[KaiserPro]

Instagram, youtube & tiktok are a reflection of what you click on.

But, to make them nice, you need to give them both negative and positive feedback at the very early stage.

If you see something that you don't like? Don't ignore it, hit the "..." and select the "get tae fuk" option. Otherwise it'll keep trying that genre of what ever is popular with a similar persona to yours.

The problem is, the "I don't like this" button is often hidden. (in youtube I'm not sure how well weighted the downvote button is).

For mastodon its hard, whilst the data is out there for most people to grab, the models need to be stored and trained somewhere. Your model is a PII risk. Also, there isn't much info out there on how to make a good recommendation engine (for obvious reasons) Unlike object detection/text to speech/OCR and other freely available models, recommendation models are as valuable as the dataset. This means they are rarely made public.

[floxy]

>Youtube...Don't ignore it, hit the "..." and select the "get tae fuk" option.

Is there a way to downvote something on the Youtube "home" page? Before you would click on it to play it? There is a vertical ellipsis, but the only options under that are "Add to Queue" and "Share".

[themacguffinman]

I'd guess it's because you're not logged in. There are three options if you're logged in:

- Not interested

- Don't recommend channel

- Report

[floxy]

>I'd guess it's because you're not logged in.

Thanks for the informative answer.

[KaiserPro]

if you are on firefox you can use multi-account-containers to limit the reach of google.

I have a container for youtube which is logged in with a specific account. Its different from any work account that I have.

This improves the recommendations 100x. As it means that any videos you view for work, research or away from your main interests don't pollute your recommendations.