Midjourney Beta – This is INSANE – The end of DALL-E 2 and Stable Diffusion?

Chapters:
0:00 Intro
0:34 Details
0:48 Art Styles
1:09 No Ratio
1:33 Less Artistic
2:03 More Artistic

The Midjourney BETA is here and it is incredibly good! You need to see there.
How to use the Midjourney Beta and get bette results
Super realistic skin, hair, details, perspective. Better than Dall-e 2 and Stable Diffusion?

https://www.instagram.com/olivio.ai/
https://www.instagram.com/robomar.ai.art/

Join this channel to get access to perks:
https://www.youtube.com/channel/UCCKx8mAHiFus-XYQLy_WnaA/join

My Facebook group: https://www.facebook.com/groups/oliviotutorials.superfan/
My Newsletter: https://oliviotutorials.podia.com/newsletter
My Affinity Photo Creative Packs: https://gumroad.com/sarikasat
My Patreon Page: https://www.patreon.com/sarikas
All my Social Media Accounts: https://linktr.ee/oliviotutorials

Midjourney, Olivio Sarikas August 23, 2022 · 100 Comments

100 Comments

  2. The reason Midjourney beta is out is because its built on top of the new stable diffusion that just came out. Stable diffusion is not going anywhere :p

    Reply

  14. I've just started using the beta and it blows me away. Seems like the characters it creates are ALOT better than previously. Shit is like digital crack, I goto sleep thinking about prompts, I dream about prompts, and I wake up screaming prompts. I need an intervention.

    Reply

  15. Adding the code of Stable Diffusion was a major upgrade to MJ, I'm actually able to get real looking alien animals to paint over, customize and improve now its not all visual noise, its amazing ! The results are so clear I think I'll even be able to use these for making sculptures. Just shows that working together is better than locking things off and hiding.
    That said, Stable is open source and will work on some GPU's so it'll definitely have its userbase, MJ costs me almost 40 bucks a month (yay taxes) and that is really expensive, you bet I will be trying either the collab or try to run it locally.

    Reply

  23. Loving being able to tag along with your MJ journey of discovery. Are we able to follow you on MJ, so we can see the images you are creating in your livestreams etc?

    Reply

  24. LOL the beta is already offline. One of the nice things about the beta was that it was uncensored which made it possible to produce traditional nude art, but they will be removing that feature. But we can make all the pictures of cats and dogs we like.

    Reply

  29. Its down already, but I can can't imagine they'll be long with the next version. I just have no idea how well they'll be able to incorporate elements of the beta into the main version, or whether they're compatible at all. It was maddening to have such a short time with it at this stage since it's such a game changer for anatomy and more literal interpretations.

    Reply

  34. Why would using Stable Diffusion destroy Stable Diffusion? This is why these kinds of videos should be left to people with a knowledge of the AI in question. So much misinformation, guys. Read up on it yourself and don't trust randomtubers.

    Reply

  36. You are not "creating"… you are lying to your self.. YOU are a lazy modern untalented "artist" who needs the help of a computer to create.

    What if there is no electricity or internet????… No "artist"? 😂😂😂😂

    If that doesn't depress you… you are not alive.

    Reply

  37. I can't believe they took it down so quickly, 24 hours is no where near enough for a "test" phase and for feedback, especially during the week when most users are working and don't have the extra time to play around with it. Poor decision on their part.

    Reply

  39. Even though it was temporary, I've fallen in love with using it, the amount of literacy it took, the accuracy, I only had enough free time to get a couple images, but holy smokes, they are incredible. Even the "unreal" sort of things like let's say godzilla. It creates something pretty damn good or on the nose. I love it and look forward to it's permanent release

    Reply

  40. Anyone have any tips getting a person from a photo approximated in MJ? It's almost impossible in normal MJ even with a high image weight and the beta didn't seem to use the image I gave it at all – it created a wonderful looking person but who looked nothing like the subject of my photo.

    Reply

  46. Midjourney does seem to have problems creating realistic looking teeth. Even in the picture of the woman you show at 0:21, you can see that the teeth are a bit off. It's a little frustrating, especially since everythin else looks so great.

    Reply

  49. There’s no “end” for stable diffusion, why? Because i can run it locally on my 3090 and not pay a dime to greedy developers like dalle and mj. Plus i can get nearly anything i want with prompt engineering already.

