AI art is for grifters and thieves, either learn to pick up a pencil and draw or commission an actual artist.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
ChatGPT Now Lets You Generate Two Images Per Day For Free
- Thread starter MacRumors
- Start date
- Sort by reaction score
Hahahaha! I wouldn't exactly go trying to copywrite those suckas...
Still requires work and discipline to use PS and LR well--I watch my photographer wife use it every night.
Were there some philosophical discussions about the nature of art? Yes. Was it still a person using the software and hardware to create something only other trained artists can? Also yes.
The problem with AI in this fashion is it has no anima or repeatability. It is always (literally) derivative. Because nothing was sacrificed (time, effort, talent, etc.,), it has little to no value.
Now, where it is valuable is as a research assistant. It can create code, for instance. It can even create decent code in a vacuum. But then a trained coder has to verify the code. See how it fits into the overall structure of the other code in the organization, etc.,
Have you seen how much power the server farms full of GPUs required by LLM queries take to run? $20 a month probably barely covers their costs.
I didn’t mean that people aren’t having their artwork copied like for like, I’m saying artists take ideas from other people’s works when coming up with their own project.
They take that idea and create something new and unique using their talents, discipline, and time. They don't just feed a prompt to a derivative AI.
Hence why it can't be trademarked or copyrighted when the AI does it.
It's not even a search engine. Please don't use this trying to get actual information. It's a text aggregator.
That's what I mean by a tool, not the end all be all of digital art. It exists and functions at the push of a button but the results are often not very good. Like PS and LR, it requires skill to achieve a 'good' result.
It's much too early to try and guess at a use case for it, though. Who knows what the final form of these generative models will be?
It is useless to me, I wanted to create selfie to spoof stupid Facebook algorithm because they have blocked my account with reasoning “we don’t think it is you, upload a selfie”. Funniest thing is that I had no photos on this account at all, but back to “AI”.
I just asked GPT generator to make a selfie, described it must be grainy for photorealism. And they have me a picture of anime-styled man that is not even a selfie but taken with mirrorless camera🤦♂️
I said again that I need realistic selfie. It gives me a selfie with Lego skin, Lego hair and zero realism whatsoever.
Then I asked it “wait, are there some safeguards or styling that prevents from creating lifelike stuff?” And it said yes, “for ethical reasons”. And it said “sorry, you ran out of two photos. Come tomorrow!”
Like, what is even the point of that generator if it cannot generate anything? Default and built-in styling had always looked creepy and ugly.
If that is a “revolution” I don’t know what to say really
I just asked GPT generator to make a selfie, described it must be grainy for photorealism. And they have me a picture of anime-styled man that is not even a selfie but taken with mirrorless camera🤦♂️
I said again that I need realistic selfie. It gives me a selfie with Lego skin, Lego hair and zero realism whatsoever.
Then I asked it “wait, are there some safeguards or styling that prevents from creating lifelike stuff?” And it said yes, “for ethical reasons”. And it said “sorry, you ran out of two photos. Come tomorrow!”
Like, what is even the point of that generator if it cannot generate anything? Default and built-in styling had always looked creepy and ugly.
If that is a “revolution” I don’t know what to say really
Dear ChatGPT,
Please make one very large image, consisting of 100 smaller images made from the following list:
1. xxxxx
2. xxxxx
Please make one very large image, consisting of 100 smaller images made from the following list:
1. xxxxx
2. xxxxx