What's the image you're trying to upload?
So I am not really seeing the hype about this being any less slop than AI gen has been to date. It has the same flaws, just pushing them into different types of local minima (eg. text still regularly fails etc).
Basically it's like lossy image compression - artifacts are always there - they're just perceptually better hidden with algorithm iterations.
But I digress.
I will admit that 'creative' part of this Astro Hot Bundle shot is neat(let's not discuss qualitatively) - note I didn't cheat by using a source image here.
It's also decidedly more copyright infringey than Dall-E ever was - which is ironic given how many safe-guards they 'try' to have
From the web or from the app?
I'm not saying it's not advancing - I think that we're pretty deep into PR hype in the increments at this point though - the discourse around it has been getting progressively worse, not better.
No argument about this - tech advancement has been on exponential curve for a long awhile now, improving like nothing we've ever seen in history before decade after decade.
No argument here either - and the most powerful usecases of AI are going largely unnoticed by the PR wheel. And in many cases the investor $.
Yep - that shoe will drop eventually.
It can be arbitrary about locking you down, sometimes won't even let you use a photo of a pet as a reference image because of privacy and content restrictions. And then start a new chat and it's a free for all.
Goldeneye running on the Frostbite engine.
A picture of myself. Would love one of my daughter in this style.
Same.
)
Is this a photo or AI?Skill issue
It can be a bit weird with prompts. Sometimes you just have to keep making attempts, wording it differently and it will just start generating it.
I was initially using the free version and it worked, but you have limited uses.
Give it something and follow up with 'recreate with Ghibli art style'
Cheers mate
I agree, though I would maybe cut your timeline in half.
As directed by J.J. Abrams?
I tried, but I was unable to make it work - it got super finicky with the "content restrictions" and eventually I got hit with a throttleIf someone has some spare time.
Could they prompt for Yoji Shinkawas’ art style.
Interested to see what A.I spits out.
I think OpenAI are tightening the grip on it while they figure how to handle the extra work load. I imagine they haven't seen this interest since the early days of dall-e or maybe the bing censorship dog
If someone has some spare time.
Could they prompt for Yoji Shinkawas’ art style.
Interested to see what A.I spits out.
It is really challenging to get it to replicate he's style. Remotely mentioning he's name along with various other Japanese video game character artists raises content policy warnings. The best I could do was get detailed descriptions of Shinkawa's art by giving it examples and then asking it to describe traits that can be replicated without breaking content policies. This is what is what I got it to produce usingI tried, but I was unable to make it work - it got super finicky with the "content restrictions" and eventually I got hit with a throttle
Bombolone
's avatar using the following:
It is really challenging to get it to replicate he's style. Remotely mentioning he's name along with various other Japanese video game character artists raises content policy warnings. The best I could do was get detailed descriptions of Shinkawa's art by giving it examples and then asking it to describe traits that can be replicated without breaking content policies. This is what is what I got it to produce using
"Create a version of this as highly stylized black-and-white ink illustration, created with dramatic, sweeping brush strokes, abstract splashes, and broken forms, emphasizing motion and energy over precision. The ink work is chaotic yet intentional, with the statue partially dissolving into bold shadows, blocks of negative space, and textural streaks. The lighting is stark, with high contrast and deep blacks. The background is minimal, consisting of abstract smudges and blocks that frame the characters without realism. The style blends traditional Japanese sumi-e brush art with modern graphic illustration and concept art flair."
I've been trying for hour or so, I've now hit a rate limit on the paid version and have to wait 12 minutes haha. I'll get it work one way or another.
I've been trying for hour or so, I've now hit a rate limit on the paid version and have to wait 12 minutes haha. I'll get it work one way or another.
My god this is so cool.Input image:
make it a LEGO video game
a moment later, he is defeated and blown into individual pieces
That's fantasticInput image:
make it a LEGO video game
a moment later, he is defeated and blown into individual pieces
