Live data from Hacker News

The Twitter Files Part 2: Twitter's Secret Blacklists

twitter.com

221–230 of 580 posts

Re: The Twitter Files Part 2: Twitter's Secret Blacklists

#221

I have been responsible a bit of content moderation myself, and interacted with mods on various forums. It's possible that this is responsible for some of my views, summarized here. - The "secret" aspect of all this can be largely ignored. Twitter is under no obligation to tell its users exactly how their Tweets are moderated. I am sure most moderation tools are secret, including whatever Twitter (or for that matter,…

Yes, the secrecy does matter. I think can immediately think of two big reasons:

1) We should censor for how things are said, instead of what is said. When censorship is carried out secretly it will inevitably trend towards abuse. One of the examples given in these leaks is that @DrJBhattacharya [1] was secretly put on a trends blacklist because of what he was saying. He's a well spoken doctor and Stanford epidemiologist, but he publicly spoke against the political measures being taken against COVID, such as lockdowns. So he was blacklisted. That's just so incredibly wrong, and enabled only by secrecy.

2) Again with the goal of adjusting how things are said, instead of what is said: If you think somebody is an idiot, saying that accomplishes absolutely nothing besides sending discourse to the level of a dumpster dive at a cheap seafood joint on a hot summer day. By contrast, engaging in good faith and simply discussing the topic not only keeps the conversation healthy, but might even actually change some minds. Secretly censoring the shitposter does not really encourage him to engage in a more productive fashion, because he may not even realize he's behaving a way that's not really considered acceptable, let alone attempt to correct it.

---

Literal spam is a different topic, and I think conflating the two is disingenuous. Doing things like blacklisting doctors because of what they say has little to do with banning somebody trying to impersonate other accounts and spamming 'give me 1 and I'll give you 2' crypto scams.

[1] - https://twitter.com/DrJBhattacharya

[2] - https://www.nytimes.com/2022/01/04/briefing/american-childre...

Re: The Twitter Files Part 2: Twitter's Secret Blacklists

#223

Earlier quoted context omitted.

It’s not the fact that these systems exist, it’s that Twitter denied that they did for years, despite all the evidence to the contrary.

Indeed and here is their tweet saying they do not https://twitter.com/twitter/status/1022658436704731136?s=46&... Isn't there a fiduciary duty to be honest with their investors?

It's generally assumed investors can read. https://blog.twitter.com/en_us/topics/company/2018/Setting-t...

Re: The Twitter Files Part 2: Twitter's Secret Blacklists

#224

Earlier quoted context omitted.

> Colloquially, I think many information suppression tactics get lumped under the term "shadow-ban" and that your definition is very narrow. It reinforces your position so I understand why you're defining it as such but it's not very convincing to anyone who doesn't share your position. Of course, that's assuming you're genuinely interested in persuasion instead of yelling into the echo chamber. Reasonable criticism.…

It's not reasonable criticism, because Twitter provided their definition of "shadow-ban" at the time, and it's unambiguously not what Weiss is talking about. You can disagree with the definition Twitter chose, but you can't retcon your own definition onto Twitter's policy statements.

I'll be frank - I couldn't give a rat's ass about Weiss or Taibii. I'm not familiar with them and they're obviously trying to squeeze as much juice out of this story as they can.

However, I see a lot of people here dismissing the wider and (imo) more valid complaints about veiled moderation tactics by beating up these two and calling it a day, case closed. I don't care what specific narrow definition twitter chose on a whim in some blog post and I'm not concerned with strictly "shadow-banning". The fact is, these global influence platforms are easy to subvert by immensely powerful interests (state actors, billionaires, etc) and the power we're placing in these people's hands is incredible, way beyond what any authoritarian dictator of yesteryear could even dream up. And when how the sausage is made is ever so slightly exposed, it's dismissed and hand-waved away.

What could go wrong?

Re: The Twitter Files Part 2: Twitter's Secret Blacklists

#225
post #211

Love to see non-employees of Twitter having access to my DMs. I don't think that's consistent with the user agreement or the privacy laws in the state of California.

I don't think they even have access to the dashboard. I assumed they were given the prepared data. An earlier tweet by Musk or a journalist mentioned lawyers would deliver the data to them direct from Twitter. Even if they did it's clear that whoever made the screenshot in the libsoftiktok screen doesn't even have access to see their email. Like all other tech admin panels, there are access controls. Just because the…

Weiss's co-author @abagailshrier commented: 'Our team was given extensive, unfiltered access to Twitter's internal communication and systems.' I'm gonna take this at face value pending clarification.

