There's a blog post with more details about the project and how it works here: censys.com/automated-hu....

Chrome has released some distribution of global traffic compared to site popularity data (zakird.com/papers/brows... Figure 1) if someone wants to do rough back of envelope calculations

It's been a few years but that's not far off from what we saw in Firefox data when we analyzed: zakird.com/papers/lets-... (Figure 5). A lot of the long tail, though may be some more widespread adoption

While there's been much work building improved models to more effectively detect threats and harassment, what users want out of these tools is more complex than what we're optimizing or evaluating for today.