My first reaction is always to insult people. You're welcome.
This is what I do viscerally, and afterward, after meditating on a somewhat more tempered reaction, I will give a reasonable answer.
AI art is moving at a speed that is currently mind-boggling. Lines are making more and more sense by the day, soon it might even be able to do a hand without 7 fingers but 5, and being an early adopter doesn't mean much anymore, because soon the entire world will use it through movies, video games, board games, you name it. You will all probably make your own damn movie.
While it's improving exponentially, it also means that the execution is improving quite drastically too. While I was reading @greenmask9 's latest Ai art post, I couldn't help but notice that he was using very short prompts with good results, while I've been very adamant about improving your prompt knowledge and refining the process by adding elements.
My first reaction was of course to say "you're lazy, that's why you're using short prompts", but after he gave me the best answer you could give, I realized that I was just being a pompous asshole, and that, maybe, he was right.
After what I've learned about AI art in the last 7 months, I can safely say that while learning about prompts is fun, and really helped me refine my style, it's important to remember that there are no rules, that's why it's exciting, and me wanting to impose "a certain way of using Ai art" is completely against the very idea of art, and by doing so I am sucking all the fun out of the process.
My point is, the size of your prompt doesn't matter, your prompt is your prompt, and the same way there are amazing long sentences in literature, there are also profound, short Haikus, with the same intensity.
Now, I'll show you some of the work I produced in the last few weeks.
Cheers, and prompt on!
Heaven's door
Simple theme, what would the door, the passageway to heaven look like? I linked it to energies, crystals, chakras, and then I wanted a melting mountain and played around with melting minerals.
Self Portrait
For the song Goodbye Yellow Brick Road, the cover of Elton John's song that I covered a few days ago, I thought of maybe releasing a version on Spotify, so I worked on a cover artwork. I used an image source, and created something from it, using the lyrics on the song "Goodbye Yellow Brick Road", very interesting stuffs...I also played around Elton John's themes...
Girl & Bear
I wanted a white polar bear and a young asian girl, ok it sounds a bit creepy now that I say it like that, but you get my point.
Starry night, mountains, lake, refections, and all.
Samurais
I had fun with this one, trying to really describe everything I wanted in my prompt, the Japanese blossoms, the intricate armor details, it turned out alright...
Cowboy & Dog
This one, people hated it on Instagram. Why so much hate I don't know, I liked it a lot.
A dog and his cowboy, and vice & versa...
Epic Battle
While I was working on a "Alex Jones vs Klaus Schwab satyre digital work", the AI went a certain route and I just went with the flow. Pretty cool, notice that one arm is missing.
Dream House
It has that Gladiator vibe...
Short Prompt
After all that discussion on prompts, I decided to put it to the test with a short one:
A nightsky that looks heavenlike, with a passageway leading to the origin of the universe, incredible reflexion on a lake, beautiful verdoyant nature
Conclusion: The conclusion is obvious, anything is possible, and let's chill out for a minute...
Next week post, will be about:
Conspiracy Theories!!!!
Great creations as always. I love the colors and vibe happening with the first batch of heavens door.
Thanks, they're my favorite too! Just a cool accident!
I appreciate you stopping by 🙏
Haha, this is so interesting, everything in this post 🙂
I think it's very helpful to learn more about different aspects of writing 'good' prompts, to try things out, to see what the effect is, and so on. It's like you'll have a bigger toolbox then. And then you can decide whether you like to use it or not.
But it's true, the most important thing, at least for me, is to have fun with this.
I love all of these. Especially the Girl & Bear ones 🙂
I love the way you presented it, I think likewise.
This is actually how Picasso described his process, he had to learn how to draw perfectly before being able to deconstruct bodies and faces through cubism.
It's the same here, you need to be pretty well versed in the prompt vocabulary, then you can choose what you want to keep.
Yes, it makes total sense!
This AI Art is so cool, I really need to get onboard the bandwagon and give it a go. I think I asked you before but what's the best free app to use?
Yeah dude no worry, it's called Midjourney the one that I use, but I use the paying subscription that allows me to commercialise what I make. Let's imagine I make a pattern for a blanket or a mouse pad with midjourney, I can sell it.
