This is wildly unlikely. the content scope of twitter is huge, the data scale is huge and they have so many engineers that it's difficult to keep "manual human

This is wildly unlikely. the content scope of twitter is huge, the data scale is huge and they have so many engineers that it's difficult to keep "manual human intervention" confidential. Even something as simple as a keyword search, ie "hunter biden laptop", would require searching tens of millions of strings per hour. Distributed across thousands of servers, integrated with whatever dev ops set up they have. It just seems like a huge pain, for little practical gain. Their scale is so huge that I doubt twitter has much manual control over what people see, even if they wanted it. It seems much easier to just multiply weights