OpenAI wants to make safe AI, but that may be an impossible taskfuturism.com1 point·echevil··0 commentsOpen articleSaveView on HN