YouTube's algorithm is built to maximize engagement for a broad audience. Especially shorts with animated "jumping" captions and videos that zoom in and out every three seconds providing no useful content most of the time.
I found it really frustrating that almost every tech conference talk ends up on YouTube, yet it's surprisingly hard to discover the ones I'd find interesting.
So I wrote a script that fetches all the talks from the channels I follow each week, groups them by conference, and sorts them by view count. That way, I spend about five minutes a week and quickly find the really good ones.
In 2024, I turned it into a newsletter. I use the exact same script and I just copy and paste the results into Substack. It turns out many people are facing the same problem because it grew to over 8,400 subscribers as of today.
Because for every one person that finds the recommendation system bad there must be ten others that find it so spot on that they spend 8 hours watching and clicking. It's the only explanation that makes sense.
Having your feed full of a thing you've recently watched at least makes sense.
What boggles my mind is videos that have been sitting in my Recommended for a few months now - if I haven't clicked it the first 45 times, I probably won't click it the 46th.
I wonder if you can delete your entire existing history to sort of reset the recommendation engine? i've definitely had to delete individual ones from my history to downrank similar ones in the past.
Because google decided to prioritise profit above all, so all their services were made to keep you engaged - more searching, more clicking, more viewing, more time wasting....by giving you bad results, forcing you to work for it and stay on their platforms.
I'm not convinced that showing me videos that I won't click is a more profitable strategy than showing me videos that I will click. I think they're just not very good at their job.
It's not about clicking, it's about time spent on the platform, looking for what you're actually after. The longer it takes you, the bigger their attention numbers get and the more ads they can show you.
it used to be a ton better but its just optimized now so we are more likely to click on things we see rather than things that are interesting or relevant. they want us to click so we see the ads lol.
YouTube's algorithm is built to maximize engagement for a broad audience. Especially shorts with animated "jumping" captions and videos that zoom in and out every three seconds providing no useful content most of the time.
I found it really frustrating that almost every tech conference talk ends up on YouTube, yet it's surprisingly hard to discover the ones I'd find interesting.
So I wrote a script that fetches all the talks from the channels I follow each week, groups them by conference, and sorts them by view count. That way, I spend about five minutes a week and quickly find the really good ones.
In 2024, I turned it into a newsletter. I use the exact same script and I just copy and paste the results into Substack. It turns out many people are facing the same problem because it grew to over 8,400 subscribers as of today.
What boggles my mind is videos that have been sitting in my Recommended for a few months now - if I haven't clicked it the first 45 times, I probably won't click it the 46th.