    Reply

  50. Dalle 2 maybe…stable diffusion no way in hell… Reason being SD is open-sourcing and also doesn't police what you're trying to create…(on your own computer run build)

    Reply

  56. One of the things that I've become super-fascinated by is that none of these ai image generators know how to create realistic looking fingers. In all of the images that I have seen created, there are always fingers missing, or they are the wrong size, fingers in the wrong places – its easy to miss, but once you see what I mean, you can't unsee it. That's why I had to stop at 31 seconds to take a closer look at the old man's fingers. Yep, you miss it at first, but the more closely you look, the more strange those fingers become. Its what I love about this new world of ai generated images – the struggle it has to understand our world from studying nothing but the history of of the images that we create. Such a strange and brilliant intelligence to get as much right as it does!

    Reply

  58. Stop using these apps that STEALS and DEVALUES illustrations and art from other peoples work!Artists will lose their jobs!!!!!Please support us!!Its immoral and unethical!!

    Reply

  60. I think so too Midjourney Beta V4 is really good. And if they maintain there current cost to use the AI think Midjourney just made a move to take #1 spot. I will say that some of my V3 prompts are just hands down better. But only certain things. If you ever do like ornate eggs etc. Like dragon eggs compare V3 and V4 it's a big difference. I think V3 is a lot better so far. However people are WAY better in V4. And I'm sure most of this is a difference in actual taste and also getting right prompts. But they are keeping V3 around so no problem regardless. I want V4 to be fully released and lets hope the final version puts all 4 variants back. I'm sure it will. Only getting 1-2 images is not fun.

    However we all know that things are changing to quickly and still great things Dalle 2 and Stable Diffusion can do that Midjourney can't and difference in personal taste too. But for now Midjourney is absolutely #1 to me. Dalle 2 is due for an upgrade actually. Or are they waiting for Dalle 3?

    Reply

  70. I feel like I just came from another planet. I'm not understanding one thing on this video or comment section. I'm so intrigued now 👍

    Reply

  71. The problem is that if you look at the details, it may look like what you asked at a first glance but most of the time you do not have real details. For instance, i simply asked for " a knight facing the entrance of of a dark cave, fullmoon scene, paper drawing style" the overall aspect was really beautifull however not even a piece of armor was a piece of armor actually, not even any symetry existed, the sword was floating beside him like if the ia could not figure out were a sword could be attached….furthermore suppose that you like a character that it has designed for you… No way you will be able to reproduce it in another scene… For now, i hope it will change soon. What it miss is the possibility to memorise a result as a named concepts to be able to reuse it in another scene and many concepts have a great lack of subdetails like my armor which could not be more accurate than the place and tje volume it is supposed to occupy on a body and the metal color. To be able to feed the ia with particular concepts would be great.

    Reply

  76. It’s funny how some people take credit for the outputs when the credit is on the engine. Just go ahead and try your fantastic “work” on DALLE mini… lol

    Reply

  77. It is basically just catching up to stable diffusion.. lol since the beta is actualy just them implimenting SD's code… Its more likely the end of MJ and Dalle if anything…. since why use them when you can jump on SD essentially for free

    Reply

  79. It's pretty and all, but, I think the images it produces are a bit…shallow? They're very "instragramable". Really extreme contrast in every image, really REALLY saturated colors….I feel like it's being trained on images that have a pretty homogenous, "internet beautiful" aesthetic. The kind of art that impresses people who don't really know anything about art, and just want things that are "pretty". Everything has to be hyper rendered, or have extreme contrast, or be surreal and dream like and "epic".

    I don't think artists should be worried about AI art quite so soon. There is a lack of restraint in these images that just makes them feel soulless. It's like these images are simply trying to impress me with extreme everything.

    It's all very cool technology don't get me wrong, it's extremely impressive. But I'm a little disturbed that so many people seem to think everything these AIs produce is just the most beautiful, most astounding thing ever. This stuff is like the Thomas Kinkade of 21st century art. It's all kind of… gaudy.

    Reply

  80. When I try the beta I get this error message "Invalid parameter
    Unrecognized argument(s): –beta" Anyone else have this problem? solution?

    Reply

  86. Yeah it's great when it works. It's down 70% of the time and instead of fixing the issue they invite more people into the beta which then strains the service even more which makes it go down more. I was a paid member at MidJourney then AFTER paying they reduced sample rate, reduced upscaling and moved me from fast mode (paid mode) to regular mode. I jumoed to Stable DIffusion and host it on my own pc now.

    Reply

  87. I will pass on all this AI art stuff. It just really takes the creativity out of creativity. No real skill involved. This is what people who don't like digital art think how digital art is made all the time. I hate that it just devalues image making in general.

