I think we've pretty conclusively proven that there's no non-dangerous option available. Maybe this isn't the best answer, but speaking as someone who's very consistently opposed these kinds of content policies in the past, I no longer think it's tenable to take a hard line against information control here.