the new image generator just generates the same image even when changing the description by 80% from the base! the pictures have ceased to be at least somehow deep, it is basically impossible to change the style!
funny, but the AI itself reacts to this and says that they have implemented a toothless comic generator! and the funniest thing is that it says don’t waste your time and use the new AI assistant and offers a choice with a full description of the advantages of competitors 
a huge minus to the development team for the degradation of the project
and I should also add here that generating even “poor” images now takes an unrealistic amount of time!
2 Likes
of course, if it helps promote and develop such an amazing project!
I thougt it will be improved and be a lot better after it launches as the DALL-E 3. But is it really get better since then? And now it is out and even replacing DALL-E 3 and my opinion on it is:
It is shaming terrible.
It is a huge step back.
Wo makes this terrible decissions?
Why everything they “improve” gets a lot worse?
Like the canvas, you have now to wait now until it is done coding for copying parts of it. I don’t like to wait.
I am mad for taking away my favourite image generator and replacing it with such a terrible thing. I prefer Image Generators of the Stoneage like SD 1.5 over it.
And the DALL-E legacy mode is no replacement, it can not generate in HD quality.
Look, this is the old DALL-E 3 HD. Beautiful.
Look, which one you think is the new generator?
What is your problem, Open AI, to make everything worse?
You just had to stay how you are and would be still the best.
Image generation now is not even close to usable. Especially the anime pictures look horrible. The worst of all picture generators I know.
To me it is terrible it can’t even do basic anime or manga styles anymore which is upsetting… I have to rely on bing image creator for the anime style I like.