Videos until now were thought to be immutable. If someone removes a point of controversy from a video, a person linking to it as proof of that controversy suddenly looks like a liar.
But that's just indicative of the HN bubble, being satisfied technically correct reasoning rather than examining the issue more holistically — that is, this question isn't about what is or was technically capable of being achieved but how regular people interpret these things.
It is generally assumed that if something was updated, people should be made aware. Updated comments, updated news articles, updated anything, there is always some indication of an edit or update.
Why should video be any different?
However, when you visit that video's youtube page, the video date was listed as something like "Live-streamed on ...", with no note of it being edited. You are given the impression that what you are about to view is what was streamed, when it's not.
The edited version is a better video, but don't tell me I'm watching the same thing as those who saw it live during the stream.
You want a more immutable source, archive the video and link to that. Same reason people archive other websites when wanting to link to something as “proof”. Because we know that the web server is not write once.
I think they're talking about the actual content hosted at a specific url, not just video editing in general.
You know that. People don't.
I dont particularly care if YouTube allows people to edit videos, but on anything with comments, allowing someone to know it was edited seems like good information management, and in this case, helps inform people not in tech that it is editable.
The more obvious the indicator that it was edited, the more people will understand that it is an option.
Sure, YouTube can do what they want, but if they do, it destroys something that is valuable about YouTube, which is that you can trust that the ratings and comments are about the video you are watching.
Good to know now.
Maybe youtube could embed something like a signature in the video metadata so that a video file could be inspected for tampering?
The average person isn't going to know how to use a command line program to download a Youtube video. The average person also doesn't have enough space to maintain large and old archives. The average person wouldn't care to maintain and prune an archive for relevance every now and then. And even if you manage to get the average person to care about all of this - you still face the hurdle of there not being any "official" hash to compare to.
this practice is quickly becoming a legal nightmare. Sites that let you download youtube videos are being taken down left and right these days with RIAA lawyers going after many of them for cash settlements. How soon before they go after projects like youtube-dl and other similar programs?
I understand if it's a third-party site, because they'd have to make a copy of the video for you to have downloaded it (and only if it's one of those copying types - if all the site did was to compute the download-link, then it doesn't fall foul of any copyright at all).
Youtube is the weakest link here though. If the scumbags working for the media industry put enough pressure on youtube to get into an arms race against the program (and others like it) things could get ugly fast. Especially as DRM worms it's way deeper into the most fundamental parts of both the internet and our devices. In the end, we'll always have the analog hole but software like youtube-dl is still vulnerable.
(I really hope you're asking out of intellectual curiosity and not because you're affiliated with the RIAA/BVMI/IFPI/ETC)
but i think the attacks on the devs will be fruitless, unless they can somehow intimidate them without actually suing them (in which case, it's a baseless suit).
Youtube disallowing the app is a possibility - after all, the T&C of youtube may be changed to exclude anyone they want. But the youtube website will need to somehow allow the browser, but restrict non-browsers (or non-compliant browsers). DRM is the only thing i can think of which could achieve this goal - and that is one reason why i hate that google managed to put in DRM into the web standard.
identifying a website vs a command line tool can be difficult. youtube-dl already supports custom user agent strings and cookies. youtube might check for things like installed fonts, look for randomly named web bugs in cache, or even try checking page load/render times. Once folks figure out what youtube is looking for they can try to correct for it but youtube can counter with something else and certainly make things complicated.
This also wouldn't really work because a YouTube video isn't a single file. Videos are available in different qualities and formats. The video component tends to be separated from the audio component. If you want to combine them into a single file, you have to mux them. However, this process is not deterministic, so the resulting hash would differ depending on a multitude of factors.
Then Youtube would have to publish the hashes for each and every file that they encode the source material to. Practically impossible and especially unverifiable if YouTube ever disappears (or the source gets deleted).
A digital signature of the hash, embedded in the metadata and computed similarly to Reproducible Builds... doesn't take much space and especially only requires knowledge of the public key for verification.
Make it private to make sure the views stay below 100k, then edit as needed out anything but the correct temperatures, and spam it all over social media when the time comes. I couldn't think of good examples to do it with words, but there isn't much you couldn't do, given that videos can be 10 hours long. It's not like anyone would mind the jump cuts much, just put those everywhere else, too.
Jump cuts. Esp if you are filming yourself speaking into camera. If you are doing audio only it would work better. The reason it works with removing a segment is because you are removing content from an existing cut to another existing cut so it will feel much more seem less.
Now if it wasn’t for the date changing when the privacy state of the video changes, why not record 100 takes, upload each and release the correct one and discard the rest. It would give a much more convincing video with no way to detect splices in the audio (a common way fake speed runs are detected) and no need to deal with YouTube’s video editor.
What I mean is if you are going to have jump cuts right at the important part of the video people are going to call you out, if you have jump cuts throughout people are going to call you out. Something like this would just shine light on the video and people will start picking it apart like a badly knitted jumper.
Also one of the first comments would be something along the lines of “nifty trick, now do it again?”.
Isn't that a known scam, actually? IIRC it was something like emailing people with the prediction of the outcome of a sports match (let's say 3 options: team A wins, team B wins, or tie) -- one third gets prediction A, another third prediction B, the rest prediction C. Keep track of whom you sent what, then do it again for the people to whom you sent the correct prediction. I think the idea is to end up with a small but very convinced pool of people that might fall for the actual scam, e.g. the offer to predict something with a very valuable outcome, at a charge.