Live data from Hacker News

Chrome, Firefox extension that blocks NSFW images using TensorFlow JS

github.com

171–173 of 173 posts

Re: Chrome, Firefox extension that blocks NSFW images using TensorFlow JS

#171
post #128

Earlier quoted context omitted.

DDG is kind of fascinating. Ever since I switched to DDG as my main search engine, I feel like the internet is 90% porn. Although that's probably not far off, I'd prefer seeing relevant search results instead.

Far from that bad IMHO, but it has a weird tendency to surface NSFW results, yes.

Its not that bad with the regular search results. Sometimes porn makes it to the first result page, usually its on the second or later.

The real problem is when searching for images. I recently had to search for a lot of images and for basically any search term I entered DDG was able to find porn. Ultimately, this made me switch back to Google since constantly having to risk showing NSFW results on your workplace in an open plan office isn't exactly a great experience.

I'm now using Google at work, Ecosia on my mobile phone and DDG/Google on my PC/Mac.

Re: Chrome, Firefox extension that blocks NSFW images using TensorFlow JS

#172

I’ve jokingly / seriously been waiting for things like this. I’m the type to go overboard with filtering and mute lists on twitter, but images are harder to deal with. I did have a 60,000 line pastebin that was used in conjunction with a Chrome extension to block Wojack and Pepe memes on 4chan using md5 hashes, but something not rely on specific hashes is obviously superior. One day someone will release the ‘detect a…

I am very interested in this wojak/pepe filtering thing, I think it's finally started to decrease but there was a solid 5 year gap where you really couldn't escape those dumb pictures

I started doing the same with md5 hash filtering but it was very slow going, an AI that can detect low quality posts and filter them out is the holy grail I guess

Re: Chrome, Firefox extension that blocks NSFW images using TensorFlow JS

#173

Earlier quoted context omitted.

I would use this for sure. Certain phrases are great flags for me to know "I won't be interested in anything this person has to say."

Do you think the issue with hyperpartisanism is that people’s existing bubbles and filters just aren’t strong enough?

I don't know about all of that, but I do know I want to use twitter for tech and not politics.
Post reply on HN