Do you happen to know if there are any project or talk of not archiving the video content itself, but instead the transcripted content instead? I feel that this would be very advantageous to archive knowledge based (versus delivery based, such as prank and stand up comedy) videos much more efficiently than the videos themselves.

Someone below brought up a very good point about many of these videos being much longer than they need to be (mainly for reimbursement and ad reasons). If the transcripted content can be archived, it could also be abridged and/or summarized as well as being combined with other similar video content as well.

I'm sort of thinking as I go here now, but I would think that perhaps Youtube has an API that lets you access the closed captions of videos?

I would be kinda worried that one day youtube just send them a take down notice because it violates something in their eula.

and you can use yt-dlp to download bits you want to save yourself.

Yeah that is what ArchiveTeam is using.

Just wonder can't we just start web torrent/torrent or IPFS these video files with some kind #YTPubArchive tag?

Don't you need to have an account for that? Already a barrier I am not willing to take.

No. yt-dlp downloads most videos without needing a yt (or any) login. A very few end up gated behind some form of login, but I suspect those are creator specific.

I find login walls on YT only when needing to confirm age for some reason: explicit or suicidal content, age checks, etc

Same.