51–60 of 168 posts
Earlier quoted context omitted.
If you mean OpenAI's Whisper, has anyone created an actual product that people can use with downloaded videos?
whisper.cpp runs locally and can output .srt, I’ve been using it a lot (takes a while though)
Really quite amazing
Earlier quoted context omitted.
If you mean OpenAI's Whisper, has anyone created an actual product that people can use with downloaded videos?
Well. I don't know, but in my experience it shouldn't be too hard. The most difficult part would be timestamps I guess.
> Your consumer can query the API on its own, and download 5 subtitles per IP's per 24 hours, but a user must be authenticated to download more. Users will then be able to download as many subtitles as their ranks allows, from 10 as simple signed up user, to 1000 for VIP user. > If you wish to offer subtitles without prompting for an opensubtitles account, please contact us and tell us about your project... For an ea…
Nice, so now we have OpenAI and OpenSubtitles that has 'Open' in their name and not open :)
A classic tactic used by conmen. Make the victim trust you, then snatch their belongings. Easier than open confrontation.
Earlier quoted context omitted.
Like some open source projects moving to a "source available" or "fair code" licenses, it is going to be bumpy finding equilibrium as to what is reasonable. Tragedy of the commons for sure. There are costs to cover, and usage profiles can vary wildly.
And when they do that people just fork and go along with their own. Tofu being a great example, opensearch somewhat less.
If you have a benevolent dictator with deep pockets, throttling usage and retrieval is less of a concern. The Internet Archive comes to mind, as does Wikipedia. Reddit's outcome happened because they were a discussion board utility masquerading as a 10-100x VC funded business, and their actions drove away the concentration of mods and contributors who were putting in all the free work (broadly speaking).
Earlier quoted context omitted.
Yes you are wrong because a movie script and movie subs are not the same thing at all
In what way? The subs are the script (often translated) + timestamps. It’s the definition of a derivative work. The subs wouldn’t exist without the original work. Edit: Yes, it’s a subset of the script and not thee script itself, and being a pedant about that isn’t a useful point. It’s just pedantic.
Earlier quoted context omitted.
Yes you are wrong because a movie script and movie subs are not the same thing at all
In what way? The subs are the script (often translated) + timestamps. It’s the definition of a derivative work. The subs wouldn’t exist without the original work. Edit: Yes, it’s a subset of the script and not thee script itself, and being a pedant about that isn’t a useful point. It’s just pedantic.
Earlier quoted context omitted.
In what way? The subs are the script (often translated) + timestamps. It’s the definition of a derivative work. The subs wouldn’t exist without the original work. Edit: Yes, it’s a subset of the script and not thee script itself, and being a pedant about that isn’t a useful point. It’s just pedantic.
Subs are not the script, they are a transcription of the finished work.
> Your consumer can query the API on its own, and download 5 subtitles per IP's per 24 hours, but a user must be authenticated to download more. Users will then be able to download as many subtitles as their ranks allows, from 10 as simple signed up user, to 1000 for VIP user. > If you wish to offer subtitles without prompting for an opensubtitles account, please contact us and tell us about your project... For an ea…
I haven't actually tried to use it. So long as they offer a dump you could use I think it's entirely reasonable to require payment for the API.
I think it's reasonable to consider them morally open even if the web UI is closed source if the subtitles are open. It feels to me like a project selling a shiny hosted app around an open source project. I think in this sort of project the core "source" is the subtitles, not the web UI.
If OpenAI let you download their models but didn't provide the tooling you'd need to run them I think their name would be much more reasonable.