My Midjourney Journey, AI art as an inspiration machine | VLOG

[@richaellr]

Midjourney is now in open beta!

[@revdancatt]

Yup: discord.gg/midjourney

[@picosdrivethru]

Watching midjourney as a commercial artist I feel threatened, I feel inspired, I feel sad, I feel hope.

[@Corelila]

Exactly the same feelings here. I’m equal parts excited and like I want to throw up. But I do have to believe that this is something that is on the nerdy side and most people will never really understand it- and will lack the vocabulary to truly make use of it. These images need an actual artist to help them become what they could be.

[@judilynn9569]

As a graphic artist, I’m excited about MidJourney. Like Dan, I’ve learned how to take MY OWN previously painted images (both traditional and computer painted) and feed/blend them into MJ to get meshed, new results. I don’t use anyone else’s style or art. Just mine. And it’s a blast! I don’t know why other artists can’t understand the advantage. I really don’t. You can’t be replaced unless you’re lazy and closed to anything new. It’s a powerful tool for the artist. Play with and learn it. Apply it to your skill set.

[@DHM91]

Love that song. Great video! Mid-Journey is an incredible tool.

[@JackWebbMusicClub]

Dan, this is an exceptionally shot, and executed video! I wish you nothing but sucess. The idea of this tool will scare a lot of people, but goodness me is it smart! Can't wait to get me mits on it!

[@revdancatt]

Thank you, and yeah the future is going to get weird/interesting with this kind of tech, super interested to see where it goes.

[@judilynn9569]

Great song at the end! We ARE indispensable!

[@revdancatt]

Hahaha, I knew right from the start of the video that I was going to have to use that music, for ALL THE REASONS.

[@bensoumokil]

I’m new to mid journey and I fricken love it! Being able to just think of something and it being created within seconds helps a lot with some of my graphic design ideas.

[@revdancatt]

I hope you're having fun with it, and yeah it's a great way to get inspiration, kinda addictive in a way.

[@RonHicks]

When you showed how to stack the prompts up that's what I've been looking for. Thanks😀

[@revdancatt]

Glad it helped. Hope you've been having fun with Midjourney and getting better results out, you kind of get into the swing of it after a while.

[@andydragonfisher6900]

Loved this workflow and I’ve just downloaded Big Magic as my next audiobook based on the extract you shared - thanks so much 🙏

[@revdancatt]

Oh Big Magic is a lovely book, interested to find out what you make of it.

[@aris_roth]

Great introduction to midjourney and an even better moral standpoint. Really appreciate your video.

[@revdancatt]

Thank you so much, trying to keep things balanced 😁

[@JeffHanke]

Great video! I really love the picture with the two faces and clouds at 13:53 .

[@revdancatt]

Thank you so much 😀 and yeah, I was kinda blasting through the outputs so I could use them for the video. There were a few which I wish I could have spent more time exploring more results.

[@Bizarro69]

It feels like we’re ghosts of the cyberpunk future. Like we’re seeing and interacting with things that are in their infancy which 100 years from now will be ubiquitous. Such a strange feeling!

[@revdancatt]

Yeah, it's an odd one isn't it. At some point this will all become personalised and remixed for us as we move from room to room (street to street).

[@ExclemationMark]

Thanks for the book recommendations, and the great storytelling. Lovely! 🙏

[@revdancatt]

You're so welcome!

[@ronnykhalil]

fantastic deep dive. thanks for sharing!

[@revdancatt]

Thanks, looking forwards to it coming out of beta so we can show the interface a bit more.

[@bastiaan7179]

Thanks for another nice video! Talking about transforming ideas, that quote from Elizabeth Gilbert sounds like Plato's concept of 'the ideal'.

[@revdancatt]

Yup, I like Gilberts take on it, it's a nice framework to operate thought.

[@JanneWolterbeek]

Very inspiring, some friends of mine are worried it will kill art and creativity, you just proved them wrong and so I sent them this video.

[@revdancatt]

Thank you, if anything I think it'll help, many more ways to get inspired now.

[@peterplantec7911]

Nice Job, Dan. I signed up for the beta and wanted an idea of how it worked. You accomplished that with inspiration. Thank you.

[@revdancatt]

Are you in yet? And if so are you having fun with it? They keep changing/improving the algorithm which has made some thing better and others not as good. I'm looking forwards to when we can talk more publicly about it.

[@peterplantec7911]

@@revdancatt Not in yet. They haven’t responded at all, unless I missed an email.

[@revdancatt]

@@peterplantec7911 if you’re on Twitter or Instagram DM me 🙂

[@LASZREN]

I'm very happy i came across your video! It's very rare to find such beautiful and valuable content like this. Thank you so much for creating, I'll be sticking around to see your progress. Greetings from Argentina!!

