Raw LLM Responses

take 2 because I ACCIDENTALLY CLOSED THE TAB AND YOUTUBE DOESNT HAVE COMMENT AUTO SAVING 1. Nightshade doesn't really affect anything in the grand scheme of things due to the pure volume of images available. This is mentioned in the original Nightshade paper. Pouring a cup of cyanide into an ocean doesn't make the entire ocean toxic. (also, there isnt enough new, poisoned artwork when compared with new, non-poisoned artwork to make a noticeable effect with the improvement of AI image generators). That's to say, Nightshade isnt "fixed", but the effect isnt really noticeable at all in the grand scheme of things. 2.1 Based on your definition (If it doesn't have an artists concious or unconcious biases/experiences/choices inside of the art, it isn't art) would disqualify photography from being art. 2.2. If you commission someone to make a piece of art and tell them to make a "droid from star wars walking", you are going to get....... an image of a droid from star wars walking. Again, by your argument, this wouldn't be art. 2.3. If you are working based off of averages (aka what the AI is doing), a heavy majority of images with that description are of Simpsons characters. If you describe a Simpsons character to someone, without saying the name The Simpsons, they would still probably think of a Simpsons character. 2.4. Imagine you look at thousands of images of Homer Simpson for days on end, you could probably draw him from memory very well. This is whats happening with the AI. 2.5. The Dune screencap example is due to overtraining on the same image many times. If you were constantly shown the same image for days on end, you could probably remember it very well. 2.6. The trend was of a specific type of style, you cant really accredit a genre to a single person 2.7. Even then, AI style lookup tables + the prompts themselves frequently have the name of the artist in them. 3.1. This image is from the Explore page (which is only the first page you get when going to midjourney.com if you are signed in). I know this by comparing the UI to the Explore page, plus the first page you get when you are not signed in doesn't have ANY examples on it. 3.2. This image was not hand-picked by MJ staff. The top images on the Explore page are selected based on community metrics. IE likes and recency. 3.3. Looking at the videos (which isnt MJs strongest suit) on the Explore page (as I cant find a way to filter them, the UX on this website could be better) these seem like good high level references. IE they seem good for general posing as long as you arent literally tracing them detail for detail. 3.4. Even then, there are better models out there that dont have these issues. IE NovelAI V4.5 or GPT Image Generation 3.5. "but what use case is there for AI references?" If you are looking for a specific concept or a specific combination of concepts. Its perfectly fine if you personally don't want to, but going after people who do doesn't really make sense. 4. Just blindly insulting and name-calling the other side doesnt help in any argument, this is an issue both sides have that shouldn't be a thing, but this is the internet so people will be immature either way. This is to say I don't agree with people just calling you a luddite. 5.1 The argument is more for accessibility in a way that is unrelated to disabilities. IE in a time and monetary way. This is also seemingly what the commenter you were replying to was saying. 5.2. A heavy majority of the people in the developed world already have smartphones, as it is basically a neccesity in the current digital world. Therefore, this wouldnt be a cost they would have to account for to use AI. 5.3. The investment isn't only monetary, its also an investment of time to get good at art. If you don't have a lot of time, AI starts to seem very attractive. 5.4. I personally have dysgraphia, and (despite lots of effort and classes) cannot make drawings that look comprehensible. This is why AI is so attractive to me. 6.1. Even if every company that works on AI suddenly stopped (which most likely wont happen, but I'll get into that in a second) AI would still continue being worked on due to open source. Look at HuggingFace or CivitAI, there are competitive models there that are free to download and were made with open tools. 6.2. Companies 99% wont stop working on AI, as the rewards from being the first to create the fabled AGI is too great. Companies that are the size of Google have the capital to subsidize the losses from working on AI in an unprofitable way. 6.3. "Then why would they work on making image/video generation better instead of just conversational or agentic AI?" Investor perception. Good looking and realistic images look a lot better on an example sheet/presentation than some numbers. This allows companies to lessen the financial load of working on AI when its not profitable, and the longer they stall the higher chance of them making the breakthrough that creates AGI. 6.4. It's corporate advertising, what did you expect? Idk what Googles marketing team is on, but they are not doing well. If I had to guess, them advertising this use case is for 2 reasons. One is to appeal to boomers and older folks, who would find it funny, the other is because it is hard to quantify and explain other, actually useful use-cases for AI (coding assistants, prototyping, drafting, etc) in a short advertisement. 6.5. A heavy majority of algorithms are based on AI. Youtube's is, TikTok's is, Instagram's is, Spotify's implementation just isnt good. 6.6. The same things being said about AI now are also what people were saying about computers, smartphones, and the Internet. That they would never take off and are just another gimmick that will die out, and look where we are now. NFTs died because they had a niche use-case, and that use-case wasn't even what they were used for. 7.1. The comparison between searching on Google Images and AI isn't really a fair comparison, this is for 2 reasons: 1. These are 2 different tools for 2 different tasks. 2. You were directly copying the end prompt, excluding any revisions to the prompt, curation, or changes to settings. What you did is more similar to downloading someone elses art, then putting it into an art program. I am also very much willing to debate on this topic, so please do respond as this seems like an interesting debate to have