Or a hypothesis...
In all the talk about opinion moderation, it looks like folks are missing a big technical point: If you were Elon Musk and walking into the ownership of Twitter for the first three minutes, you might wonder to yourself...
"Why do you need 16,000 employees to run a website?" (or whatever the staffing number is or was)
Evidently the previous administration DID need them for something. Ordinarily you'd think that a programming and development department plus some sales and back office would be all you'd need to run a big conversational text board. I'd peg that at around 400 people. But the others I suspect were at least a vast division of cat herders, breaking up the worst arguments among a mass of humanity with too much time on its hands, getting onto the site basically to argue with each other.
If there ACTUALLY IS an effective way to do that, you have two choices: either use computers to automate the process (which automation would be the main justification of the computer in the first place) or do it with manual, analog people. Fine - but what happened? It seems both choices were made. They have both automation AND a few aircraft carrier loads of people. What the hell??
It might be that in pursuing machine learning, artificial intelligence and algorithm creation goals, they didn't make as much progress as their social media competition (which also may not be making much progress either). If their AI is failing, that's why they need the other X thousand people - to amend the decisions the AI misinterpreted, failed to identify or got wrong some other way. If you can't just script this stuff with a keywords list then you have to automate - but if that's inefficient, you have to have an army of moderators TOO.
This would explain Twitter having to find a buyer. Ad revenue isn't enough to support a contingent of AI developers that make six-figure incomes plus an army of moderation people at the same time.
