It is a bit ironic that a paywalled article like this will have a top level comment with the archive link, which can then be easily scraped by AI (along with the comments)
I think a big difference is that there’s no micro transactions or compulsory licensing for content, so it always feels patently unfair to buy a subscription to read one article.
You can see artifacts when their servers are at queue load and you see the URLs, a few resources have the JWT with the account details in the URL. IIRC the clearname of the account in the token is Masha Rabinovich, with an email account masha@dns.li, an identity that has cropped up in various investigations [1][2].
[1] https://gyrovague.com/2023/08/05/archive-today-on-the-trail-...
[2] https://webapps.stackexchange.com/questions/145817/who-owns-...