The biggest risk of large language models is that they may bury us in lies | Dark Hacker News