If I'm not mistaken, you created some graphical assets for your brand identity.
This headline is very clickbaity.
You already created a logo and a color palette, which is almost all work done.
On top of that, the visual assets are mediocre at most. This 'brand identity' is like 1 in a thousand. This only supports my opinion that midjourney isn't capable of real creative work.
Slightly OT: I still don't understand why did they choose Discord to be the interface of Midjourney. So weird. It's like the early ~2010s Battlefield games where you could only launch them through the web browser + a plugin. Even if it has to be an instant messaging platform then why Discord, why not Element/Matrix or anything else?
Personally I only Stable Diffusion since there is a standalone self hosted version with Krita. But of course you need a decent PC/Mac for that which is a huge difference.
The approach is interesting, my main take is that this type of design has become prevalent and quite standard: it felt to me like I had seen that identity on 5 different brands. With that in mind, is it really an identity if it's the same as everyone else?
Design is hard, so doing variations on a theme, or formulaic processes to get to a point where you say "I like it, it feels right, and it looks good" is a success. If you can repeat a process like the OP lays out, that's a strong foundation to build a website on.
The identity part comes from the divergence from baseline, and having a good foundation lets you experiment and "craft" easily. Nobody is gonna get a knockout web design in 30 minutes following a tutorial, but you might spend a week or three building on what you do in that first 30 minutes to achieve something unique and stellar.
What struck me is that it lacks a design vision, and so you cannot reach identity. Brand Identity isn't easy, it requires critical thinking about what/why/who.
I can see tough AI improves the baseline for the "do it yourself" people that have no design skills. But using a humanly (well-) made pre-designed template would yield a better result currently.
I am wondering if the output will improve if the user has a good understanding of what they are and how this should translate into visual language, or that the network would just be capable of discriminating between some design trends.
TLDR: This is not making a brand identity in midjourney, this is making a logo and color palette and getting Midjourney to create some similar images.
I was somewhat expecting this to be more of a creating a brand identity based on the company, but this is not what they did.
They took a bunch of existing "themes", but who says those are representative of the brand?
> In our case, I have already designed the logo and color palette.
So, the core brand identity had already been decided before midjourney. Then they just used the blend feature to get a bunch more images that are a mix of the 2.
This headline is very clickbaity.
You already created a logo and a color palette, which is almost all work done.
On top of that, the visual assets are mediocre at most. This 'brand identity' is like 1 in a thousand. This only supports my opinion that midjourney isn't capable of real creative work.
Personally I only Stable Diffusion since there is a standalone self hosted version with Krita. But of course you need a decent PC/Mac for that which is a huge difference.
Design is hard, so doing variations on a theme, or formulaic processes to get to a point where you say "I like it, it feels right, and it looks good" is a success. If you can repeat a process like the OP lays out, that's a strong foundation to build a website on.
The identity part comes from the divergence from baseline, and having a good foundation lets you experiment and "craft" easily. Nobody is gonna get a knockout web design in 30 minutes following a tutorial, but you might spend a week or three building on what you do in that first 30 minutes to achieve something unique and stellar.
Its like how everything was flat for a bit and now my phone icons are slowly drifting back to gradients
One trend people probably have noticed is that many brands over the last decade went from a script or serif font to a san-serif.
People also hate "corporate memphis" (https://en.wikipedia.org/wiki/Corporate_Memphis) which has been a trend for a while (probably on the way out but still used heavily).
There aren't a lot of unique brands out there, and if they have some kind of success they're not likely to be unique for long.
I can see tough AI improves the baseline for the "do it yourself" people that have no design skills. But using a humanly (well-) made pre-designed template would yield a better result currently.
I am wondering if the output will improve if the user has a good understanding of what they are and how this should translate into visual language, or that the network would just be capable of discriminating between some design trends.
I was somewhat expecting this to be more of a creating a brand identity based on the company, but this is not what they did.
They took a bunch of existing "themes", but who says those are representative of the brand?
> In our case, I have already designed the logo and color palette.
So, the core brand identity had already been decided before midjourney. Then they just used the blend feature to get a bunch more images that are a mix of the 2.