There are a plethora of other Ai art apps: Night Cafe studio (which is ok, not the best), stable diffusion (a bit harder to understand at first if you aren't tech savvy), Midjourney that you use through Discord, there is a new one called BlueWillow and it looks decent! You can still use Midjourney for free, but you won't be able to commercialise what you make, though I am sure if you post what you do here and make hive, nobody will say muffin hahah...
Thanks a million man for all the details. I'll def check it out.
Lol, you are so rude Ed. Good to let people know what to expect when they meet you, trying to remember how rude you were to me lol.
The art is pretty wild. That short prompt, seemed pretty long to me, but I haven't tried it yet, I now wonder how short you can go, to get a picture hmmmm.
Looking forward to the next series.
Hope you are all keeping warm. Much love xxxx
Hahaha I am the rudest you can be, but also very likable. Like a donkey...
To show you the difference in size, here is my most recent prompt about an MMA fighter that just retired yesterday.
It did this:
You can go as far as just one word.
Some Ai art software have a word limit, like Night cafe.
Wow, I don't think my prompts will ever be this long! Do you feel that the AI is able to take everything into consideration?
Lol, I was just messing around and testing. Sometimes I copy and paste other artists' prompts, and see what happens (shhhhh it's between you and I, and the rest of the blockchain).
Nope, I think it takes what it can, it's hard to know exactly how much.
Haha. I do use some of my old prompts sometimes just to see what happens. And I definitely 'borrow' stuff from other people. I guess that must be pretty common, haha.
No but seriously, it really helps me. Sometimes I don't even know what stuff in someone's prompt does, I mean what kind of effect it has. Then I look it up, learn something, and sometimes I test it myself 🙂
You sure have taken to this form of art and expression. Some of your creations are awesome and the skills you have obtained over the past 7 months are unbelievable.
Pretty soon you'll be making music videos and incorporating your music with your other new passion, AI art.
Great stuff, my friend.
I can't believe it's been since August that I've been doing this!
Funny story, a friend is proposing a marketing campaign for Nissan on Monday, and asked me to generate a few images yesterday for his proposal, I don't know if it will lead to anything, but I can't believe that people could be even remotely interested in something that I consider like a hobby!
I think there is a learning curve, but if I want to be completely honest, the tools are just getting better and better, very rapidly, so even someone that is just starting today could have great results...
It's interesting and inspiring what you said, yesterday I humored myself, and for my next music video, I wanted to see if I can use one of my own images (a picture from that video), and transform it with AI.
The whole video for the music video is roughly 6000 images (24 frames/s) so it will take me a while to create something, altering all the images.
For now the software that I have is a little bit too random in term of result, but maybe next month or so, with the next update, it will be possible.
And then the sky is the limit, interesting what you said the other day "cursed with talent", because my creativity goes in all directions, a bit like a film maker, I can't wait to be able to tell stories on screen!
Anyway rambling a bit here, thanks Beef!
Hobbies, if one spends enough time on them, you never know where they will go.
One of our boys speaks of the time it takes to become a master at whatever it is that you're into. I don't remember the scholar's name, but he talks about 10,000 hours being the magic number.
The tools just keep getting better, which is mind-blowing, I'm not sure where this AI stuff is headed but what can be created with it is already magical.
If you pull that off, altering 6,000 images may give you the 10,000 hours needed to become a master of this ever-growing form of artistic expression.
Good luck with it all, it sure seems to have tickled your fancy.
That's a good way to see things! I want to keep it fun for now, I saw what "working" in the music industry did with my perception of music, and maybe keeping it fresh is the only to still enjoy what I do in the future. Once you start putting the "I make a living with this" factor in front of something that you enjoy, the pressure becomes too much.
Yes, I agree with the 10 000 hours needed to become a master, totally true!
Nice work!
True, there is not a single right approach to getting awesome results. I tried your "short prompt" and made it shorter. The result is different but still beautiful.
As per my usual technique, I take some of the pictures I like and use them as part of the prompt. This is the result.
Normally, I would proceed by adjusting weights and other parameters or shortening the prompt further, but I got some Werewolves to worry about.
Cheers!
Oh I really like what you came up with! In general, using an image source in the prompt is pretty efficient, especially to fix issues like the number of fingers on a hand, etc...
Did you notice that we can't change the aspect ratio anymore on the V4 of Midjourney?
We can use 3:2 or 2:3 AR. Which is gross since I most of the time need 16:9, but it is still better than rectangles.
Hahah yeah, I am also waiting for Image weight to be available again.