[@revdancatt]

Hey, thank you, and best wishes to Argentina!

[@JonathanWendt]

Really loved the shared passages. Thanks, Dan!

[@markg890]

Came to your channel for the pen plotting, stayed for the explorations.

[@revdancatt]

Hah thanks. Switching back to pen plotting for the next video.

[@dinahfromkabalor]

Thanks for giving your thoughts on this and showing some of your process as you explore it. What’s missing for me here is who made that red robot image. I think the credit part of stealing like an artist is what’s getting lost and what is the core of many peoples discomfort (And is what Midjourney and its kin are actually incapable of doing.)

[@revdancatt]

Thank you. I think the red robot is a bit of a red herring. Currently artists will use Pinterest very much as a mood board, so that red robot would be on one of those, along with a whole bunch of other artwork. Many artists use mood boards like this, and then create their own original price, having studies and been inspired by the art collected on Pinterest. This is another tool for an artists to investigate an artwork to get ideas. I'm not suggesting they just just take the output from Midjourney and go "Here's the finished artwork", but rather use the results as a jumping off point for once again creating their original artwork. In both cases the artist should be referencing the sources of inspiration.

[@dinahfromkabalor]

@@revdancatt You are slowly convincing me, video by video, that things like Midjourney used literally in the middle of the journey to an original piece can be a benign tool. It does still pose a problem for all the usual capitalism reasons-particularly lack of consent (from artists whose work was used to train it) and beiing extractive (in that those artists without which it wouldn't work are never compensated even as Midjourney brings in money)-as well as probably damaging (and maybe gutting) the market for entry-level artists to get started by charging for "good enough" art. Color me wary, but less uniformly negative than I was. Thanks!

[@lynneholman891]

Great video - wish you could do voice overs on all youtube videos lol - so easy to listen to

[@revdancatt]

Thanks, I'm going to start with my own, and then move onto other people's videos 😁

[@Cytryz]

Quite a lovely video. Was enjoyable to watch and learned some stuff while im eagerly waiting to try midjourney. I got the beta access, i just haven’t had a chance to sit down and try it yet. So excited.

[@revdancatt]

Thank you... you're going to have good fun with midjourney, it can get quite addictive sometimes!

[@StevesMakerspace]

Wow, really cool stuff. Another well done video. Thanks!

[@revdancatt]

Glad you enjoyed it!

[@SuperCool898Gaming]

whoa using pinterest to find more images similar to the midjourney images is brilliant, I will need to start doing that

[@revdancatt]

Yeah it's kinda fun. "An AI made this, lets go see some similar things that humans made" - an interesting way of discovering art.

[@1970smm]

awesome video. Just curious, did you keep adding a prompt along with the images?

[@revdancatt]

Sometimes I do. So you could give it an image and they keywords "cyan magenta yellow cubist" to give it something else to work with.

[@LedZepElo]

Hi Dan. Great video! What commands do you use on MJ to mix two photos together? Thanks 😊

[@lejoueurcon]

Very curious about this as well!

[@revdancatt]

​ @Martin Paré @BunniBean you just throw in a url to the image you want to mix in, you can even do more than one at a time. So if you'd used the prompt... /imagine a spaceport in a brutalist style --seed 23 ...you would get some form of spaceport, always the same selection because we're using the --seed command. If you then threw in the url of an image, like this... /imagine a spaceport in a brutalist style --seed 23 picsum.photos/seed/12345678/500/500 --iw 1 That would tell it to also use the image from the url, the '--iw 1' is how much weight the systems to apply to the image you're passing in, so 0.5 is a "normal" amount, 2 is a lot. But, it isn't a style transfer system, although close. So it's going to understand the words you've written more. It's not that good at analysing the image you've passed in, although it'll give it a go. So the image part is more about adding "flavour" than anything. Hope this helps.

[@LedZepElo]

@@revdancatt That does help me a lot! Thank you so much for your detailed explanation!

[@andreasdietrich2064]

How do I get to this page with the cube painted by several artists?

[@josephstanski5180]

Wow! Well done with the creative process - continue please. :o)

[@revdancatt]

Will do!

[@el.rascador]

very good video. good energy

[@FelipeHoenen]

Did you build that cube and that crowd style libraries yourself? So fascinating to look at!

[@revdancatt]

Ah, I wish I did. No, that's part of Midjourney and all the imagery it's been trained on. It's very good at emulating different artists' styles.

[@sharonxhazan]

@@revdancatt where i can find this page?

[@RetropunkAI]

Great stuff thanks for sharing.

[@swhitham]

That would entertain me a lot, too.

[@revdancatt]

Yeah, it's good fun, the kids really like it.

[@theblowupdollsmusic]

Awesome work. What text descriptions did you give for the initial first round of red robot generation if you can remember? They looks really cool.

[@revdancatt]

Ah I can't remember, probably something around 'space' and '1950s'. I should go back and give it another whirl.

[@taraaliyeva1813]

Can you make a tutorial on how to add photos from PC and start aiding ideas

[@squibblets]

Really great overview of your process! Just got into Midjourney myself and I'm having loads of fun. May I ask, how did you feed midjourney the robot picture?

[@revdancatt]

You just throw in a url to the image you want to mix in, you can even do more than one at a time. One image... /imagine cymk picsum.photos/seed/12345678/500/500 --iw 1 ...two images... /imagine watercolour picsum.photos/seed/12345678/500/500 picsum.photos/seed/87654321/500/500 But, it isn't a style transfer system, although close. So it's going to understand the words you've written more. It's not that good at analysing the image you've passed in, although it'll give it a go. So the image part is more about adding "flavour" than anything. Hope this helps.

[@wetdreamstravel1865]

Hey Dan, How did you insert the robot image to have MidJourney make variations of it? I cant seem to find out how to do it. and then how do you merge it with another image you have already created

[@mrbennett]

I'd love to know these myself.

[@revdancatt]

When you write the prompt `/imagine something cool` you tag the URL you want it to deal with on the end, along with how much "weight" to want to apply to that image. So something like: `/imagine something cool some.website.com/image.jpg --iw 2` Midjourney is good with turning words into images, but it's less good at understanding what's in an image you feed it, so it won't do amazing jobs with your images, but it will try it's best.

[@revdancatt]

@@mrbennett Hey Emmanuel, see my other reply to the above comment, should help you out.

[@robfrost1]

Great video. A. What commands does one use to use some image as the starting point, and B. How many "in"s are there in "Pinterest"? ;)

[@revdancatt]

To use an image you just paste in the URL pointing to the image. Everything starts with a slash command, so it'd be `/imagine foo.bar/someimage.jpg`. And at least 2!

[@robfrost1]

@@revdancatt thananks

[@matteomeloni704]

Amazingly inspiring video!! What do you think would be the best way to transpose those ai-generated images into a physical format?

[@revdancatt]

Thanks. For straight up printing, the output isn't high enough resolution right from the system, although hopefully it will be soon. But things like Photoshop and specific scaling software is getting much better at scaling things without resorting to just stretching the pixels. So I think a work flow of, Midjourney 👉 Scaling software 👉 Sharpening 👉 Print does a pretty good job now.

[@heartshinemusic]

Wow, the crossbreeding with your own input as a starting point is a really powerful way to "direct" the AI. Really impressive. Thanks for this video!

[@revdancatt]

It's only going to get better and easier too, going to be an interesting few years.

[@heartshinemusic]

@@revdancatt Hey Dan, I wanted to add that (so far) you are the only one talking about (and showing) that you can feed picture to Midjourney to cross with the generated images. It would be very worthwhile to show in a bit more detail how this actually works. (Do the pictures have to be an URL on the internet, or can you feed a file from your own computer, for example.)

[@revdancatt]

I'm on it. When I made that first video we weren't allowed to show the UI, which makes showing how to feed them in tricky. I'm less worried about showing the UI so I think I'll make a short video on how to do it, as that's the question I've been asked the most.

[@heartshinemusic]

@@revdancatt That's really great to hear! Good luck with the video. Can't wait to see that!

[@romenpheonixx7005]

How the heck do I bring my work into mid journey? If I have it on my computer without a url? Even when I tried a url it gave a warning message that it couldn’t do it. I paid for the subscription to. Great work as well

[@revdancatt]

The images has to live somewhere on the internet for the cloud part of it to access. If you have the paid subscription then you can talk to Midjourney in discord vi DM, what I do is just upload the image I want to use in Discord in a DM to midjourney (or any discord chat/group/channel you want). Once it's uploaded in discord click on it and then hit the "Open original" link just under the image, that'll open up the image in a browser with a URL, which you can then use to pass into midjourney. The url will be something like: "cdn.discordapp.com/attachments/123456789.jpg" Hope that helps.

[@FilmFactry]

What paper do you print the original background? Thanks!

[@revdancatt]

Most of the time I print onto Hahnemühle German Etching paper, but here it's Ilford Cotton Textured 310gsm. Hope that helps.

[@FilmFactry]

@@revdancatt Thanks! Your work on IG and on the Tube really inspired me to get a plotter. I didn't even know you were into AI as well. I'm deep down the Rabbit hole learning Midjourney which is so addictive. And it is just in its infantry.

[@revdancatt]

Yeah, it's good run, and new stuff coming out all the time.

[@IkedaDogbo]

How do you tell MJ to use the image from the internet as a starting point?

