Midjourney still has the edge in quality, but it's a moot point if it takes you 1000 v-rolls to get to your original vision.
If all you're generating is anime waifus then MJ/NovelAI/Niji will suffice, but generating prompts particularly featuring relatively complex scenes or actions are amazing on DALL-E 3.
And of course unfortunately, it goes without saying that open AI DALL-E is going to be the most restrictive in terms of censorship.
I generated these from DALL-E 3 instantly. Try to generate them in any other commercial offering. Go ahead. I'll wait...
Descriptions:
A 80s photograph of the Koolaid Man breaking through the Berlin Wall.
Comic illustration set at a festive children's party. The main focus is on the magician who looks uncannily like a well-known fictional wizard. He's trying to say abracadabra but accidentally uses the killing curse.
For pure prompt coherence though I think ideogram is not far behind dalle 3.
Ideogram comparisons at bottom:
1. Generate initial draft image in DALL-E 3 (iterate as necessary)
It's essentially the ONLY good InstructPix2Pix model.
2. Bring into InvokeAI
Inpaint with stuff that might be considered censored in DALL-E 3.
I'd like to see some proof of Ideogram - it looks... very mobile/instagrammy from the landing page. If you have an account, try out my prompts I'd like to see what you're able to produce.
I can corroborate this. I wanted about 6 images for a presentation. I rolled ~300 MidJourney images. Most of them looked great, but none of them did what I wanted. I rolled ~50 DALL-E 3 images.
In the end, I only picked DALL-E 3 images. They were qualitatively not as good as MidJourney. For example when you zoom in then you see distortions. Or they're a bad fit for 16:9 format. But only DALL-E 3 was able to draw the things I wanted.
For interest's sake, this is from the second /imagine on MidJourney (so one of the second set of 4 images):
While yours is what you'd want, this arguably looks more like the super cheesy children's TV commercials back in the day and beats the ideogram take.
The Midjourney generations all appear to be referencing Halloween costumes or terrible cosplays, as if there are no trademarked koolaid men in their training set.
I remember hearing that the first versions of MJ used the LAION image set for training data - I'd be curious to see if it has any training data containing the Koolaid man.
I did a search through my MJ history from the past year and added the results to the imgur link to include my attempts at generating the Koolaid man from v3/v4/v5.2.
https://laion-aesthetic.datasette.io/laion-aesthetic-6pls/im...
And only three hundred for berlin wall:
https://laion-aesthetic.datasette.io/laion-aesthetic-6pls/im...
I am using these tools to send custom personal messages to close friends and family.
You can't even guarantee that you're getting the checkpoint that is known as GPT-4-Turbo via the API, and in fact you can almost guarantee you're not because now they do side-by-side tests between different versions of their models in the web UI.
The "new" model was being used months before they announced GPT-4 Turbo, you could similarly tell by the speed differential in some of the newer tool using modes.
Even the "original" is likely still being used in some responses in the web ui, you can tell because some of the side-by-sides have a drastic speed differential that matches GPT-4 vs GPT-4-Turbo.
-
You're trying to compare models using a tool that doesn't actually tell you what models you're comparing past which family they belong to, and represents more models than you're trying to discuss.
In other words you're talking about a topic you clearly don't understand and the equally clueless mob is glad to roll with it.
Good thing then that I was comparing the chat offering to 7b and 13b open models then, maybe you didn't catch that, try a third reading.
Also, this kind of is a chatgpt thread. I know it's not a perfect comparison, but still. It's a discussion point I'm presenting, not a research paper.
GPT-4 has an API, when comparing models you use the API otherwise you can't even name the checkpoint of the model that you're testing.
The web ui uses different models, unknown parameters, and even uses RAG... so it's not even an unfair comparision, it's not a comparison.
At the end of the day you're free to convince yourself whatever you want on whatever flawed process you want, but when you're corrected going full angsty sarcasm isn't going to change reality.
-
AI is in such a strange place now that the masses are interested in it: you can say whatever deluded thing you want as long as it aligns with what people want to hear.
I guess while people masturbate to this self-affirming nonsense the rest of us can build useful things on superior tech?
What you call "tone policing" I call "violating numerous HN guidelines":
> Be kind. Don't be snarky. Converse curiously; don't cross-examine. Edit out swipes.
> When disagreeing, please reply to the argument instead of calling names.
> Please don't sneer, including at the rest of the community.
> Please don't post shallow dismissals, especially of other people's work.
You may not personally agree with it, but there's nothing in the HN guidelines or comments from mods saying that I can't do it
> it adds precious little to the conversation
The conversations would have ended a long time ago if you're vitriolic tone was considered acceptable and normal
See dang's response to your comments: https://news.ycombinator.com/item?id=38493663
https://news.ycombinator.com/item?id=38490123
https://news.ycombinator.com/item?id=38484632
We have to ban accounts that post this way. It's not what this site is for, and destroys what it is for. Moreover, we've had warn you about this more than once over the years.
I don't want to ban you, so if you'd please review https://news.ycombinator.com/newsguidelines.html and stick to the rules from now on, we'd appreciate it. That means no swipes and no personal attacks, among other things.
Edit: you've unfortunately been doing this in other recent threads too:
https://news.ycombinator.com/item?id=38453843
https://news.ycombinator.com/item?id=38453381
https://news.ycombinator.com/item?id=38377068
We really are going to have to ban you if this keeps up, so if you'd please make whatever change is needed to not post like this again, that would be good.
If AWS wants to torpedo their business for marginal gain they can. And you won't be able to tell just because your workload falls on its face.
Your comment is the textbook definition of FUD in a comical way.
We do know of course that ChatGPT is most likely using 4-Turbo from the decrease in latency and increase in unhelpful answers.
We cannot say that the models are "converging down" though. I don't remember the marketing materials but from the model side we all realize that the Turbo models have some type of quantization/optimization that makes them cheap and fast. 4-Turbo is 3x cheaper than 4, substantially quicker and provides better results than 3.5-Turbo. Amazing progress in my arena.
(That said, perhaps I'm Clever Hands-ing myself by only using SDXL for what it's good at. It's terrible at dragons every time I've tried that…)
A year is a good timeframe to evaluate things: the rest of the world seems to lag behind OpenAI by around 12-18 months, at least with LLMs and image generation.
On the other hand open source tech usually has additional features for controlling output that OpenAI never bothers to implement, like llama.cpp’s grammars or ControlNet. So in that sense open source is usually ahead of OpenAI in terms of customizability.