> Longer answer: Basic data analysis can pick up most if not all the shenanigans. Doesn't matter how much they "age" the accounts or hide their tracks, eg IP address etc. Don't even necessarily need high powered hardware or even machine learning. They do help in particular ways if you choose to use them.
> Fake vote detection is now a choice not to rather than a technical blockage.
I don't believe this is correct. Bots can create simple sentences that very much appear as a human. Spread these out across enough accounts, and filtering noise becomes very difficult.
There are also apps that activists use and I'm certain PR firms as well which alert humans to go and vote in certain ways.
These and other technologies combined with IPv6 make it exceptionally difficult for site admins and moderators to discover manipulators. Plus, many sites have few resources to put into the task anyway.
HN has at least put one step against it by not having downvotes on article submissions, though that doesn't help against upvoting manipulators.