4 ms·

There are a lot of problems with this response, but it makes one argument in particular that seems very foolish and I have not seen otherwise mentioned: that In

There are a lot of problems with this response, but it makes one argument in particular that seems very foolish and I have not seen otherwise mentioned: that Invidious is somehow not subject to YouTube terms because it doesn't use the public API. But it does load the videos from YouTube's media CDN - so if the author is making the claim that Invidious is not using YouTube "the service" when it still relies on YouTube infrastructure for the heavy lifting this seems... Actually worse? It takes a contract issue over the terms and makes it into something more like a CFAA issue, since Invidious is consuming YouTube resources without, the author claims, using anything intended as a public service. Obviously this analogy is inexact but it feels like telling the electrical utility that they can't sue you because you aren't using their electrical service, you're tapping the wires before the meter. You kind of want to be governed by the terms of a service because that tends to mean that whatever you're doing is something you're allowed to do, and not outright abuse.

pcthrowaway 3y ago
So Google scrapes the entire internet, and even caches thumbnails of images. If I have a website where I host some of my own images, and I outline a terms of use regarding use and/or caching of those images, is Google obligated to follow those terms of use? Also, I don't think Invidious (the software repository) loads anything from Youtube's CDN, any more than curl does, though it seems like the default configuration does do this, and they give out plenty of advice on how to make it work as a "proxy" for the youtube video data. If the people deploying Invidious are infringing, then I agree there's a good argument that Invidious (the software) has some liability. Though it's not clear to me if people doing this are infringing. As far as I can tell, no one has been successfully sued for hotlinking images hosted by another website. This seems like the same precedent to me. Furthermore, if hotlinking was considered infringement, then a case could be made that most websites would be guilty enough to be de-indexed from Google. Someone (in the UK) tried to sue Google a while back because Google was prominently displaying search results that hotlinked his website, and he lost the lawsuit (https://www.theregister.com/2020/01/16/google_search_image_results_hotlinking_high_court/ https://www.theregister.com/2020/01/16/google_search_image_r...) Until there's clarity on whether it's OK for websites to embed resources from other hosts, I don't think it's clear that even Invidious instances are infringing here. And even if they are, Invidious (the software) could very well make some small changes to their deployment, to make it configurable, for users to (theoretically) use their own video servers. Users can then configure their own instances to use Google's CDNs if they want, and I don't believe the people making Invidious would be any more responsible than Daniel Stenberg is for the ways people might use curl
- circuit10 3y ago
  I’m the other hand you could claim a DDoS is just visiting the site a lot of times but that’s illegal, I think this is somewhere in between so it’s not that clear
  - pcthrowaway 3y ago
    DDoS is illegal because of the intent. No one gets in trouble when a site gets hugged to death
    - circuit10 3y ago
      And that’s the thing here as well: linking to a website has a different intent to embedding an internal resource from it that isn’t meant to be accessed through anything other than the main site in a way that takes away revenue from the site To be clear I’d like projects like this to exist but I’m not convinced that it’s legal under current laws
- jcrawfordor 3y ago
  Sure, but Google's crawler (mostly) obeys robots.txt and other directives. It seems to be one of the most responsible crawlers this way. Here, Invidious is explicitly being told to stop and refusing to do so. In a lot of these situations there's not much wrong with doing it until you're told to stop. Invidious has just been told to stop. Google used to have services that felt a lot more like this, like Google Images before it got aggressive about using cached versions and old, old Google Videos. But they got rid of all of those, I suspect because of thinking along these lines. It was very much an open legal issue around newspaper publishers, for example, and while Google is far from innocent they did make changes to mitigate their legal risk. Google seems to be pretty consistent in the idea that serving up content from someone else's service, in their products, is not acceptable. All of AMP kind of got built around that issue. And I don't think Google wanted it that way, but the legal landscape is the legal landscape. If Google is bound by these rules others have to be as well.
- dspillett 3y ago
  > So Google scrapes the entire internet, and even caches thumbnails of images. If I have a website where I host some of my own images, and I outline a terms of use regarding use and/or caching of those images, is Google obligated to follow those terms of use? It isn't quite the same as the hypothetical case you lay out, but a number if news organisations have argued that Google should follow some rules about how they used the scraped information. Any ruling against Google here may be relevant against Invidious and similar products.
photonerd 3y ago
Google is unlikely to make that argument given that their entire search monopoly rests on abusing exactly that.

