6 ms·

All the accounts seen on the popular "ignore previous instructions" screenshots have paid for twitter. There's no reason to believe that these bots don't operat

All the accounts seen on the popular "ignore previous instructions" screenshots have paid for twitter. There's no reason to believe that these bots don't operate at scale and don't all spend $8 a month. On top of that, they are effectively paying per word to post through paying for tokens. This does not stop them.

rglullis 2y ago
Bots being paid by a state that is a misinformation campaign is not the same as spam. 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.
- sumeno 2y ago
  Bots spamming misinformation isn't spam, gotcha. Really moving the goalposts there
  - rglullis 2y ago
    I meant spam as shorthand for "bulk send of messages of similar content". Cheap to produce and to send. If bots paid by some government count as spam, then we should also say the same about sponsored content or PR releases. 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.
    - jen20 2y ago
      I classify both sponsored content and PR-speak as spam, personally.
      - rglullis 2y ago
        Should we start complaining about HN's "spam problem" because of all the "Launch HN" posts? And what about YC companies that have job ads pushed to the top of the frontpage? 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.
        jen20 2y ago
        > Should we start complaining about HN's "spam problem" because of all the "Launch HN" posts? No? That doesn't mean they aren't spam, though.
- latexr 2y ago
  > they could simply run their own instances And then what? No one is going to navigate an instance that is just spam, and other instances will promptly block them.
  - rglullis 2y ago
    Right, the incentive will be to create accounts on the popular instances, not on the paid ones. 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".

rglullis 2y ago

Bots being paid by a state that is a misinformation campaign is not the same as spam. 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.

sumeno 2y ago
Bots spamming misinformation isn't spam, gotcha. Really moving the goalposts there
- rglullis 2y ago
  I meant spam as shorthand for "bulk send of messages of similar content". Cheap to produce and to send. If bots paid by some government count as spam, then we should also say the same about sponsored content or PR releases. 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.
  - jen20 2y ago
    I classify both sponsored content and PR-speak as spam, personally.
    - rglullis 2y ago
      Should we start complaining about HN's "spam problem" because of all the "Launch HN" posts? And what about YC companies that have job ads pushed to the top of the frontpage? 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.
      - jen20 2y ago
        > Should we start complaining about HN's "spam problem" because of all the "Launch HN" posts? No? That doesn't mean they aren't spam, though.
latexr 2y ago
> they could simply run their own instances And then what? No one is going to navigate an instance that is just spam, and other instances will promptly block them.
- rglullis 2y ago
  Right, the incentive will be to create accounts on the popular instances, not on the paid ones. 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".