A 21st century Manhattan Project – is it our last line of defense against an AI apocalypse? What experts are saying about the future of AI?
– The AI arms race is accelerating, with predictions of a state-sponsored superintelligence project launching in the US by 2026-2028, regardless of who's in the White House
– The scale of investment required is staggering – potentially reaching $1 trillion per year by 2027, dwarfing historical precedents like the Manhattan and Apollo programs
– Automated AI research could lead to an "intelligence explosion," compressing years of progress into months and potentially achieving full superintelligence by 2030
– The challenge of superalignment – ensuring superintelligent AI/AGI aligns with human values, remains a critical and unsolved problem
– New players like Safe Superintelligence Inc. are emerging, focusing solely on creating secure superintelligence
– Some organizations, like MIRI, are calling for drastic measures such as a global "power switch" to halt AI development if critical risks emerge
Top comments (0)