1/ Working on a thing about this but consider two numbers:

-# of people who post/consume public radical content online

-# of people genuinely interested in harming strangers b/c of radical ideology

People believe these are directly correlated but I'm not sure that they are.

2/ Let me use an example that's (((close to home))): online anti-Semitism. If I were crowned Emperor of the Internet (any day now) I would view the goal of reducing the amount of public online anti-Semitism as *separate* from the goal of reducing internet-inspired anti-Semitic

3/ real-world attacks. That's because there might actually be an argument that if you aggressively deplatform everyone who posts anti-Semitic garbage, sure, you get rid of the 90% of them who are just in it for the shitposting and who harass journalists etc., but the 10% who hold

4/ deeply-held views, are deeply disturbed, or both, are going to flee underground to environments that are much more dangerous pressure-cookers. That's an example of how these should be seen as *potentially* different goals. But this is somewhat speculative because no one knows!

5/ I just don't think there is strong evidence that we can correlate mere exposure to public online radicalism to terrorism in some zoomed-out sense. I know that seems counterintuitive but think about the last 20 years. Exponential increase in exposure to radicalism, pretty flat

6/ or declining violence, with blips. I wish people would think this through more carefully rather than assuming that if the racists posts go away, the threat of racist violence goes away with them. It is good to fight both but those efforts might sometimes be at cross purposes.

7/ Oh and the one part I left out is how public sites of radicalization can feed people into private ones, which is of course worth considering and which does run counter to my thesis in certain ways.