Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I like the way Jeff signed off the article, pointing out that whilst the video has been pulled for (allegedly) promoting copyright infringement, Youtube, via Gemini, is (allegedly) slurping the content of Jeff's videos for the purposes of training their AI models.

Seems ironic that their AI models are getting their detection of "Dangerous or Harmful Content" wrong. Maybe they just need to infringe more copyright in order to better detect copyright infringement?



> Youtube, via Gemini, is (allegedly) slurping the content of Jeff's videos for the purposes of training their AI models

If by "allegedly" you mean that google admitted it

> Google models may be trained on some YouTube content, but always in accordance with our agreement with YouTube creators (https://techcrunch.com/2024/05/14/google-veo-a-serious-swing...)

Where "agreement" likely means "you accepted some tos 15 years ago so shut up".

> the video has been pulled for (allegedly) promoting copyright infringement

the irony...


> Where "agreement" likely means "you accepted some tos 15 years ago so shut up".

I am not a content creator or business on yt but i am 99.9% certain as soon as you enter your business credentials to make money they pretty much are allowed to do as they please and change the terms without notice (to which you must agree). And because as pointed out into the article, yt is a monopoly in all but name you have to agree to it as there are no viable alternatives.


The hoovering of data for ads went about the same. They consume my data and told me it was for better ads - the most visible result is that I get ads for things I've already bought and it conflates searches made only in the spirit of understanding with desire. On the bright-side it's produced quite a few good jokes. "I googled Breitbart and I'm getting ads for testerone treatment and viagra!" [my wife, 2014]

The least these creeps could do if they're going to treat us like this is deliver the experience they say the evil justifies.


Gemini doesn't need his video as train data, google can just torrent any content and use it as training data, just like facebook.


Uh? Veo 3 is arguably the result of owning YouTube and tapping into its content. No need to torrent much if you store the largest amount of footage on Earth.


The Veo demos I saw all looked like Hollywood productions. Not like YouTube videos which are 99% garbage you wouldn't want to train off of.


Those Hollywood style videos are just more impressive, especially for people who will pay. Veo can produce any style or quality of video, it's just not impressive to demo a video that looks like one any run-of-the-mill YouTuber can make in their bedroom.


If you want to use content-ID on youtube, you need to upload a copy of your content to youtube itself.

Both legit and illicit, there's all types of videos on youtube.


You mean in the way Meta is being sued for, and bluntly are almost certainly going to lose and have to pay out lots and lots and lots of money for?




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: