Meta CEO Mark Zuckerberg’s newly announced mission to deliver “personal AI superintelligence for everyone” represents his second attempt in four years to redefine humanity’s technological future. The freshly formed Meta Superintelligence Labs, staffed with high-profile acquisitions from OpenAI and other rivals, aims to reach the “frontier” of AI capability within a year, positioning this as “the beginning of a new era for humanity,” according to an internal employee memo. Yet this bold vision bears a striking resemblance to Zuckerberg’s 2021 metaverse proclamation, which promised a revolutionary virtual universe where users could “do almost anything you can imagine” as “the next version of the Internet”.
The Ghost of Metaverse Past
The parallels between both initiatives trigger skepticism as Meta’s metaverse ambitions yielded minimal tangible success despite massive investment. The company rebranded from Facebook to Meta in 2021, signaling Zuckerberg’s conviction that the metaverse would replace mobile internet. Pre-rendered concept videos depicted users playing zero-gravity poker with robot avatars, teleporting to Tokyo concerts from their living rooms, and elderly friends playing chess across continents.
Reality diverged sharply from these projections. By 2025, Meta’s flagship Horizon Worlds platform faced widespread criticism as a “depressing” and “lonely” experience plagued by “empty” virtual venues 1. Early adopters reported environments overrun by unsupervised children, with one user describing it as “kindergarten” during a virtual date. Technical limitations compounded these issues—employee tools were reportedly avoided internally, while the codebase was described as “a 3D version of a mobile app”. A $50 million creator fund failed to incentivize sustained development, leaving Horizon Worlds dwarfed by alternatives like VRChat.
Visually, Meta’s metaverse faced brutal reception. When Zuckerberg unveiled his avatar in 2022, critics slammed its “soulless” and “dead-eyed” appearance. Artist Grimes declared Zuckerberg “unfit to run the metaverse,” while others noted indie games offered superior aesthetics. Corporate ventures like JPMorgan’s Onyx lounge in Decentraland became ghost towns featuring “just a tiger, a picture, and a random avatar”.
The AI Ambition Reset
Zuckerberg now positions AI as the true heir to internet evolution. His April 2025 interview outlined a near-future where users interact dynamically with AI entities instead of passive content consumption: “You can talk to it, or interact with it, and it talks back… Or you can jump into it like a game”. The vision expands beyond entertainment. Zuckerberg anticipates AI solving diseases, advancing science, and creating “a world of abundance” where everyone accesses “superhuman tools”.
Meta reports one billion monthly users for its existing AI products, dwarfing metaverse engagement 1. The company aggressively recruited AI talent, offering compensation packages reportedly reaching $300 million over four years (though Meta disputes these figures). Unlike the metaverse’s hardware dependency, AI integrates directly into Meta’s core revenue engines: advertising algorithms, content recommendation systems, and user analytics across Facebook, Instagram, and WhatsApp.
Skepticism and Structural Challenges
Industry observers note troubling similarities between Zuckerberg’s execution patterns: grandiose promises of societal transformation, aggressive talent acquisition, and downplayed technical hurdles. “Reading that memo, I couldn’t help but think of another ‘vision for the future’ Zuckerberg shared not that long ago,” wrote one tech analyst, referencing the metaverse hype cycle.
Technical barriers remain substantial. Current AI models still hallucinate facts, struggle with reasoning, and falter on basic tasks like children’s games. Agentic AI systems, which autonomously execute real-world tasks, face projected failure rates exceeding 40% by 2027 due to cost and unclear ROI. Energy demands also pose existential questions, with advanced LLMs emitting 50× more CO₂ per query than smaller models.
Dr. Evelyn Tan, a Stanford AI ethicist, questions the feasibility timeline: “Superintelligence requires breakthroughs we haven’t even conceptualized. Framing this as a one-year sprint dangerously oversimplifies the challenges.” Meanwhile, former Meta engineers cite reduced internal skepticism compared to the metaverse era, when leaders like John Carmack openly discussed technical constraints.
The Credibility Calculus
Meta’s pivot carries significant reputational stakes. Having invested roughly $60 billion in metaverse development with minimal return, Zuckerberg’s leadership faces intensified scrutiny. The AI transition strategically leverages existing strengths, data networks, user bases, and advertising infrastructure, unlike the metaverse’s hardware-heavy reinvention.
Yet the specter of “meta-failure” lingers. As one industry observer noted, today’s promise that we’re at “the beginning of a new era for humanity” risks aging as poorly as Meta’s former vision of a metaverse where “you’re gonna be able to do almost anything you can imagine”. With shareholder patience wearing thin and competitors advancing rapidly, Meta’s superintelligence labs must deliver concrete results where the metaverse could not. The company’s future as an innovator, not just an acquirer, may hinge on this high-stakes redemption arc.
Subscribe to my whatsapp channel
Comments are closed.