4 ms·

daringfireball has 30 tweets in the last 10 days. It's also a 15 year old account with 100k followers. It seems unlikely that any remotely competent spam detect

daringfireball has 30 tweets in the last 10 days. It's also a 15 year old account with 100k followers. It seems unlikely that any remotely competent spam detection algorithm or human review process would suspect it of being spam.

lapcat 4y ago
This is missing the point, which is "The account did not violate Twitter rules in any way." If Gruber's accounts are not spam, then Thompson's account is not spam either. The age of the account and other statistics are irrelevant, because the same rules ought to apply to everyone. I think it ought to be uncontroversial that an account shouldn't be suspended if it hasn't violated the rules, but somehow many people in this thread are defending the punishment of the innocent. There's actually no rule against bots on Twitter. In fact there are some accounts now explicitly labeled "Automated". The goal ought to be not "detecting bots" but rather "detecting rule violations".
- tshaddox 4y ago
  I'm not all that concerned with what Twitter's rules supposedly are. If the only complaint here is that Twitter should be more clear about their rules and/or more consistent in enforcing them, then sure, I can agree with that (although there are obvious reasons why you shouldn't be completely clear about what your exact thresholds are for spam detection). I'm just talking about what I think looks clearly like spam, and what I think reasonable spam policies would look like. That whole Twitter thread of accusations has now been deleted, so I'm not necessarily addressing the exact text of it. I'm just saying that one of those posting patterns looks very clearly like spam and the other very clearly does not appear to be spam.
  - lapcat 4y ago
    > I'm just saying that one of those posting patterns looks very clearly like spam and the other very clearly does not appear to be spam. I disagree. Tweeting links without interacting in any way with other accounts is to me clearly not spam.
    - tshaddox 4y ago
      It’s all about age of the account, number of followers, and rate and origin of the automated tweets. Perhaps a way to think about it is to ask, if there was a new account created just to have a bot spam out links at a high rate, how would it look different than this account?
      - lapcat 4y ago
        > rate and origin of the automated tweets There were no automated tweets. https://twitter.com/benthompson/status/1583163548708241410 https://twitter.com/benthompson/status/1583163548708241410 "Account was not scripted. Everything was posted by a human" > if there was a new account created just to have a bot spam out links at a high rate, how would it look different than this account? I think we differ on the definition of "spam". Tweeting links on your own profile is not spam. Tweeting links in replies to other users can be spam. Ben's account was not replying at all, just tweeting in its own profile. A Twitter user would generally only see those tweets if they followed Ben's account. The was no unsolicited pushing.

lapcat 4y ago

This is missing the point, which is "The account did not violate Twitter rules in any way." If Gruber's accounts are not spam, then Thompson's account is not spam either. The age of the account and other statistics are irrelevant, because the same rules ought to apply to everyone. I think it ought to be uncontroversial that an account shouldn't be suspended if it hasn't violated the rules, but somehow many people in this thread are defending the punishment of the innocent. There's actually no rule against bots on Twitter. In fact there are some accounts now explicitly labeled "Automated". The goal ought to be not "detecting bots" but rather "detecting rule violations".

tshaddox 4y ago
I'm not all that concerned with what Twitter's rules supposedly are. If the only complaint here is that Twitter should be more clear about their rules and/or more consistent in enforcing them, then sure, I can agree with that (although there are obvious reasons why you shouldn't be completely clear about what your exact thresholds are for spam detection). I'm just talking about what I think looks clearly like spam, and what I think reasonable spam policies would look like. That whole Twitter thread of accusations has now been deleted, so I'm not necessarily addressing the exact text of it. I'm just saying that one of those posting patterns looks very clearly like spam and the other very clearly does not appear to be spam.
- lapcat 4y ago
  > I'm just saying that one of those posting patterns looks very clearly like spam and the other very clearly does not appear to be spam. I disagree. Tweeting links without interacting in any way with other accounts is to me clearly not spam.
  - tshaddox 4y ago
    It’s all about age of the account, number of followers, and rate and origin of the automated tweets. Perhaps a way to think about it is to ask, if there was a new account created just to have a bot spam out links at a high rate, how would it look different than this account?
    - lapcat 4y ago
      > rate and origin of the automated tweets There were no automated tweets. https://twitter.com/benthompson/status/1583163548708241410 https://twitter.com/benthompson/status/1583163548708241410 "Account was not scripted. Everything was posted by a human" > if there was a new account created just to have a bot spam out links at a high rate, how would it look different than this account? I think we differ on the definition of "spam". Tweeting links on your own profile is not spam. Tweeting links in replies to other users can be spam. Ben's account was not replying at all, just tweeting in its own profile. A Twitter user would generally only see those tweets if they followed Ben's account. The was no unsolicited pushing.