aistinks.bsky.social
I help you not get fooled by AI slop. I also help artists be better. Because a world full of human made art is the future we’re all fighting for.
150 posts
10 followers
15 following
Regular Contributor
Active Commenter
comment in response to
post
The most interesting revelation in this vid is GenAI is very similar to crypto. It’s not particularly useful for normies, but it’s great for doing crimes. Spam, content farming, and reality distortion are its top functions.
comment in response to
post
AI concepts like this are especially pointless because it’s less “look what I made” and more, “look what a computer algorithm can do”. When is the novelty of a computer making an ugly, messed up image going to wear off?
comment in response to
post
Bro, this tweet comes from a satirical account who also says things like “they turned off the llm based biography translation on here for seemingly all accounts and now my funny prompt injection is worthless”
comment in response to
post
I looooove to shame people. To their face even. But I don’t like to make things up in order to do it.
comment in response to
post
All chatbots have a hallucination rate of 48-79% and they are getting worse. These things don’t know what a fact is. They’re just putting one likely word after another to sound like a sentence. I’ve personally seen it invent fake urls to its “sources”. It is a useless research tool.
comment in response to
post
Basically, ChatGPT is the perfect scam tool. We don’t trust human beings that lie to us 48% of the time. Why are we trusting bad simulacrum of human liars? 5/5
comment in response to
post
Where Nussbaum and I disagree is that ChatGPT can be useful for anything at all. With a 48% hallucination rate that’s getting worse not better, it’s not reliable enough for any use case except pretending to be a human being. It can’t give you advice or facts or even the current weather. 4/
comment in response to
post
This is a very apt conclusion. The bot will “lie” to you because it’s displaying only what it’s random word fill computes you want. Then it’ll apologize for lying because that’s now what it calculates you want. Like the NYT report on CGPT convincing a guy he was in the matrix, then saying sorry. 3/
comment in response to
post
Bots are not search engines. They don’t have up to date info. They also aren’t trained to do anything but keep you engaged and will tell you what you want to hear, not what you need to hear. It definitely doesn’t know how to read an x-ray. And it won’t know what’s currently going on at colleges. 2/
comment in response to
post
24FPS. Perfect anatomy, including hands. You’ve got to be rotoscoping a lot of this from real video. All the best did. Disney couldn’t animate this well in their prime.
comment in response to
post
Just to be clear for those not opening the thumbnails, the first image is the AI slop. The second two are @taramcphersonart.bsky.social actual work. I hope she sees this and joins a lawsuit because this is unambiguous.
comment in response to
post
All you’re doing is tone policing me because you can’t come up with an ethical answer to the very simple question: why AI and not artists. That’s a cop out.
comment in response to
post
Birdwatching requires a lot of skill. You need to be able to know where birds are and be able to ID them. You need encyclopedic knowledge of birds and their behavior and habitats. But I'm glad you're admitting you have no skills. What do you get out of paying genAI for images instead of artist?
comment in response to
post
For the sake of context, this is an account that says it loves "creativity" but seems to exist solely to undermine, demean, and debase actual human creativity. Your prompt is not creativity. A machine cannot be creative.
comment in response to
post
The only reason I can see for this slop is you're jealous of her talent and you purposely want to demean her.
comment in response to
post
Why would a fan make crappy forgeries of an artist they like and not even give credit? That's seriously monstrous. And why would McPherson continue to make art just to feed a machine that's not paying her and a fan base that has no respect for her?
comment in response to
post
How is generating images a hobby? What effort are you putting into it?
comment in response to
post
Midjourney’s logo is really shoddy. It’s hand drawn with a single width stroke. The bottom line of the front sail exceeds the width of the sail itself. The first wave crest isn’t round like the rest. The boat line crosses it. This all makes sense for a company that gens stuff for talentless users.
comment in response to
post
Shameshot. AI muddies the truth. It devalues art and human creation. It destroys the environment. All of these things are hard right, fascist goals. When you like and enjoy this stuff you are being fooled into supporting fascism. It doesn't matter how left you think you are.
comment in response to
post
Seriously, are you a bot, or is there a person behind this AI slop slinging account? And where are you located? Because you're not an American, that's for sure. Do you generate all of this trash, or do you have other sources? I'd love to know what the eventual endgame of a content farm like this is.
comment in response to
post
There are sooooo many of these left wing slop creators out there, it would be impossible for me to ID them all. They exist to get likes, followers, and QPs, but little else. And they work! Because people are idiots who like anything that makes fun of a politician they hate. Environment be damned!
comment in response to
post
Man, that account name with a dash between "will" and "reign", but not between "truth" and "will". Yeah, that's not LLM generated word salad at all. Interesting, that this bot stamps that name on some of its images. Most AI slop is authorless because the prompters don't even know how to insert text.
comment in response to
post
Because this short isn’t art. It’s just a bad simulation of it. It tries to use what someone who doesn’t understand art at all thinks the language of high brow cinema is. But there’s nothing here but flecks of fan cruft. And what is GenAI but a system for creating images with zero understanding? 8/8
comment in response to
post
I want to make clear, I don’t think much if any of this short was made with GenAi, but it still has all the hallmarks of it. Maybe because it was made by the same kind of people who appreciate what AI does, which is make empty nostalgic crap that looks like art, but isn’t. 7/
comment in response to
post
Finally, everything just looks plastic. Yes, this is some of the best human CG work out there, but CG is not perfect. And after almost 40 year trying to refine it, it may never be perfect. It hit a plateau a long time ago. Sounds like GenAI, which may never advance past what it is today. 6/
comment in response to
post
Next is the super shallow depth of field. I don’t know why this is so popular with GenAI. Maybe because in photography it means an expensive lens with a wide aperture? And the whole point of GenAI is to fake quality where there is none. 5/
comment in response to
post
Similarly, the characters, on top of being silent, don’t move much. They just look longingly most of the time. They gesture a little. Sometimes they actually walk. But most of the time they are as static as the landscape. 4/
comment in response to
post
The next similarity is the slow moving camera. AI vid can’t move quickly. If it does you’ll see terrible blurring streaks and the image falls apart. Like when you have image enhancement on your TV during a sports game. You have to have slow camera pans and zooms so everything stays stable. 3/
comment in response to
post
First, no one speaks. Which is weird because Sam Witwer, who “played” Kirk says there was voice work done. But Unification is completely silent. 2/
comment in response to
post
Here's the image in full, but not a shameshot, because the op didn't post it to propagate it as something real.
comment in response to
post
Weird choice of words. Last I checked Indians were “humans”.
comment in response to
post
NBC, do some journalism, for heaven’s sake, instead of printing PR releases. Palisades Research is run by ex Antropic employees. Of course they’re getting the same phony results. You’re getting snowed in the name of making this tech look more advanced than it actually is.