Tried Andrej Karpathy's latest video -- https://www.youtube.com/watch?v=7xTGNNLPyMI It said "too long video" Exactly, duh ;D
[dead]
TL;DW: Too Long; Didn't Watch Distill YouTube Videos to the Relevant Information
201–210 of 213 posts
Re: TL;DW: Too Long; Didn't Watch Distill YouTube Videos to the Relevant Information
#202[dead]
Re: TL;DW: Too Long; Didn't Watch Distill YouTube Videos to the Relevant Information
#203What I'd really like is a service that edits down YouTube videos by removing all the stock footage and talking head crap, then speeding up the audio to fit over the remaining novel information—whether that's new battlefield footage, electron micrographs, demonstrations of machining techniques, or just elephant toothpaste. The talking head filler seems like it would be easy to recognize, but stock footage recognition…
how would you remove stock footage? B-roll has voice over it, it's kind of a necessity.
Re: TL;DW: Too Long; Didn't Watch Distill YouTube Videos to the Relevant Information
#204What I'd really like is a service that edits down YouTube videos by removing all the stock footage and talking head crap, then speeding up the audio to fit over the remaining novel information—whether that's new battlefield footage, electron micrographs, demonstrations of machining techniques, or just elephant toothpaste. The talking head filler seems like it would be easy to recognize, but stock footage recognition…
I don't watch YouTube but if I would / I would if I'd cut out anything with faces or speech & use an LLM to summarize what's technically relevant from the transcript in a way that fits length of what remains.
Pipeline such content, but use weighted random videos, with low weights for types of content with clickbait headings & perhaps blacklist for words like meme or lol in transcript to cut out things with stock footage. I am not sure of exact best way to remove it, actually, other than "using the transcript for some computational technique of probabilitistic stock footage prediction" which I bet would be most effective.
Re: TL;DW: Too Long; Didn't Watch Distill YouTube Videos to the Relevant Information
#205Earlier quoted context omitted.
One of the best things in SponsorBlock is Highlight segments. If the video is 10 minutes of filler building up to a single interesting/exciting moment you can often see exactly on the timeline where to jump to.
That's also present in the regular YouTube video scrubber. Maybe less prominently?
I think that only shows up when a video hits a critical number of views so it knows where people watched and rewatched.
Re: TL;DW: Too Long; Didn't Watch Distill YouTube Videos to the Relevant Information
#206I thought maybe I'd finally be able to through a Wendigoon video with this, since I'm generally interested in the things he talks about but can't stand some of his linguistic tics. Unfortunately looks like most of his videos are too long for this, which I guess is ironic considering the name. TL;DP
You could try my app https://github.com/rmusser01/tldw
Supports arbitrary length videos and also lets you choose what LLM API to use.
Re: TL;DW: Too Long; Didn't Watch Distill YouTube Videos to the Relevant Information
#207What I'd really like is a service that edits down YouTube videos by removing all the stock footage and talking head crap, then speeding up the audio to fit over the remaining novel information—whether that's new battlefield footage, electron micrographs, demonstrations of machining techniques, or just elephant toothpaste. The talking head filler seems like it would be easy to recognize, but stock footage recognition…
Cool.
What's stopping you from doing it?
Re: TL;DW: Too Long; Didn't Watch Distill YouTube Videos to the Relevant Information
#208What I'd really like is a service that edits down YouTube videos by removing all the stock footage and talking head crap, then speeding up the audio to fit over the remaining novel information—whether that's new battlefield footage, electron micrographs, demonstrations of machining techniques, or just elephant toothpaste. The talking head filler seems like it would be easy to recognize, but stock footage recognition…
Cool. What's stopping you from doing it?
Watching too many YouTube videos, probably.
Re: TL;DW: Too Long; Didn't Watch Distill YouTube Videos to the Relevant Information
#209It seems to have some reasonable length limitations. It refused to distill this epic analysis of dungeons across 3 games, since the video approaches 4 hours long: https://m.youtube.com/watch?v=PajArJbPfpE
Re: TL;DW: Too Long; Didn't Watch Distill YouTube Videos to the Relevant Information
#210What I'd really like is a service that edits down YouTube videos by removing all the stock footage and talking head crap, then speeding up the audio to fit over the remaining novel information—whether that's new battlefield footage, electron micrographs, demonstrations of machining techniques, or just elephant toothpaste. The talking head filler seems like it would be easy to recognize, but stock footage recognition…
how would you remove stock footage? B-roll has voice over it, it's kind of a necessity.
By "speeding up the audio to fit over the remaining novel information", as I said. "B-roll" doesn't necessarily imply worthless stock-footage filler; in https://www.youtube.com/watch?v=woj4vfMLpao or https://www.youtube.com/watch?v=DdF_nzMW_i8 or https://www.youtube.com/watch?v=Eu_crbcBdNM, for example, the B-roll is the remaining novel information. The third of these also includes the kind of talking-head filler I'd like to remove.