Comment by godelski | Hacker News Reader

Comment by godelski | Hacker News Reader

Parent Full thread

> Why is anyone surprised that the major/globally operating social media companies have extensive internal systems for controlling the content that appears via their platforms?

No one is. These Twitter Files are just building on a right wing prior of them being shadow banned. But they have yet to demonstrate that right wingers are specifically targeted. They don't show why these accounts got flagged (the crux of the issue and none of this means anything without demonstrating this knowledge) but just show prominent right wingers that did get flagged. We just get left with more questions. Do prominent left wingers also get flagged? Do prominent left wingers act in ways that would similarly be flagged? Is flagging biased? (I'm sure it is, but lets' measure it)

It is a common statistical manipulation through leading someone to causal conclusions via selection bias (sometimes called "malinformation"). While not directly said these tweets (and the previous ones) are targeted to make the reader believe that Twitter silences right wing voices. This may very well be true, but they have yet to provide strong evidence of this. Just samples. Problem is, we don't know anything about the distribution of samples so we have no means of determining if they are cherry-picked or not.

This is exactly right.

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&...

> If this[0] is accurate, then 99.7% of political contributions made by Twitter employees went to Democrats in 2022.

You're jumping a step here because you're making a sampling error. Like I said before, we know nothing about the distribution.

Let's put this in an extreme example. Suppose there are 100 people. 99 of them have the favorite color green. 1 of them has the favorite color red. Does that mean the logo will be red? What assumptions are implicit here to say yes? (this is where the logic fails) If a singular person controls the color of the logo and that person is not chosen at random then it really doesn't matter how many people have a favorite color green. The logo will be the favorite color of the person that makes the choice. Everyone else is noise.

Coming back to Twitter it would be naive to assume that the choices are being made democratically or uniformly at random. So what does the median employee's opinion matter? We know that higher incomes tend to lean more conservative but this is hard to say here because those studies generally aggregate out all salaries >$100k/yr, which might as well be a random sample. If the trend continues (reasonable assumption, but naive to take as fact) then it is reasonable to assume that managers (who we are assuming make more money than the median employee) are more conservative. But again, we just don't know.

The data you are linking to is actually irrelevant to the conversation at hand (it also has a huge yearly variance, making it even more noisy).

Yes, it is unreasonable. This is not how companies work. The personal preferences of employees are not what govern the operations of their employer.

Sure would be easier if the world was that black and white but unfortunately there are a lot of gray areas employees encounter which are left up to their subjective judgement.

Tesla employees gave 93% to Democrats; do you think the cars work better if the owner is a Democrat?

Employees of Trump Entertainment Resorts gave 100% to Democrats!

It isn't an unreasonable leap for people in our position with very limited information, but it is still just a guess. It's a shame that both of these reporters, with eyes inside the machine, seem to want to HINT that there is systemic bias without providing any supporting evidence.

> It isn't an unreasonable leap for people in our position with very limited information

No, this is exactly WHY it is unreasonable. This is how conspiracy theories work. Humans want causal understandings but they have limited information and their imaginations fill in the gaps. If you have little information you should also have low confidence. Look around here at how many people are certain of conspiracy. Look in the Twitter thread too. Much of what people are saying is highly unreasonable and this is not healthy for us as a community.

That tweet is as good as worthless imo. For all we know the entire $167k was from a single executive who had no moderation power.

[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.

Source? Most mass shootings were related to gang violence last I checked.

The vast majority of mass killings are not politically motivated and are in fact gang related. The vast majority of politically related mass killings are committed by people with severe mental illness and the political leanings of the perpetrators are irrelevant. Of the politically related mass killings done by people of sound mind, the distribution is fairly even politically speaking.

Reply on news.ycombinator.com