> Basically they are choosing to ignore security vulnerabilities because they had to read too much slop last year.<p>That's not only an uncharitable take, it's also wrong.<p>If 999 out of every 1000 "reports" from a specific source is wrong, then it is not irrational to disregard all 1000, especially when they can be generated faster than you can read.<p>I mean, it's just probabilities, right? If you're okay trusting output from an LLM, you should be okay with using statistics in general as a source for informing decision-making.