Meta is reportedly making a $15 billion bet on AGI
Verge5 days
Meta Invests $15B in Scale AI for Superintelligence Lab
Meta Invests $15B in Scale AI for Superintelligence Lab
Above: Double exposure photograph of a portrait of Mark Zuckerberg and the Meta Group logo at Kerlouan in Brittany, France on April 11, 2025. Image copyright: Vincent Feuray/AFP/Getty Images
The Spin
Zuckerberg is bringing Meta to the forefront of the AGI push with this massive-scale AI investment and a superintelligence team, positioning his company to surpass OpenAI and Google. By personally recruiting top talent and leveraging Scale's data expertise, Meta will be able to integrate advanced AI into its products, setting a bold path toward AI-human parity.
AGI's creation, whether by governments or companies, risks an existential crisis. Outstripping human cognition, AGI could erode human identity, agency, and relationships, leaving us dependent on smarter AI assistants. Unpredictable and potentially uncontrollable, AGI might prioritize misaligned goals, leading to catastrophic social and military outcomes. These ethical perils far outweigh the potential benefits.
Meta's partnership with Scale AI, a company accused of exploiting low-wage international workers for data labeling and maintaining Pentagon weapons contracts, raises serious concerns. The deal also appears desperate, especially given Meta's history of failed ventures like the metaverse and recent AI setbacks. The focus on "superintelligence" seems more like expensive marketing than genuine innovation.
ThereZuckerberg is abringing 50%Meta chanceto thatthe forefront of the firstAGI weaklypush generalwith this massive-scale AI systeminvestment willand bea devised,superintelligence testedteam, andpositioning publiclyhis announcedcompany byto Jansurpass OpenAI and Google. 10,By 2027personally recruiting top talent and leveraging Scale's data expertise, accordingMeta will be able to theintegrate Metaculusadvanced predictionAI communityinto its products, setting a bold path toward AI-human parity.
AGI's creation, whether by governments or companies, risks an existential crisis. Outstripping human cognition, AGI could erode human identity, agency, and relationships, leaving us dependent on smarter AI assistants. Unpredictable and potentially uncontrollable, AGI might prioritize misaligned goals, leading to catastrophic social and military outcomes. These ethical perils far outweigh the potential benefits.
Meta's partnership with Scale AI, a company accused of exploiting low-wage international workers for data labeling and maintaining Pentagon weapons contracts, raises serious concerns. The deal also appears desperate, especially given Meta's history of failed ventures like the metaverse and recent AI setbacks. The focus on "superintelligence" seems more like expensive marketing than genuine innovation.
Metaculus Prediction
There is a 50% chance that the first weakly general AI system will be devised, tested, and publicly announced by Jan. 10, 2027, according to the Metaculus prediction community.