[orf]

Seems like a lot of out of context clips to push a narrative, standard project veritas stuff.

Like the clip about banning certain groups of people: the engineer says "it's going to ban a certain way of talking". You might take from this based on the video that he's talking about inadvertently banning right-leaning users. Well, what if he's talking about banning racism or messages that use ((subtle dog-whistles)) instead of overt racism?

Also, not many people like Trump, especially not in left-leaning areas (like where Twitter is based). Are employees not allowed to voice their dislike of a very dislikable man, while drinking at a social event?

[trs80]

If you watched the whole thing you'd see the context of each interview.

[orf]

I did, and it's just longer out of context videos that cut off at convenient points:

> Journalist: "Would you say the algorithms block liberal or conservative users?"

> Engineer: "I would say the majority of it are for Republicans, because they are all from Russia and wanted Trump to win"

> Journalist: "So you would mostly just get rid of Conservatives?"

> Engineer: "Yes"

> Immediate cut.

Woah, what a compelling argument. The guy is saying they are using machine learning to block Russian bots masquerading as American conservatives that are influencing the narrative in favor of Trump (that totally don't exist btw because we love Russia now and they would never attempt to meddle in any other countries elections). The video attempts to twist that into him saying they block conservatives.

This study is an interesting read: http://comprop.oii.ox.ac.uk/research/polarization-partisansh...

And an associated news article: https://www.theguardian.com/technology/018/feb/06/sharing-fa...