pcthrowaway 3y ago

So Google scrapes the entire internet, and even caches thumbnails of images. If I have a website where I host some of my own images, and I outline a terms of use regarding use and/or caching of those images, is Google obligated to follow those terms of use? Also, I don't think Invidious (the software repository) loads anything from Youtube's CDN, any more than curl does, though it seems like the default configuration does do this, and they give out plenty of advice on how to make it work as a "proxy" for the youtube video data. If the people deploying Invidious are infringing, then I agree there's a good argument that Invidious (the software) has some liability. Though it's not clear to me if people doing this are infringing. As far as I can tell, no one has been successfully sued for hotlinking images hosted by another website. This seems like the same precedent to me. Furthermore, if hotlinking was considered infringement, then a case could be made that most websites would be guilty enough to be de-indexed from Google. Someone (in the UK) tried to sue Google a while back because Google was prominently displaying search results that hotlinked his website, and he lost the lawsuit (https://www.theregister.com/2020/01/16/google_search_image_results_hotlinking_high_court/ https://www.theregister.com/2020/01/16/google_search_image_r...) Until there's clarity on whether it's OK for websites to embed resources from other hosts, I don't think it's clear that even Invidious instances are infringing here. And even if they are, Invidious (the software) could very well make some small changes to their deployment, to make it configurable, for users to (theoretically) use their own video servers. Users can then configure their own instances to use Google's CDNs if they want, and I don't believe the people making Invidious would be any more responsible than Daniel Stenberg is for the ways people might use curl

circuit10 3y ago
I’m the other hand you could claim a DDoS is just visiting the site a lot of times but that’s illegal, I think this is somewhere in between so it’s not that clear
- pcthrowaway 3y ago
  DDoS is illegal because of the intent. No one gets in trouble when a site gets hugged to death
  - circuit10 3y ago
    And that’s the thing here as well: linking to a website has a different intent to embedding an internal resource from it that isn’t meant to be accessed through anything other than the main site in a way that takes away revenue from the site To be clear I’d like projects like this to exist but I’m not convinced that it’s legal under current laws
jcrawfordor 3y ago
Sure, but Google's crawler (mostly) obeys robots.txt and other directives. It seems to be one of the most responsible crawlers this way. Here, Invidious is explicitly being told to stop and refusing to do so. In a lot of these situations there's not much wrong with doing it until you're told to stop. Invidious has just been told to stop. Google used to have services that felt a lot more like this, like Google Images before it got aggressive about using cached versions and old, old Google Videos. But they got rid of all of those, I suspect because of thinking along these lines. It was very much an open legal issue around newspaper publishers, for example, and while Google is far from innocent they did make changes to mitigate their legal risk. Google seems to be pretty consistent in the idea that serving up content from someone else's service, in their products, is not acceptable. All of AMP kind of got built around that issue. And I don't think Google wanted it that way, but the legal landscape is the legal landscape. If Google is bound by these rules others have to be as well.
dspillett 3y ago
> So Google scrapes the entire internet, and even caches thumbnails of images. If I have a website where I host some of my own images, and I outline a terms of use regarding use and/or caching of those images, is Google obligated to follow those terms of use? It isn't quite the same as the hypothetical case you lay out, but a number if news organisations have argued that Google should follow some rules about how they used the scraped information. Any ruling against Google here may be relevant against Invidious and similar products.

photonerd 3y ago

Google is unlikely to make that argument given that their entire search monopoly rests on abusing exactly that.