Hacker News Reader

Comment by eigenvalue | Hacker News Reader

I’d be worried if I were MidJourney now. This seems just significantly better for all sorts of things other than pure art stuff. Anything that requires text or strict following of instructions is immeasurably better in Dalle3 than in MJ. And ChatGPT can take your vague, bad prompts and turn them into pretty good instructions for the model. I actually downgraded my MJ plan since it became clear to me that I’d be using it far less now. Hopefully they can come up with a response to it, but it’s a tall order to integrate an LLM the way OpenAI has done here.

View on HN

I'd be worried if I were Midjourney specifically because Discord is a terrible user interface. Trying out DALL-E 3 on a regular web interface ia so much better.

I'd try out MJ again if they had a regular website. I don't even like OpenAI as a company, but I can't stand using Discord like that.

This is also my big complaint with Midjourney, and the reason I’m going to cancel my subscription after testing DALL-E 3 through ChatGPT over the last week. The discord chat slash command interface is a horrible user experience, especially on mobile.

> I'd be worried if I were Midjourney specifically because Discord is a terrible user interface.

Yeah, Discord is just nasty for that. It got a little better when I found out how to create my own "server" (what?) so my stuff didn't get lost in a sea of other messages, but it's still not good.

I'll probably continue to put up with it until the moment when OpenAI images are of equivalent quality, but not after that.

Why do they even do it that way? It’s so annoying to scroll through the mass of images to get to a previous one you generated.

See the documentation on making your own "server" (as they call it, I don't know why... I'd call it something like a "channel"). That gives you a semi-private space where only the things you've done show up. Your images still show up in the main stream, I think, but you won't lose your stuff in a flood of other people's stuff.

Discord still kinda sucks, but not nearly as much.

Overall I prefer MidJourney's styles and feature set. But it's really, really hard to make MidJourney draw the things I want, especially when there's specific/detailed scenery I want to depict. The latter is now quite doable using DALL-E 3 even though the drawing itself may not be as good as MidJourney.

I recently generated images for a presentation. It took about 30 tries to generate 5 suitable images. But I burned 60 MidJourney generations and in the end none of the results were satisfactory. But because they were ugly but because they didn't properly depict the concept I wanted.

Now, if I can import a DALL-E 3 image into MidJourney and then use Zoom Out from there, that would be wonderful.

MidJourney seems to have a much wider range of style more easily accessible. I'm having trouble getting DALL-E 3 to produce less "plastic-y" figures.

What prompts are you using? Unsure about what exactly you mean, but seems it's capable of generating different styles pretty easily. Here is three (x4) examples: https://imgur.com/a/X0IbtIX

> Create four images made with as different styles as possible (eg plastic, metallic, organic and something else). Make the images about a PC desktop computer standing on a desk, and a screen that shows Hacker News frontpage

> Create four new images with more different styles

> Create four similar images with the style "out of this world"

“Photorealistic cat” or “photo of a cat”

They all look like characters from the WALL-E movie. None of them would be mistaken for an actual photo of a cat.

I'm guessing that it's giving you that kind of image because usually photograph captions don't use keywords like "photorealistic", it's mostly when you're doing illustrations, concept art and similar, that you use it to convey it's trying to imitate a photo. So it's creating images that look like those images, rather than actual photos.

Here is "Create four photographs of a cat": https://imgur.com/a/dRlCwqy

Iteration "Make the photos more realistic so they look like pictures taken with a real DLSR camera": https://imgur.com/a/WX0LEjD

Seems to produce OK results.

This one, to me, could be mistaken for not only an actual photo of a dog, but my own dog. (aside from the cyborg implant, which is a bit of a giveaway :) )

https://www.karmatics.com/stuff/cyborgdog.jpg

I didn't even bother saying "photorealistic", but I did give lighting hints:

"tricolor english shepherd future cyborg parts on head and front leg, very futuristic tech, replaces side of head and eye, black anodized aluminum, orange leds, big raised camera eye where normal eye would be, with colorful rich deep teal artificial iris, in worn future urban park with pond and stone bridge but pretty near dark with streetlight above beautiful dog and technology"

I don't tend to get plastic-y images. I think DALL-Es images are less plastic looking than either typical CG movies (even Pixar) or Midjourney.

Do you consider these plastic looking?

https://www.karmatics.com/stuff/dalle.html

One important thing is to give long prompts with lots of adjectives.

Reply on news.ycombinator.com