Nate Jones and I dig deep into Twitter’s decision to delete Rudy Giuliani’s tweet (quoting Charlie Kirk of Turning Point) to the effect that hydroxychloroquine had been shown to be 100% effective against the coronavirus and that Gov. Whitmer (D-MI) had threatened doctors prescribing it out of anti-Trump animus. Twitter claimed that it was deleting tweets that “go directly against guidance from authoritative sources” and separately implied that the tweet was an improper attack on Gov. Whitmer.  So where did Twitter find the “authoritative guidance” that Giuliani was supposed to be “going directly against”? Of course, Twitter isn’t explaining itself, which raises questions about the basis for its action. (I offered two of its representatives a chance to come on the podcast to offer a defense; they didn’t respond.) In short, all the people who’ve been telling us our freedoms are at risk as a result of the health emergency might be right, but the source of the danger isn’t government. It’s Silicon Valley. Nate thinks (probably correctly) that Kirk and Giuliani were wrong about the “100% effective” claim, and that people like them and the president are going to get people to take dangerous drugs without medical advice if they aren’t policed. It’s a spirited exchange. In contrast, Paul Rosenzweig and I find a fair amount of common ground outside this week’s media consensus that Zoom is either evil or stupid, maybe both, for its handling of privacy and security of users. No doubt there are a staggering number of privacy and security holes in the product, and the company will get sued for several of them. But we suspect that many of the problems would have been exposed and fixed over the course of the three years it would have taken Zoom to reach the levels of use it’s instead reached in three weeks. One error, exposing LinkedIn data to unrelated users with the same Internet domain, seems to have hit Dutch users especially hard.  The DOJ inspector general has found widespread gaps in the FBI’s compliance with its now-famous Woods procedures. Matthew Heiman and I try to put the damaging report in perspective. It’s hard to know at this point how serious the gaps are, though the numbers suggest that some will be serious. Meanwhile, the FISA court has ordered a rush evaluation from Justice of more or less exactly the same questions the IG is asking. We manage to agree that the court’s June 15 deadline is not realistic given everything else the same group of lawyers will be doing between now and November.  Matthew tells us that the Saudis are suspected of a phone spying campaign in the United States. I point out that foreign location collection is pretty much built into the SS7 phone system, so the worst that can be said about the event is that the Saudis were caught doing “too much” spying in the US. Paul comes down agreeing with a new court ruling that violating a site’s terms of service isn’t criminal hacking. And now that that’s settled, I have a research proposal for the Hewlett Foundation. Washington State has adopted a facial recognition law that Microsoft likes, Nate tells us. No surprise, I suggest, since the law will only regulate governments, not the private sector. I’m not a fan; it looks like a law that virtually guarantees that any facial recognition system will be forced to “correct” empirical results in favor of quotas for “protected subpopulations.” This leads, in light of Zoom’s problems, to the question whether that includes the Dutch. Who is hacking the WHO? Who isn’t? Matthew notes that Iran has joined what must be a crowd of eavesdroppers in WHO networks. Nostalgic for the days before the coronavirus? How about this blast from the past: Marriott has revealed a data breach exposing (some) personal data for up to 5.2 million c