On top of that, they are effectively paying per word to post through paying for tokens. This does not stop them.
Assuming the Fediverse was as popular as Twitter/FB, they could simply run their own instances, no need to pay for access through any specific server.
And then what? No one is going to navigate an instance that is just spam, and other instances will promptly block them.
There was a spam wave some months ago. The spammers were using Mastodon and Sharkey instances that had open registrations. They were not signing up to paid-only instances to become "legit".
Anyway, my point is that charging for access to a network where access is already open is good enough for a filter to avoid spam originating from your node. Spammers are not going to be interested in paying $29/year to be able to send posts via Mastodon when they can just create a bunch of accounts on servers with open registrations or simply running their own botnet.
Hell, I don't even have to be a complete cynic to make the argument that basically any news piece today only gets to be written if it serves the economic interests of its publisher. That is valid from the NYT and Washington Post to an indie game developer talking about their project on Mastodon.
If everything is "spam", then there is no ham. If there is no ham, there is no way to build an classifier. If there is no way to build a classifier, then what is the problem we are trying to solve in the first place?
Bots spreading misinformation are bad and a problem on Twitter/Facebook/TikTok, sure. But the reason that an entrance fee does not solve this problem is because these networks are built on the idea of controlling what content people get to see. On the "open social web", people are in charge, there is no "algorithm" and manipulation becomes a lot harder.
No? That doesn't mean they aren't spam, though.