zlacker

Google absolutely can do all of those things without an algorithm. What they can't do is accomplish that without impacting profit margins (or at the minimum, executive bonuses). "If it impacts business as usual, then it is impossible" is a naive/flawed/libertarian stance.

>>ilikeh+(OP)
I’m usually a proponent of the “wall garden” when it comes to applications and strict sandboxing for most users, since software can harm your computer.

But in the case of YouTube, there is absolutely no way that they can curate it and it still being as open as it is.

>>scarfa+Z
There is no need to curate every video, only the ones qualified enough to be recommended/showcased to the public who is not explicily looking for them.

>>ilikeh+(OP)
YouTube does human curation already. They are refered to as "playlist" and every user has the ability to create and share them. So what you are asking for is Google to create their own playlist? Would this also entail removing that ability from other users?

>>ilikeh+(OP)
You do realize that to cover current needs (400h uploaded every minute), YouTube would need to employ more than 72000 people working full time right?

>>scarfa+Z
> But in the case of YouTube, there is absolutely no way that they can curate it and it still being as open as it is.

So?

If YouTube exits the space and allows oxygen back into the video sharing market, we might actually get some different video sharing services that do different things (a la NicoNicoDouga).

>>ilikeh+(OP)
You made me curious so I did some back-of-the-envelope math. An average of 576K hours of video is uploaded to YouTube every day [1], which is 4.032M hours per week. If the reviewers watch all the video at 1x speed and work 40 hours per week, you'd need about 100K reviewers to do the job. (This is just to watch the video -- not including any additional work done to annotate the video with whatever information you want out of your reviewers.) If each one costs $30K a year (probably a lowball estimate including salary, insurance, etc.) it would cost a total of $3B per year. YouTube makes $4B in revenue per year and roughly zero profit AFAICT, so there's no way this is feasible.

[1] https://www.quora.com/How-many-videos-are-uploaded-on-YouTub...

>>ilikeh+y1
Say I watch a video on a topic like "nes video game speed running". Right now I'd see other nes video game speed running videos, it's very useful. In a curated world, what would be recommended? It's probably too much of a niche topic to yield results that would be very useful.

>>xondon+P1
And these people would inevitably make some number of mistakes in categorization too, or miss something, or just be unable to quite hit some baseline universal standard that doesn't upset a group. Then YouTube still gets the bad press.

>>xondon+P1
But 99.9% of all videos uploaded never gets more than a few handfuls of views so those are irrelevant. Of the remaining 0.1%, you don't need to watch every second of every frame - speeding it through at twice the speed should be doable. So by your own calculations, 72 000 * 0.001 * 0.5 = 36 people working full time.

>>bsder+42
Video streaming, processing, and storage at scale still costs a lot of money. I don’t think even Google is doing it profitably.

>>bjourn+S7
You can set that 0.001 factor as big or as low as you like, but then we’d get the same nytimes hit piece saying this is intentionally being done by humans.