This morning I had a brief episode of what I have taken to calling "Being out of phase with yourself." It felt as if I was parroting my own actions as I was performing them, as if I was suddenly and briefly cognizant of the delay between my unconscious movements...
Current AI safety approaches have a structural problem: they ask the same system to both generate outputs and judge whether those outputs are safe. RLHF trains a model to produce outputs that a reward model likes. Constitutional AI has the model critique its own outputs against principles. In both cases,...
My eyes are burning from the strain of how aggressively I am not-crying right now. For as long as I can remember I have been the odd one out. I never cared about what everyone cared about, or at least not in the "normal way" that people were expected to...