Meta has developed plans to create a brand new synthetic intelligence analysis lab devoted to pursuing “superintelligence,” in keeping with reporting from The New York Occasions. The social media big selected 28-year-old Alexandr Wang, founder and CEO of Scale AI, to hitch the brand new lab as a part of a broader reorganization of Meta’s AI efforts below CEO Mark Zuckerberg.
Superintelligence refers to a hypothetical AI system that will exceed human cognitive talents—a step past synthetic common intelligence (AGI), which goals to match an clever human’s functionality for studying new duties with out intensive specialised coaching.
Nonetheless, very similar to AGI, superintelligence stays a nebulous time period within the discipline. Since scientists nonetheless poorly perceive the mechanics of human intelligence, and since human intelligence resists easy quantification with no single definition, figuring out superintelligence when it arrives will current important challenges.
Computer systems already far surpass people in sure types of info processing similar to calculations, however this slender superiority does not qualify as superintelligence below most definitions. The pursuit assumes we’ll acknowledge it after we see it, regardless of the conceptual fuzziness.
AI researcher Dr. Margaret Mitchell advised Ars Technica in April 2024 that there’ll “seemingly by no means be settlement on comparisons between human and machine intelligence” however predicted that “males in positions of energy and affect, notably ones with investments in AI, will declare that AI is smarter than people” whatever the actuality.
The brand new lab represents Meta’s effort to stay aggressive within the more and more crowded AI race, the place tech giants proceed pouring billions into analysis and expertise acquisition. Meta has reportedly supplied compensation packages value seven to 9 figures to dozens of researchers from firms like OpenAI and Google, in keeping with The New York Occasions, with some already agreeing to hitch the corporate.
Meta joins a rising listing of tech giants making daring claims about superior AI growth. In January, OpenAI CEO Sam Altman wrote in a weblog put up that “we are actually assured we all know how one can construct AGI as now we have historically understood it.” Earlier, in September 2024, Altman predicted that the AI trade would possibly develop superintelligence “in a number of thousand days.” Elon Musk made an much more aggressive prediction in April 2024, saying that AI could be “smarter than the neatest human” by “subsequent yr, inside two years.”
