Forget the ill-fated Metaverse, stablecoin venture Diem/Libra, and internet access project Aquila. Mark Zuckerberg is out with a new vision, and this time it's centering around AI. In a 6,500-word manifesto, Zuckerberg argues that artificial superintelligence should be widely distributed and based on individual empowerment rather than the control of elite labs or governments. A similar opinion piece on the matter was published two weeks ago in the WSJ, but the fresh essay goes into much more detail to get out his message.
Snapshot: To achieve his vision, Meta (META) plans to deliver personalized, superintelligent agents to billions of people free of charge or through dynamic compute auctions. "Your agent will work 24/7 on your behalf to improve your relationships, health, career, finances, home management, hobbies, and more," as well as providing "incredible tools to express your ideas... create new businesses... and contribute to scientific progress." Invention, not automation, will be the greatest contribution of superintelligence, as long as capabilities are not restricted—and checks and balances are guaranteed—by the distribution of powerful AI tools to everyone. Meta unveils open-source AI model Muse Glimmer
Zuckerberg also dismissed fears that AI would eliminate most jobs and humanity's relevance, calling it "doom" discourse and a narrative that threatens to concentrate power in the hands of a few. "Novel ideas and major steps forward rarely originate from established institutions alone. Putting power in people's hands to pursue their own aspirations is how humanity has made the most progress. As everyone gains more powerful tools, each person will become more capable of shaping the future, not less."
Critics say Zuckerberg is merely looking for a new in, with Meta (META) far behind in the AI wars. Looking to play catch-up, he's advocating for open-source models and "distillation," where less capable models are trained on the outputs of stronger ones. He may also be promising personal agents in return for data collection and the ability to target users with advertising, while increased adoption would benefit the company's infrastructure and data center investments. Elsewhere, Meta has been hammered recently over social media addiction and harm to children, so it might make sense to shine a spotlight on new priorities in a well-timed AI manifesto.
Thought bubble: Whichever side is correct, Zuckerberg is clearly pivoting back to his early roots of "move fast and break things." He's advocating for a system of disrupting incumbents, iterating in the wild, and letting the market sort out the consequences. This contrasts heavily with his stance in recent years that advocated for strong centralized infrastructure. Those efforts saw the widespread filtering of news and content, along with his calls for government regulation of internet platforms, triggered by events like the U.S. elections and the Cambridge Analytica scandal.
