Fun fact: Midjourney bans images of genocidal dictator Xi Jinping, but not other world leaders.
They don’t want to get hacked by the Chinese :)
They want to make as much money as they possibly can, so they take all steps to ensure they cannot hurt the feelings of Poo-Bear :)
It's fucking pathetic to watch a company choose to kowtow to a dictator for the hope of profit
Every company does everything possible to maximise profit
and most are pathetic for it
I feel like the cat is way out of the bag on this. Stable diffusion can be run locally. Hugging Face has over 15,000 models labelled with text to image generation.
I'm sure there will be (or already are) Trump and Biden Loras.
Dont even need them. Theyre such big names that theyre already baked into the base model.
If not the two of them are so public that it would take a bad actor no time at all to train one.
What do you need, 20 pics.
1 pic and an IPAdapter
Just looked that up, it's amazing stuff.
I guess it's a matter of increasing the friction to making one. This will slightly decrease the number of them?
Or they just don't want to be associated with it
But that's not necessary a better thing to do then. It might decrease awareness on the ease of generating fake images for the general public 🙁
This election cycle will be full with AI generated propaganda. Fun times ahead.
I hope they start posting nudes of Biden.
I want a steamey sex tape with both candidates.
Thanks I hate it.
Which one bottoming?
I’ve had family member ask me about AI since I’m younger and I tell them every time that they aren’t ready for it at all. Especially in this years elections, AI will be weaponized immediately. Like in a few months the machines will turn on from countries around the world to generate fake images and mountains of fake articles and videos and sound bites. Basically, the average Facebook user will be flooded with misinformation and fakery to the point of overload.
AI stuff is survival of the believable, not reality. So if you can believe it, even just a little, you will reinforce an AI model on the edge of your disbelief and it is designed to perfect delusion and extremism.
This year will end in tragedy. I guarantee it.
the average Facebook user will be flooded with misinformation and fakery to the point of overload.
If you put it like this, it does not sound that bad - since that is already the case. /not sure if s or not.
When I say misinformation and fakery, I mean that this information won’t be the normal type. It’ll be closer to fabrication or lies where the truth is not even partially represented. Same with fakery. It’s the difference between propaganda editing and straight up CGI or photoshop. The propaganda editing is common. Skew the narrative to the point of misinforming. The other type is not currently common. Because making a believable lie vs a parody of the truth takes more time. But AI can and will change that.
For a while now Facebook has been a place of delusion where a different reality exists for people. This will double or triple that disconnect and will likely drive the whole platform into full post-truth but then also delusional truth. I hope I’m wrong but I doubt it.
Take 1d4 psychic damage
Here is an alternative Piped link(s):
What about videos of Biden and Trump eating spaghetti?
Piped is a privacy-respecting open-source alternative frontend to YouTube.
I'm open-source; check me out at GitHub.
That's nice, but without a law and some kind of real enforcement, nothing is going to change.
That seems like a reasonable decision
All well and good for text prompts but Midjourney lets you upload an image and modify it, or smash two images together. Are their filters going to detect trump and biden’s faces in those images? Not sure AI is there yet.
100% AI is already there yet but I don't think it's worth the cost to run it on every gen for them
fucking photoshop exists. all of this is moot point
EDIT : stable diffusion too. if you want to make propagada, you'll make propaganda.
How do i get midjourney to do this?
When you do the /imagine command, paste in a URL after the prompt
And to smash?
When two urls love each other very much…
You can use the \blend command if you want to combine two images without any prompt.
If you put two image urls after the prompt then it will combine those images together.
Hmm, doesn't seem to work for me. Bummer.
It has to be a link to a direct image, the image can't be on a webpage or such. Generally that ends in .png (or another photo-related extension) for convention (but not strictly required, depends on how the server is configured).
You can also upload your own images to discord instead:
Too bad it'll still work for a world leader who looks kinda like his name rhymes with hidin' or disgraced cheeto
Oh no... if they didnt went public they might have stopped all the people who are generating in advance
This is a most excellent place for technology news and articles.
