If AI relies on human-curated data, what's stopping it from learning to influence its human gatekeepers to better achieve its goals? A theoretical risk that calls for proactive oversight and robust alignment research. #AISafety Alignment

A cover of a pat benatar banger

Seems legit

I often wonder how much sitcoms impact our day to day interactions. Conventional wisdom tells us a conflict/resolution makes a good story. However, I would imagine conflicts to be rare, except that my brain thinks reality is that conflict is normal bc I watch all these sitcoms.