To make AI safe, we must develop it as fast as possible without safeguardsalignmentalignment.ai2 points·louisbarclay··2 commentsOpen articleSaveView on HN