital expenditure for the year, increasing its forecast to $69 billion. Meta’s chief financial officer Susan Li said employee compensation is one of the biggest drivers, as well as the AI infrastructure. Although the company spent money to build its AI products, it also predicted that the outlook for the third quarter was better than expected, with revenues expected between $47.5 billion and $50.5 billion.
