The original algorithm rewarded engagement absent dates, but this resulted in old classic hits mopping up revenue while newer stuff struggled to grab anyone’s attention. You’ll never make a music video more popular than Rick Astley’s Never Gonna Give You Up, so why bother trying?
Then the algorithm shifted to fresh-first bias, which incentivized streamers to constantly churn out new content. But it still contended with users who stubbornly wanted to see the old content. So you got a bunch of content that tried to imitate historical hits or play on trends. This ended up producing 10,000 videos named some variation of “Rick Astley’s Never Gonna Give You Up, Explained” with a digitally edited picture of the singer with big eyes and a soy face.
Now we’ve got this deluge of AI generated crap that nobody wants to look at or search for, piling up in YouTube’s back catalog. The only way to justify hosting it is to jam it into someone’s feed. So every user is being A/B Tested once again, with a new procedurally generated wall of garbage that will eventually narrow down what any given individual is most likely to click on and watch. Then we can solve both of the problems above. Always have new content, but its technically “fresh” rather than a rehash of some prior release.
We are doing Monkeys On Typewriters because someone at YouTube HQ decided it was better than letting anyone watch the Rick Astley video one more time.
I desperately with YouTube had real competition.
There are other places to host video, but they tend to be very boutique or with an abundance of very low quality content. That, plus YouTube leveraging economies of scale and the networking effect means there’s nowhere else you’d ever want to try and host a video, unless you were looking to reach a very boutique audience or you were putting out material you didn’t really expect anyone to watch.
The original algorithm rewarded engagement absent dates, but this resulted in old classic hits mopping up revenue while newer stuff struggled to grab anyone’s attention. You’ll never make a music video more popular than Rick Astley’s Never Gonna Give You Up, so why bother trying?
Then the algorithm shifted to fresh-first bias, which incentivized streamers to constantly churn out new content. But it still contended with users who stubbornly wanted to see the old content. So you got a bunch of content that tried to imitate historical hits or play on trends. This ended up producing 10,000 videos named some variation of “Rick Astley’s Never Gonna Give You Up, Explained” with a digitally edited picture of the singer with big eyes and a soy face.
Now we’ve got this deluge of AI generated crap that nobody wants to look at or search for, piling up in YouTube’s back catalog. The only way to justify hosting it is to jam it into someone’s feed. So every user is being A/B Tested once again, with a new procedurally generated wall of garbage that will eventually narrow down what any given individual is most likely to click on and watch. Then we can solve both of the problems above. Always have new content, but its technically “fresh” rather than a rehash of some prior release.
We are doing Monkeys On Typewriters because someone at YouTube HQ decided it was better than letting anyone watch the Rick Astley video one more time.
There are other places to host video, but they tend to be very boutique or with an abundance of very low quality content. That, plus YouTube leveraging economies of scale and the networking effect means there’s nowhere else you’d ever want to try and host a video, unless you were looking to reach a very boutique audience or you were putting out material you didn’t really expect anyone to watch.