3 ms·

If the openai board can't get humans to align, what hope do they have of "aligning" ml models?

• catchnear4321 3y ago
  people in alignment with each other is agreement. a language model in alignment is control. the model does not need to be aligned with the desires of people. just person. it could be people, but getting alignment with people is…