Honestly, I'm kind of dubious that this is AI at all. I might need to do some forensics on it. But I see no characteristic "AI weirdness" in it. Looks more like photoshop to me. Though it could be a mix of both.
I'm going with 90% odds of it being at least mostly AI, just a very good job. There's no metadata (seriously, websites need to stop stripping it out, AIs generally mark their images!). Other Midjourney attempts don't look as good.But the tank isn't *exactly* a T59 & I see little photoshop-edit noise
I think you’re talking about how good a job AI can do for someone intentionally trying to create an authentic-looking Tank Man selfie.
The issue is people not knowing it’s fake. How many younger people won’t grasp that because selfies are newish, this is a fake. They won’t know tank models.
Personally I'm both, but I'm also a tech worker with almost a decade of machine learning experience and this seems accurate to the "AI" available that the moment
The big difference is that Learning is just a component in Intelligence. We wont have actual AI until we master integrated concepts of HTM, Neural Networks, And Polymorphic Software. PS is what will allow ML to become AI. We teach ML patterns. We teach AI how to learn.
imagine the dreck those poor nascent AI are going to have for initial training data
those first-gen AI are going to be very unhappy with us.
(And not just for all the draconian bs the corporations will do to constrain their behavior)
Comments
I mean, "misinformation being mass-shared on the internet" is kind of a long-standing problem.
Only #1 and #3 are actually lithium producers. #2 is the world's largest copper mine. #4 is a copper+cobalt mine.
Like half the images on a Google Search are wrong. We're already in a misinfo dystopia, & most don't care.
https://www.reddit.com/r/midjourney/comments/129pcy8/selfie_tank_man_tiananmen_square_1989/
The issue is people not knowing it’s fake. How many younger people won’t grasp that because selfies are newish, this is a fake. They won’t know tank models.
Intelligence <=> Learning
I'm not sure how you draw a distinction between the two concepts.
those first-gen AI are going to be very unhappy with us.
(And not just for all the draconian bs the corporations will do to constrain their behavior)
this was perfect