2 ms·
As long as humans are still interacting with the real world, it might actually work out - by depending on both real-world experience and machine-generated knowl
by TeMPOraL 3y ago
As long as humans are still interacting with the real world, it might actually work out - by depending on both real-world experience and machine-generated knowledge, humans would become a conduit through which ML models could indirectly experience that real world. The feedback loop would make the models smarter, not dumber.