Just because the 'delete website' button is visible to you doesn't mean your account has permission to use it.

Well, nothing looks greyed out. Again, I'm going to assume they had access unless there's specific information to the contrary.

Re: The Twitter Files Part 2: Twitter's Secret Blacklists

#226
These “Twitter Files” have been so disappointing. I was really expecting to see some crazy stuff, but all I’m seeing is a company that was moderating their content. Even the incident that Weiss points out of someone being doxxed and Twitter doing nothing about it happened under Elon’s rule, and was probably not moderated because of the chaos that Elon created. Even the people that were “shadow banned” makes sense, I’m actually surprised that they didn’t moderate accounts like Libs of TikTok even more considering that they’re actively shutting down children’s hospitals and directing hate and violence towards transsexual people. All in all, a big nothingburger.

Re: The Twitter Files Part 2: Twitter's Secret Blacklists

#227
post #224

Earlier quoted context omitted.

It's not reasonable criticism, because Twitter provided their definition of "shadow-ban" at the time, and it's unambiguously not what Weiss is talking about. You can disagree with the definition Twitter chose, but you can't retcon your own definition onto Twitter's policy statements.

I'll be frank - I couldn't give a rat's ass about Weiss or Taibii. I'm not familiar with them and they're obviously trying to squeeze as much juice out of this story as they can. However, I see a lot of people here dismissing the wider and (imo) more valid complaints about veiled moderation tactics by beating up these two and calling it a day, case closed. I don't care what specific narrow definition twitter chose on…

The thread is about the false claims Weiss is making, not about your own opinions on moderation. What you should do, if you want to share those, is start a new top-level thread on this story. I'm sure what you have to say is germane to the story. It's just not germane to the thread you're commenting on.

Re: The Twitter Files Part 2: Twitter's Secret Blacklists

#228
post #206
post #24

Twitter calls it "visibility filtering" [1] Facebook gives mods a "Hide comment" button [2] TikTok calls it "visible to self" [3] [4] Truth Social does it [5] [6] [7] Reddit shows all removed comments to authors as if they're publicly visible [8] Open source tools are built to do it [9] [10] Textbooks advocate "disguise a gag or ban" [11] I call it Shadow Moderation [12]. The system intentionally does not show users…

> Reddit shows all removed comments to authors as if they're publicly visible [8] This is not true. There are cases where comments are silently removed without any public indicator; To the author, it appears as if the comment is posted normally. To other accounts however, there is no indication that the comment was ever posted.

There may be some wires crossed here. Reddit does show all removed comments to the people who wrote them as if they're not removed. I built Reveddit to reveal this. You can lookup a random user with this link [1]. There's a 50% chance they have a removed comment in their recent history, and chances are they weren't told that it was removed.

> To other accounts however, there is no indication that the comment was ever posted.

This is another important detail, so thank you for bringing it up. Removed leaf comments (comments with no children) do not show any marker at all. I mention it in my talk at 1:23 [2],

> Note that other users only see one marker here. They are not shown a marker for replies that have no children. So there is no indication to them that there was even a second comment here. The vast majority of removed comments fall into this category. Anything removed automatically for containing a phrase or link from a subreddit's ban list has no chance of receiving a reply.

[1] https://www.reveddit.com/random

[2] https://cantsayanything.win/2022-10-transparent-moderation/

Re: The Twitter Files Part 2: Twitter's Secret Blacklists

#229
post #28

Earlier quoted context omitted.

It's not trustworthy when a system actions a user's content without telling them. Free speech principles are also a thing. The question is not what is getting demoted, it's how . In this case, it's being done secretly, and Twitter isn't the only one doing it: https://news.ycombinator.com/item?id=33916414

This outrage is hilarious coz HN shadowbans people too. Which increases site quality which is why all the complainers are on here.

HN has an option to be able to see the removed content.

Re: The Twitter Files Part 2: Twitter's Secret Blacklists

#230
post #207

Earlier quoted context omitted.

If this[0] is accurate, then 99.7% of political contributions made by Twitter employees went to Democrats in 2022. Is it unreasonable to suspect people on the right are targeted more often than the people on the left? Wouldn't that be the default assumption and the opposite position would need evidence to prove otherwise? [0] https://twitter.com/mtaibbi/status/1598829996264390656?s=20&...

[flagged]

What is your source for this? As far as I understand the vast majority of mass killings in the US have no political motive.
Post reply on HN