6 ms·

> “…according to expert quantification and analysis in January 2022, over half of Twitter’s 8,000-person staff was authorized to access the live production envi

> “…according to expert quantification and analysis in January 2022, over half of Twitter’s 8,000-person staff was authorized to access the live production environment and sensitive user data. Twitter lacked the ability to know who accessed systems or data or what they did with it in much of their environment.” Half of any company having access to prod is insane, but half of a giant like Twitter is unimaginable to me. It really explains how compromising single employees seems so effective there.

rmatt2000 4y ago
In the political Twitter neighborhoods that I inhabit, there have been many cases over the years of people having their accounts yanked, with the only message being "You have been banned for violating the Twitter rules. This decision is final and cannot be appealed." When these bannings get publicized, however, the bad publicity often leads to the "permanent ban" being reversed, with Twitter simply explaining that the original ban "was a mistake". How does this sort of thing happen so often, unless random employees have access to powerful tools that they are not authorized or trained to use?
- SilverBirch 4y ago
  Here's the obvious counter - social media companies are by nature trying to chase super-linear cost:return ratios. So they can't actualy accept that N users mean N/x moderators. So instead, you hire Y moderators, and any time anyone points out you've shit the bed you wave your hands in the air and whisper "AI" when really what's happening is Y < N/x so you have a shit moderation team.
  - rmatt2000 4y ago
    If there aren't enough moderators, wouldn't the likely result be fewer rule-breaking accounts banned rather than more non-rule-breaking accounts banned? Maybe, instead of hiring N skilled moderators for $D, they could be hiring N * \delta unskilled moderators for the same price.

rmatt2000 4y ago

In the political Twitter neighborhoods that I inhabit, there have been many cases over the years of people having their accounts yanked, with the only message being "You have been banned for violating the Twitter rules. This decision is final and cannot be appealed." When these bannings get publicized, however, the bad publicity often leads to the "permanent ban" being reversed, with Twitter simply explaining that the original ban "was a mistake". How does this sort of thing happen so often, unless random employees have access to powerful tools that they are not authorized or trained to use?

SilverBirch 4y ago
Here's the obvious counter - social media companies are by nature trying to chase super-linear cost:return ratios. So they can't actualy accept that N users mean N/x moderators. So instead, you hire Y moderators, and any time anyone points out you've shit the bed you wave your hands in the air and whisper "AI" when really what's happening is Y < N/x so you have a shit moderation team.
- rmatt2000 4y ago
  If there aren't enough moderators, wouldn't the likely result be fewer rule-breaking accounts banned rather than more non-rule-breaking accounts banned? Maybe, instead of hiring N skilled moderators for $D, they could be hiring N * \delta unskilled moderators for the same price.