I've been following Rekieta's stream for about a week, it was ok until today. I doubt that it was automated, probably reported by someone.

Hadn't thought of that. I've been listening intermitently, and the fact that only today - and during Jury instructions no less - does the channel go down make me think that yeah, someone reported it, and the automation probably kicked in after that. Simplest explanation I think.

You are missing the fact that automated copyright take-down systems aren't perfect and are likely designed to learn towards taking fair things down, and not living infringing things up.

I'm not missing anything. If this is automated, you're telling me they can't put in a rule that says "if public access, then take-down not apply"? Maybe you're making a different claim that a network stream (from CBS or another) is not eligible for re-stream? That I find more compelling, but then again it also shouldn't be hard to add another layer to that rule.

Is all the metadata really in place such that CBS' and other networks' streams are able to be labeled as sourced from public access? I don't know, but my assumption is that it isn't, and that large media companies wouldn't bother to use it even if it's technically feasible.

Given the lack of such metadata on the streams which typically have protected content, the system taking these down probably just looks at similarity of content to streams which are proprietary and silences them "to be safe".

I think the conclusion I'm drifting toward is thus "YouTube could make it possible for wonkish commentary on live public access topics by creators on the platform, but they don't care enough to do it", which makes sense.

How exactly would they know that it's public access? Their algorithms don't understand what content is, it just recognizes patterns in data, and that one snippet of video is sufficiently similar to another as to be considered the same snippet. Putting in a rule to deal with an edge case is non-trivial, and there are an infinite number of edge cases.

Pretty sure YouTube employees have already classified all government channels and public access channels as such. There's a notice below those videos that indicates it. And if not, it should be done anyway.

They are pretty smart, I guess they could reuse the same technology they use to not to take down videos with "correct" views that they agree with?

Yawn. Just a conclusion looking for its argument. I'm not sure why this kind of talk is tolerated here on these kinds of topics but other topics its widely downvoted for being the baseless and specious nonsense that it is.

You don't have to make the algorithm do it, is what I'm saying.

It might not be an issue of the difficulty. All of these will inherently be error prone, so why automate something that isn't going to significantly going to reduce your workload in the end, while also simultaneously annoying the advertisers that actually pay YouTube.