[glenstein]

So on the one hand, we have a massive systematic problem across virtually all online platforms. This problem involves (at least) thousands of new examples every day and so far has proved nearly impossible to control. And next to that we have a different problem based largely on idiosyncratic examples, and extrapolations from those idiosyncratic examples to hypothetical worst case scenarios.

I find this way of engaging with the problem to be profoundly misguided in two ways. One, it's a failure to correctly evaluate the relative scale of the two problems, and to consistently think and speak clearly about them in terms that reflect their relative scale.

And secondly, it mistakenly sets up the two problems as being in a relationship of interference with one another, such that talking about one is used to mean we should stop talking about the other. Instead of saying "this statement that racism is bad and sexual harassment is bad is a statement I do not support" it would be more helpful say "yes, that is a problem, I agree, we need to solve it. And meanwhile here's this other thing, but don't let this other thing detract from the importance of the first problem or imply we don't need to actively work on the first problem."