Reaction
Can AI become wildly superhuman at every task?
One post argues that persuasion and election predictions face external limits; a reply says physics constrains computational tasks, too.
TLDR
A post argues that AI's superhuman performance in math and cybersecurity need not extend to persuasion or election forecasting, which it says face external limits. It also argues that developing cures requires real-world experiments, not just computation. A reply praises the distinction but adds that physics limits even the computational tasks presented as open-ended.
Combined views
10.7K
4 Sources, first seen ago
