The field of making sure AI does what humans actually want, not what we literally said. Either the most important problem ever or a great salary, depending who you ask.
Alignment team says it's safe. Alignment team also hasn't slept since 2022.
+1219