rand

the problem i have with llms is that they mess with the distrust between humans and machines. before llms, if the human and the machine both agreed that something was correct, it usually was.