Note on Left and Right in the Twitter Files…

Mainstream outlets try to grind every news story into grade D partisan hamburger and Twitter Files coverage is no exception. The Washington Post even called me a “conservative journalist” for a few minutes.

This isn’t a left or right project. The question that interests me, how these companies have been absorbed as intelligence arms, is more future/dystopia than blue/red. But that story is hard to sell, se we’re getting the usual stupidity.

Still, well-meaning people have questions:

Obviously, I know voices on the left have been “moderated” a lot, because I’ve been interviewing them for years: https://taibbi.substack.com/p/meet-the-censored-paul-jay… https://taibbi.substack.com/p/meet-the-censored-olivia-katbi-smith…

But we’re all trying to triage time and instead of looking up individual accounts, most of us have been looking at broader search terms to start, like “FBI,” “Covid,” “DHS,” etc.

I did look up account(s) connected to Julian Assange. His “PV2” page doesn’t show anything unusual, just automated suspensions of an abandoned account, and I found no record of, say, a government-initiated action.

That doesn’t mean such intervention doesn’t exist. It may be in a Slack or an email somewhere. It’s a big haystack. We’ll keep looking.

A few actions hit both Republicans and Democrats. Some FBI offices were clearly running searches of “November 4” to catch people trying to trick others into not voting, and this snared both Biden and Trump voters in silly numbers.

There are interesting/ambiguous details, like a decision we found in which the company considered restricting all moderation decisions involving Biden or Trump to four senior executives:

But Twitter did have a clear political monoculture. I ran searches for both “RNC” and “DNC,” cross-referenced against senior executives. “RNC” turned up pages about Republicans suing the company. “DNC” returned mountains of insistent moderation demands.

Some of the latter were quite funny and revealing.

In multiple instances Twitter initially decided not to remove videos lampooning Joe Biden because they were obvious parodies “unlikely to cause offline harm or generate confusion.”

These included a "deceptively edited" video of Biden coughing: https://www.reuters.com/article/uk-factcheck-biden-coughing-video/fact-check-video-of-biden-coughing-has-been-selectively-edited-to-imply-he-has-covid-19-idUSKBN26N37R… And a "Todos Con Biden" Trumo spoof: https://www.youtube.com/watch?v=n-CgjeGJUFQ…

Usually Twitter was responding to complaints from one very voluble DNC staffer and still applied warning labels to such content. In one case, they refused to do either.

“Because the video is an unaltered excerpt of the Vice President's speech, our teams consider it to be out of context, but not deceptive,” Twitter told the staffer, who fumed, "These rules need revision."

In the process, they sent a graph of their bizarre moderation flow chart, which among other things showed they can still apply labels to non-deceptive material. This seemed more interesting than the fate of a Biden coughing mashup.

If this kind of mechanized speech control can be used one way today, it can be used in another tomorrow, especially if unseen enforcment officials are pushing on the levers.

People will try to make this story about who did and did not benefit and argue this endlessly. But the utility of the project is showing everyone how the machine worked - which we're trying to do.