[@revdancatt]

When you are entering the prompt you include the url to the image you want it to work with. "/imagine url.to.the/image.jpg in the style of [artist]" Worth noting though that the system is designed to turn text into images that it's already been trained on. So it's not very good at understanding new images it's never seen before. That type of thing is more of style-transfer AI systems, which this isn't. Hope that makes sense.

[@goilo888]

Patiently waiting to get admitted to the beta like a kid in July waiting for Christmas…. In the meantime can you tell me what is the maximum size output for images? The one thing that’s disappointed me with all the AI sites/app’s I’ve tried is that the final size is quite small.

[@revdancatt]

This is around 1,024 by 1,024, although you can go a little wider or taller. But they're changing it all the time and I think larger sizes are coming soon.

[@revdancatt]

They've just updated the max size to 1,664 x 1,664. They say they can now do 3 Million pixels output, so width * height < 3,000,000 seems to be the limit.

[@goilo888]

@@revdancatt Thanks for the info.

[@ScholarsonScreen]

hey Dan, Great educational and informative video. I have a question if you or anyone could help me out. I want to know how did you use your image references you collected from Pinterest and combine the two in midjourney. Thanks

[@revdancatt]

I'm going to be making a video on this soon. But the short is you need to grab the URL to the image, and then when you tell Midjourney to `/image` you add whatever words you want in as normal, and then paste in the image URL. Midjourney doesn't really understand what's in the image, so you may not get what you expect, but you can also play with the '--iw 2' flags (image weight), if you use high values `/imagine url.to.image.jpg --iw 5` it will trying and use the main elements/forms of the image over the words. If you use `/imagine url.to.image.jpg --iw 0.5` it'll use the words over the image. Hope that helps a bit, at least until I make the next video 😁

[@ScholarsonScreen]

@@revdancatt Thank you for the information. i will give it a try. When will we be execting this video. Thanks

[@revdancatt]

Maybe at the end of the week, if I get my act together 😁

[@ScholarsonScreen]

@@revdancatt can't wait. Thanks

[@ScholarsonScreen]

@@revdancattthank you for making the video. Good work as always.

[@hakimsetbel1774]

very interesting thank you

[@revdancatt]

Thank you for watching!

[@picosdrivethru]

what ai system would you recommend for processing my own images, to create variations of my own images? I would be very interested in one of those.

[@revdancatt]

I'm not sure. These ones are good at using word prompts to build images, but not so much at understanding and being able to work with images you upload. There are other systems that use style-transference which would most likely be better, but alas I'm not familiar with those.

[@picosdrivethru]

@@revdancatt ah thanks for the lead Dan.. yeah I just came across those, seems they could be what I'm look for!

[@walaazahran29]

how can i download this app?

[@revdancatt]

At the moment it's based on Discord, you can sign up for the beta here: www.midjourney.com/app/

[@diegoramosg]

I love the way you go about creating work! Midjourney sounds like an excellent tool to bounce ideas. Do you know of any other programs that work similarly? I’m currently waiting to get in the beta

[@revdancatt]

Dall-E is the one to look at, although I think it's still closed beta: openai.com/blog/dall-e/

[@FireballVFX]

It is such a potential for creative visual storytelling. Incredible tool! Cant wait to get my hand on it : ) Anybody know what is the chance to get invited and test its potential limits?

[@revdancatt]

If you're on twitter or instagram send me a DM (both @revdancatt) and I'll drop you an invite.

[@FireballVFX]

@@revdancatt wow, that is amazing! :) I really appreciate it 😍

[@revdancatt]

Happy to help 😁

[@LuckyClovers]

Can I have a hug? You rock

[@revdancatt]

Always happy to send over a hug 🫂

[@DarkForcesStudio]

"Pininterest"?

[@revdancatt]

Hahaha, I was hoping no-one would notice! There's only so many times to can record everything before just moving on without getting everything perfect. I feel like an old man "All these kids on their pininterest"

[@TheRCNut]

Dan, been following you for a while as a fellow studio artist here. I've been on the Midjourney waitlist for quite a while. Is there such a thing as a private invite you'd be willing to share? As always, thanks for the content!

[@revdancatt]

There is such a thing, unfortunately I've already sent out both of mine. They keep expanding the beta so hopefully you'll get in the next couple of waves.

[@amslu]

Signing a painting generated by AI does take some ego, doesn’t it? Why you modify the output by using some archaic methods like the pen plotter, when the AI can generate thousands of such iterations in the same time? Seems like the ‘human’ part of the creative process starts to become a limitation…

[@sergioaguis]

The AI doesn't have intention, a human does, so he can sign it because he was behind the intention and also intervened. A image printed will never be the same to "very old" and and "rudementary" techniques. All these costly and slowly techniques for making an image are very valuable to the right market, just like engraving and etching. The market doesn't care about your opinion thought. Nor your opinion to the intention of the artist.

[@LASZREN]

Don't forget art is all about limitations!