• 0 Posts
  • 37 Comments
Joined 3 years ago
cake
Cake day: September 2nd, 2023

help-circle






  • It’s a hyperbole. It’s not unethical. But many people don’t like it/hate it.

    1. An image is just more expensive than text. Text is usually ~1 byte per character. An uncompressed image is ~3bytes per pixel. Of course, most images are compressed, and compression does a lot of heavy lifting. But still. Also, you don’t need to uncompress plain text. More amount of bytes means more storage expense, more network expense, and worse user experience due to latency.
    2. We have plenty of tools that work for text. You can copy-paste it. You can easily edit it with just a keyboard. It is easily configurable via fonts and font sizes.
    3. Accessibility: text can be read by screen readers. Images (without alt text) cannot. Maybe there are some fancy screen readers that OCR images, but then it’s the case of point 1.
    4. Text is easily indexable. Which means that it’s searchable. If the reddit search tool were any good, it could find the post. Not for images. Alternatively, 3rd party search tools such as google and DDG work.

    There’s probably many more points.




  • You went through my comment history and quoted me, to just not read the whole quote.

    Here, I’ll help you:

    It’s fine if someone already answered with what you were going to answer. You can just upvote that guy and move on.

    As I said, there are already 3 top comments explaining to you why you’re being downvoted. I don’t need to explain myself when I mostly agree with them, I just upvote them.

    If everyone had to explain every downvote, we would have hundreds of comments on each post, and most of them would say the same thing.



  • If you approve someone, you don’t need to explain yourself, you would just say “I agree with this guy”. There’s no substance to it.

    However, if you downvote, you are saying “this is wrong”. Which is much different. When you accuse someone of being wrong, you should explain yourself, otherwise you’re being a dick.

    It’s fine if someone already answered with what you were going to answer. You can just upvote that guy and move on.

    EDIT: I’m absolutely in awe that this comment specifically gathered so many downvotes. And this is a good example of what I was referring to. Lots of people downvoted, and 0 ppl said why.



  • Well yes, the LLMs are not the ones that actually generate the images. They basically act as a translator between the image generator and the human text input. Well, just the tokenizer probably. But that’s beside the point. Both LLMs and image generators are generative AI. And have similar mechanisms. They both can create never-before seen content by mixing things it has “seen”.

    I’m not claiming that they didn’t use CSAM to train their models. I’m just saying that’s this is not definitive proof of it.

    It’s like claiming that you’re a good mathematician because you can calculate 2+2. Good mathematicians can do that, but so can bad mathematicians.





  • Yes, privacy should be the default. However, there is a limit for what is practical.

    A VPN adds latency to your connection. It adds an additional failure point. If you don’t need it, why use it? Most people don’t need privacy at that level.

    Additionally VPNs are not free.

    For the ones that claim to be free, you have to remember that if you’re not paying, you are the product.

    What privacy do VPNs provide to the average user when not doing anything illegal? Absolutely zero. You might claim “but VPNs hide my traffic from my ISP!” And it is true, but in doing so you expose it to the VPN provider. In the end, unless you operate your own network, you will have to tell someone where you are going. Using a VPN is just kicking the can.

    Of course if your ISP is known for doing shady stuff and there is a VPN that you fully trust, it may be worth it.

    But I swear the VPN industry has wiped people’s minds with millions of ads so they’re not thinking anymore.


  • The wine thing could prove me wrong if someone could answer my question.

    But I don’t think my theory is that wild. LLMs can interpolate, and that is a fact. You can ask it to make a bear with duck hands and it will do it. I’ve seen images on the internet of things similar to that generated by LLMs.

    Who is to say interpolating nude children from regular children+nude adults is too wild?

    Furthermore, you don’t need CSAM for photos of nude children.

    Children are nude at beaches all the time, there probably are many photos on the internet where there are nude children in the background of beach photos. That would probably help the LLM.