And this is perhaps the most profound irony of the executive order: It criticizes the sweeping immunity provided to the tech industry by Section 230 of the Communications Decency Act, the controversial 1996 federal law that prohibits online intermediaries from being treated as the publishers or speakers of content posted by internet users. But the order doesn’t address the core problem with the law that scholars and advocates have highlighted for years—namely, how its immunity provision not only fails to encourage online intermediaries to address harmful content but rewards them for indifference. Trump’s order does not acknowledge the ways that this immunity has allowed online intermediaries to ignore, encourage, and profit from abuses—harassment, privacy invasion, deadly misinformation—directed at vulnerable groups, especially women and people of color. It does not recognize, in other words, the similarities between Twitter and Trump.

In the standoff between Twitter and Trump, siding with Twitter is easy. Yesterday, the company intervened again. To the president’s tweet about the protests in Minneapolis that said “When the looting starts, the shooting starts,” Twitter added a warning label indicating that the apparent threat of extrajudicial killings violated the company’s policy against glorifying violence. The official White House account reposted the same language that Trump used, and Twitter put a warning label on that tweet, too. In its willingness to flag Trump’s posts, Twitter has taken a much braver and more principled stand than, say, Facebook has in countering misinformation and abuse.