I am 100% certain at this point that Netflix does something similar with their "thumbnails" to appeal more to what they think the user's taste profile is. For instance showing female characters for a show that's as misogynistic as it gets to appeal more to a user that's primarily watched shows with female leads: e.g. Suits.
You make it sound like it's just a theory -- it's a well-documented fact. It's not hidden or nefarious or anything.
Yes, they absolutely generate a wide selection of thumbnails and show different ones to different users based on what kinds of shows they like.
E.g. if you like action, thumbnails will emphasize action. If you're into romance, thumbnails will emphasize romance.
I don't really see anything wrong with it. Any given show may be "about" lots of different things, and any single thumbnail will invariably emphasize one thing and totally ignore other things. It's not like there's any "one true thumbnail".
>This blog post sheds light on the groundbreaking series of A/B tests Netflix did which resulted in increased member engagement. Our goals were the following:
>Identify artwork that enabled members to find a story they wanted to watch faster [0]
Although it doesn't say that they give different thumbnails to different users, just that they tested different thumbnails for overall performance using A/B testing. So I'd like a source for customized per-user thumbnails. Although I only skimmed the article so I might have missed it.
Personally I'd rather not have a streamer buy, collect, trade my data so it can build a profile on me to optimize which ads they show me, what shows goes on my home screen, and what thumbnails they think will get me to watch what they want me to watch.
One of the major benefits of streaming services is that they give you personalized recommendations based on the profile they've built of what you've watched in the past.
Very few people want that feature to go away. It's not a bug, it's a feature. There's far too much content to sift through on your own, without personalized recommendations trying to help.
And if you don't want it used for targeting ads, then pay for the tier without ads.
Pretty sure many pr0n sites are doing something similar. I think they sample random timestamps to use as thumbnails for a/b testing and are also adding some AI filter eg. to make the actors look younger. Interesting to consider if the developers for these sites ever end working for FAANG/streaming companies which is kind of a scary thought.
> “The real challenge for all image tools in this space is battling Midjourney,”
I have a MidJourney subscription. I do not use it as much as I had hoped. Once in a while I can get it to generate something useful, but I still feel like it is hit or miss.
Today I was trying to get it to create a sort of cartoon vector art of a line of people waiting on a single bank teller. No matter what I did, I could not get it to keep a single line. It kept putting random people on the other side of the teller.
