Can also you Dataturks do a similar thing, hooking up with Mturks/your own team..also allows overlapping selection and NER+Classification etc. Adding extra data and much more.
I get disgusted by thinking why does Netflix want people to not sleep. Why auto-play in 5 sec? Isn't this morally wrong, like dangling cocaine in front of a drug addict.
I agree that people have free will but designing products so that people waste days just watching TV is very very sad IMHO.
Well I do, that's part of why I stopped watching broadcasts altogether.
The autoplay-next-thing can be disabled in the settings though, although it'll still auto-play a trailer after 15s if you complete a movie or whole series, which I loathe since it's a real mood killer after movies (I very much enjoy chilling or discussing while credits roll).
> that's part of why I stopped watching broadcasts altogether.
For me it was not because of linear programming, but because the ads became ever more obnoxious and interrupting - first it was simple ad blocks. 5min, nice, take a pee. Then the stations increased the loudness of ad spots, well fuck your ears if you decided to not take a pee. Then stations started putting layers of ads above the content and the last fad before I decided to tv-exit was that stations PiP-ed the content right in the middle of action scenes or whatnot - at 25% of screen estate, and the rest 75% was ads. Also, quality of content that was not US cinema radically decreased - Germans who know the Galileo of pre-2000 know what I mean... it's all about the TV form of clickbait, aka "eyeball-baiting", now.
"with almost half quitting after the first week"...to me this sounds super critical...no amount of funding or a partnership with even god him/herself, can't make it work unless this is fixed.
BTW we had a question in general to anyone who can help us,
Does hosting such a dataset cause issues with SEO etc? Anything else we should be aware of?
Makes a lot of sense, actually its really difficult to get a large enough dataset for moderation tasks to make a decent inhouse model for a fair enough comparison.
Sure, we can try scraping that from pornhub etc but fee then the negative classes would be very domain specific, using stock images may not provide a good measure.
Also, its really weird to assign such a task to any of your employees, feels kinda strange :)
Yeah, it's definitely not a nice task but what's stopping someone (well, besides potential legal issues) from using these commercial APIs to create datasets programatically and training a cloned model from that?
I'm curious what the profit margins are on these APIs because I think they are way overpriced.
