Email providers very explicitly do care. You can be banned from gmail for especially egregious things for example. If an email account was serving me nazi propaganda I would inform the platform, and if it was coming from a specific email provider they would likely be blacklisted. Keep in mind this already happens when it comes to spam; the difference is that nazis are not blasting random emails with nazi stuff so it becomes much harder to track.
> Do you care if your web host happens to host their websites?
Generally speaking, yes. We went over this before with Cloudflare and sites that were full of people harassing and harming other users. Responsibility eventually falls on someone and that someone was the host provider.
> Why should a "federated Reddit replacement" have to be any different? If you subscribe to subs X and Y and Z, and none of that stuff is in those subs, does it matter to you if the instance (or other instances it federates with) contains other subs–which you never visit–that do have it?
Because you have a fundamental misunderstanding of how federated instances and content works, which is why I recommend you actually interact with those services before trying to make arguments. A user isn't subscribing to sub XYZ, they're a user of an instance which has subs XYZ and which federates to instance ABC. This is important because the instance is serving you that info, so if I like XY but hate AB, the instance can serve me content, users etc which are a part of AB. In fact, this is why Reddit does exactly what you seem to think they don't do. Many subreddits have bots, automod etc that will ban people that are from certain subreddits due to the content of their posts or how likely they are to troll. Defederation is a more explicit form of that.