    Reply

  89. Not as great as you'd think. It didn't know who Blackbeard was. Doesn't come up with what you describe a lot of the time. Can't create a basic button for UI. It's not bad at concept art and if you need free stuff but it's still got a long way to go. Also not sure if it "learning" from actual artists by stealing their work or characters and changing them around is ethical or legal to use commercially. Time will tell. But for 120 dollars per year you could buy some stock graphics for your project and feel good knowing you're actually supporting artists directly.

    Update: played around a bit more with it and it depends on how you ask. Still if you are looking for detailed commerical work, still best to go with a real artist who is actually drawing the fine details or why not pay them to fix an AI image and donate some of your profit to an art school to balance things out. PS: putting "in the style" of certain well known artists and using that for commerical work is very unethical. But asking it to do a hyper real or sketch isn't as much because at least the pixels are drawing something more original at that point.

    Reply

  90. This software is amazing. So many artists on Twitter are keen to shove their radical left wing views down people’s throats … well, now you go what you wanted – making beautiful art is now available to anyone How’s that for a socialist concept in its purest form? 😂

    Reply

  91. As a lifelong artist and sculptor fully aware that art is subjective, this AI software I see alot of visually beautiful art and imagery but with no depth and no meaning outside of visuals…Purely aesthetic beauty…I aim to use AI as a tool to speak out against the bills of our world and get people thinking (or even make them feel a bit uncomfortable whether by taking a shot a religious dogma or mobile phones at a restaurant table)…I still feel the primary work of the artist is to expand minds through meaning and depth of human experience…Not just to paint a hollow pretty picture…Which is the greater usage of this AI tool…

    Good upload

    Reply

  95. I was using the beta version which is the standard 2D rendition. It's ok. I didn't get to experience the 3D or 4K version that seems to be what's promoted. For that you need to register as a paying member.

    Reply

  99. The further it developes the more do I believe that people are to naive about it's potential and effects on society. Right now, Ai is treated by many more like a "toy" or just another "tool" to use that gives you very quick and cool results. But when you look actually at what it does and the discussions happening right now in the creative industry? This will put a hell of a lot of people out of work. Low level entry jobs will become non existant. Business will drastically lower their creative teams. Making an income as a creative person was already very difficult before. But now? I fear it will be impossible for most. And some might argue that Ai opens up the possibility for "non-artists" to be creative as well but what it will lead to is oversaturation. Once everyone can create a work of art on the level of the most skilled artists out there with a few seconds maybe minutes at most? Then it's meaningless. If everythings a masterpiece nothing is a masterpiece, so to speak. It's like watching a game where no one can lose. There is a reason peoople don't really find it very interesting to watch two chess computers playing against each other. And with Ai there will at some point be the issue that you will see less and less art created by humans which means the Ai will take art created by "Ais" to learn and process. The effects of that are not even really clear yet beacuse Ai has a very different learning-process compared to humans, Ai does not place any value on what it learns, has no experience it doesn't give meaning to anything it looks at because Ai has no concept of it, it only mimics ideas and concepts the forms and aesthetics that humans put value into. This has advantages. But also very serious disadvantages. To an Ai a dragon has as much or little value like a stool or a person.

    I feel people are way to excited right now to think about the real impact here that Ai will have on society as a whole. And it's developing fast. Incredibly fast. Right now, it's the creative business. Because this is the easiest one to push Ai in, a low risk, high reward situation. But you have other areas where you have a lot more research and investements being thrown in. From military applications to health, business, pretty much any blue collar job, even high paying ones(!) will be effected. Some experts believe that every second job can be one way or another affected by the automation trough Ai. And there is a potential to remove 30 or even up to 40% of the current work force. And it is very unlikely that we will see a high number of new jobs here to replace it. The target of Ai is to make things more cost effective, efficient and productive.

    To Ai humans will be what cars have been to horses. They didn't give horses new "cushy" jobs. They pretty much replaced them entirely. In the 1900 horses have been the main means of transportation in cities and rural areas. But today? They are pretty much gone. Because they are obsolete. And this will be true for a large number of jobs out there as well. And they will not be replaced equally by new profissions. There simply won't be enough of it to go around. And with the way our society and economy works right now? It will hit the job market like a nuclear bomb. And we're not ready for this. If you are very excited about Art Ai right now? Think about your own job. And what an Ai specifically designed for your profession with the same level of evolution.

    Reply

  100. ….and in the not too distant future this will end up being fully animation capable !!!
    Then, any 5 year old without modeling, texturing, choreography, and editing skills will be able to coax this AI into creating magnificent films and animations. The final step in commoditising visuals and film. The future looks brighter than ever….

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *

© 2024 AI Art Video Tutorials