One argument that AI companies have made
\n\nis that AI is such a transformative technology
\n\nthat if we get to AI, it will then allow us
\n\nIt will even, before we've reached, you know,
\n\ngeneral intelligence or superintelligence,
\n\nthat make up for the excess power use.
\n\nHow is that math borne out in your reporting?
\n\nI don't know if that's so much math
\n\nlike some folks have researched this,
\n\na lot of AI companies that are touting changes,
\n\nyou know, like, energy savings from AI,
\n\nor like climate improvements from AI,
\n\na lot of the projects that have borne fruit
\n\nwere just kind of like everyday machine learning.
\n\nThey weren't building agents to solve the climate crisis.
\n\nthat eventually solves the climate crisis,
\n\nbut on our way there, we are building out
\n\nlike some of the biggest gas plants in the world.
\n\nAnd again, like anything is possible.
\n\nI'm gonna be the best-faith advocate here.
\n\nIn five years, someone could make, you know,
\n\nthere are a lot of nuclear companies, for instance,
\n\nstarting up making small modular reactors.
\n\nThe industry could do a complete about-face and say,
\n\nOr they could be like, We got more efficient.
\n\nYou know, we don't…
Original source: https://www.wired.com/