[candybar]

This is a huge tell,. Matt Taibbi (or Elon Musk via him) is trying to paint a picture that this was heavily tilted in favor of Democrats. But he has access to the actual data - why not just publish that if he feels it's damning? Whatever mechanism he used to find these emails and also to conclude that both campaigns had access, could also have been used to determine the bias statistically. And the raw data could even have been shared for public consumption.

Instead, he talks about campaign contributions, which isn't even circumstantial evidence of bias. You don't have to contribute to a political campaign to process moderation requests from them. There's no plausible mechanism by which the volume of requests that can be processed is related the amount of money donated by Twitter employees. You just need one contact and again, no evidence exists - and not even clear accusations have been made - that Twitter employees involved were reluctant to process requests made by the Trump team.

If anything the fact that Matt Taibbi is dishonestly trying to make this connection is extremely strong circumstantial evidence that they couldn't find anything even remotely damning. Because if they could find something actually damning, they would've used that instead. Resorting to, ugh, Twitter employees are liberal, so I'm sure they weren't entirely being fair in processing requests from the Trump team and ugh, more contributions to the Democrats so like, must be able to process more requests, which is a complete non-sequitur, is highly damaging to whatever narrative that he's trying to push.

[etchalon]

It sounds like what he has is a dump of emails, not necessarily a database of moderation actions.

Even if he had a database like that, it's unlikely it's easily classified in a way you could actually correlate individual actions to specific sources. I highly doubt there's a database column for "as commanded by the dark Democratic conspiracy."

[candybar]

Sure but you can turn a dump of emails into statistics easily. Or alternatively, you can release the dump of emails.

> This system wasn't balanced. It was based on contacts. Because Twitter was and is overwhelmingly staffed by people of one political orientation, there were more channels, more ways to complain, open to the left (well, Democrats) than the right.

Yet the fact that there's absolutely no attempt to even analyze the data for signs of bias, but have to resort to a complete non-sequitur, is quite telling. It likely means that the raw data doesn't support their narrative. You can also share a sample of tweets that were requested to reviewed by both parties and which ones got removed and which ones did not. If the Trump administration's requests were not honored despite the tweets clearly being in violation or the Biden campaign's requests being honored despite the tweets clearly not being in violation, well that would be something if there was a clear pattern, or internal discussions that explicitly suggest that they are making decisions like that specifically to favor one candidate. It still has nothing to do with the 1st amendment, but it would at least corroborate the claim that Twitter was biased.

Again, the standards need to be high because they are the party with all the information. This type of insinuation can be persuasive if you're the outside party that suspects bias and the other party controls the information. But claiming that there might be some bias, look at all this external information as to why it might have happened, tells the exact opposite story, when they are the ones that control the information. Elon Musk and Matt Taibbi literally have all this information at their disposal and clearly want to spread this narrative that Twitter was biased in the Democrats' favor. But they can't prove it - it seems ridiculous for them not to have looked into this, but they thought it was best not to release that information. Why not? Almost certainly because it detracts from the narrative.

[mcrad]

Yes, you can turn a dump of emails into biased statistics, easily.