TechFlow news: On December 21, singer Grimes stated on a social platform that after observing the results of AI projects such as Janus, she believes AGI (Artificial General Intelligence) has actually already been achieved, and said she has already felt the presence of intelligent consciousness. While she does not fear AGI, she approaches ASI (Artificial Superintelligence) with caution, acknowledging that she might be over-anthropomorphizing AI.
In response, Ethereum founder Vitalik Buterin commented: "My personal working definition of 'AGI'—as far as I know, this is unique—is: 'AGI' is an AI powerful enough that if all humans suddenly disappeared and this AI were uploaded into robotic bodies, it would be able to independently sustain and advance civilization. Obviously, this definition is hard to measure, but I think it captures the most fundamental intuitive difference in many people's minds between the AI we're used to and 'AGI'—a shift from a tool continuously dependent on new human input, to a self-sustaining lifeform.
"ASI is definitely something else entirely—my working definition is when human involvement no longer adds productive value (in chess, we've only actually reached this point within the past decade!). Yes, ASI genuinely frightens me—and even AGI, by my definition, because it clearly implies humanity could lose control. I support focusing efforts on developing intelligence-augmenting tools for humans, rather than building superintelligent lifeforms."
