Not with the crrent ones, it’s easier to difference ads by code than by content in most Vids, eg, divulgations, news, influencers, etc. YT, to not destroy the own business modell, also avoiding that there later also 5 years old vids with outdated ads, must use some kind of dynamic insertion, that means, it can be discovered and skipped by some userscripts, for sure not in the extension stores, but in Greasyfork or OpenuserJS, which are independent from Google influence.
Can we train a local AI model to recognise ads?
Not with the crrent ones, it’s easier to difference ads by code than by content in most Vids, eg, divulgations, news, influencers, etc. YT, to not destroy the own business modell, also avoiding that there later also 5 years old vids with outdated ads, must use some kind of dynamic insertion, that means, it can be discovered and skipped by some userscripts, for sure not in the extension stores, but in Greasyfork or OpenuserJS, which are independent from Google influence.