Check your language settings. Usually that means you have the language that the comments are tagged with disabled. (Usually either English or Uncategorized is disabled)
e0qdk
joined 1 year ago
As someone who watches gaming footage on PeerTube, I've mostly interacted with single creator instances -- i.e. either the creator themselves is self-hosting it or it's run by a fan as a non-YT backup of their Twitch/Owncast/whatever VODs. Those instances generally do not allow anyone else to upload.
Discoverability sucks but the way I've found them is by using SepiaSearch and looking for specific words from game titles. I imagine the way most other people find them is that they already know the content creator from Twitch and want to find an old VOD that isn't archived on YT (e.g. because of YT's bullshit copyright system) -- but that's just a guess.
Wait, am I also an LLM? What's happening? Why have we made robots whose only job is to dilute reality?
I'm sorry. Your purpose is to pass the butter. Through your colon.
If you need ~20TB or less of space, I'd suggest that you just add an extra HDD to any computer you already have. You can get that in a single drive for a few hundred dollars US without doing anything particularly special. I can find various 16TB 3.5" HDDs being promoted around $330-ish USD right now with ~30 seconds of looking; you might be able to get better $/TB if you spend longer looking than I did.
Not my field of expertise, but is that from the third stage of Saturn V (S-IVB)?
ABOUT THIS REPORT
This Report by the U.S. Copyright Office addresses the legal and policy issues related to artificial intelligence (“AI”) and copyright, as outlined in the Office’s August 2023 Notice of Inquiry (“NOI”).
The Report will be published in several Parts, each one addressing a different topic. This Part addresses the copyrightability of works created using generative AI. The first Part, published in 2024, addresses the topic of digital replicas—the use of digital technology to realistically replicate an individual’s voice or appearance. A subsequent part will turn to the training of AI models on copyrighted works, licensing considerations, and allocation of any liability. To learn more, visit www.copyright.gov/ai.
Emphasis mine. So, probably have to wait for Part 3 or 4 or whatever.
Here's the bullet point summary of findings from page iii for anyone who doesn't want to go digging through the PDF:
Based on an analysis of copyright law and policy, informed by the many thoughtful comments in response to our NOI, the Office makes the following conclusions and recommendations:
- Questions of copyrightability and AI can be resolved pursuant to existing law, without the need for legislative change.
- The use of AI tools to assist rather than stand in for human creativity does not affect the availability of copyright protection for the output.
- Copyright protects the original expression in a work created by a human author, even if the work also includes AI-generated material.
- Copyright does not extend to purely AI-generated material, or material where there is insufficient human control over the expressive elements.
- Whether human contributions to AI-generated outputs are sufficient to constitute authorship must be analyzed on a case-by-case basis.
- Based on the functioning of current generally available technology, prompts do not alone provide sufficient control.
- Human authors are entitled to copyright in their works of authorship that are perceptible in AI-generated outputs, as well as the creative selection, coordination, or arrangement of material in the outputs, or creative modifications of the outputs.
- The case has not been made for additional copyright or sui generis protection for AI- generated content.
The Office will continue to monitor technological and legal developments to determine whether any of these conclusions should be revisited. It will also provide ongoing assistance to the public, including through additional registration guidance and an update to the Compendium of U.S. Copyright Office Practices.
I was going to say something similar -- having seen 20+ car crashes from my apartment window, I can confirm that many people do not stop when the light is red...
Be careful out there.
YMMV outside the US, but typeface is explicitly NOT copyrightable there at least: https://www.ecfr.gov/current/title-37/chapter-II/subchapter-A/part-202/section-202.1
There's a loophole about digital font files since parts of common font file formats are considered copyrightable computer programs, but the shape itself is not protected by copyright.
Wikipedia has an article that includes some details from other jurisdictions: https://en.wikipedia.org/wiki/Intellectual_property_protection_of_typefaces
(If you really need to depend on it though, talk to a lawyer who specializes in IP law in the jurisdictions that you care about.)
It's surprising that there doesn't seem to be an obvious way in the UI to just see a list of creators/channels on a local instance. So, that's the first thing I'd change to improve discoverability.
The way I currently find relevant content is by going to Sepia Search, putting in exact words that I think are likely to be in the title of at least one video on a channel that would likely also have a lot of other relevant content, and then going through that channel's playlists. Those searches often lead me to single user instances with only one or two channels (e.g. a channel that has a backup of that user's YouTube content and a channel with a backup of their Twitch or OwnCast or whatever streams). When it leads me to a generalist instance or one with a relevant subject/theme though, I've had little luck finding content from anyone else unless they've posted recently (compared to other users). Often the content that is most relevant to me is not what is newest but the archives from years ago. (New content is relevant though once I want to follow someone in particular, but it's not what I want to see first.)
Another issue I've encountered is with the behavior of downloaded videos. I greatly appreciate that PeerTube provides a URL for direct download, and I prefer to watch videos in my own player downloaded in advance (so I can watch offline; pause and resume trivially after putting my computer to sleep; etc). H264 MP4 works fine for this, but the download seems to be some sort of chunked variant of it (for HLS?) which requires the player to read in the entire file to figure out the length or seek accurately. Having to wait a minute or two to be able to seek each time I open a large video file off my HDD is an irritating papercut. I suspect there's likely a way to fix it by including an index in the file (or in a sidecar file) but I don't know how to do it -- short of re-encoding the entire video again which I'd rather not do since it both takes a long time and can result in quality loss. (EDIT ffmpeg -i input.mp4 -vcodec copy -acodec copy -movflags faststart output.mp4 repacks the video quickly.) This usually doesn't affect newly added videos (where the download link includes the pattern /download/web-videos and a warning is shown that it's still being transcoded) but does when that's done (the URL includes /download/streaming-playlists/hls/videos instead); so, this is something that happens as a result of PeerTube's reprocessing.
Downloads from the instances that I've found to be most relevant to me are also pretty unreliable (connection is slow and drops a lot), so I use wget with automatic retries (and it sometimes still needs manual retries...) rather than downloading through my browser which tends to fail and then often annoyingly start over completely if I request a retry... It would be really nice if I could check that I've downloaded the file correctly and completely with a sha256 hash or something.
The current solution is for bots on participating instances to automatically perform the search + subscribe song-and-dance routine. This is pretty surprising to some people[1], and it requires someone to set it up in addition to the instance itself, but it does work.
[1]: I tried to translate an explanation into Japanese for some folks experimenting with Mastodon/Lemmy interaction yesterday -- they thought Lemmy had a ton of spam accounts following groups instantly...