Lexica is super consistent and accurate to the prompt and almost never makes "mistakes" like creating extra limbs. But im not feeling the plastic look of the images. also they seem like they ut the cfg scale to the max, aka no creativity haha. but the potential is huge! dreamlike is amazing. It can turn any prompt into a decent image. Highly recommended. instant art... never heard of it! if it 100% comfirmed its free? like no credits or monthly subs? thanks for the video!
Lexica is doing great work but everything is so smooth and vibrant I had to change the promts a lot to get a minimal vintage look with them. But still a good tool to have and is fast to, thank you, good information well put together!
You keep upping your game! Every video is spectacular. So educational. The only problem is, people are so sucked in the video they forget to like, share, or comment.
Thank You ! You are a treasure for everyone who , like me, can't afford paying for any applications ! It's so difficult times these days , so you are my Savior ! All the Best of the Best to You !😊
Thank You So Much Matt... You are totally blowing me away on how awesome your overviews, detailed observations and expertise on explaining everything in plain English. Finally, A One-Stop-Location for Everything Dealing in the AI Arena. I am so glad I found you and your free offers for we who have little to spare but the knowledge we are picking up from you, is like, if we won the lottery! You Are Truly GOD Sent and Greatly Appreciate All That You Do... Bravo ! Bravo !! Bravo !!!
Matt Thank you SOOO much. I have been having frustrating issues with my designer from Fivver. Because of you I sign up for the $10 Lexica. (that's all I need). I just made the cutest Easter Bunny design. It does cartoon vector images too and they come out so detailed. sometimes it gets it wrong ( three legs, or two heads) but i tweak it. Its fun to play with. thanks a bunch!
Great video Matt. Really handy to see these examples. One thing that comes to mind ... even putting the same prompt into the same system (site) can generate wildly different results. The only way to get the same results is to grab the seed and include that in the prompt. I have no idea, however, if a seed from one system will give the same or similar results in another system. Presumably if they are using the same (or very similar) model then including the seed in the prompt should result in relatively similar results across the sites/services.
Matt, thank you very much for your to the point explanations. It was through your channel that I really understood what all the AI stuff was about, how to use it, etc. So I thank the RU-vid algorithm for flushing you into my timeline. And what's especially fine - now that I understand quite a bit more, your videos are still a real inspiration. Thanks for that too. Please continue at this level.
Thanks Matt. I've just discovered you while looking for a midjourney alternative as I find MJ platform to be so busy and confusing. Well not MJ but discourse. The two questions I have are 1.what about commercial useage rights on these platforms and 2. how do you ensure that no one else uses the same prompts and your generated artwork for their own projects? If you are selling your generated artwork for eg: as a coloring book you don't want anyone else using the same images. I will of course research and attempt to find these answers myself but thought these questions (& your answers )may help someone else.
I just spent over an hour with midjourney and after making 25 images I was immediately told I would have to upgrade to continue. I find midjourney interface confusing and irritating, I know it will take me some hours to get use to it. I will seriously consider using one of these Ai art platforms, they seem easier to use and the art looks great.
Yes I'm also here because I find the midjourney platform through discourse immensely busy, messy and hard to navigate. These look like fabulous options.
Instant Art doesn't offer privacy but the worst part was that all the prompts I tried produced roadkill images. Even the humans looks mauled and mangled. I will keep trying I guess. Love your videos by the way.
Thank you for this amazing video, Matt. I appreciate the time and effort you put into creating and sharing valuable content. Your channel is a great source of inspiration and motivation for me. Thanks again!
Tried Lexica, looks to have a limited set of png's that it overlays to make an image together with some colour changes. A bit of fun for 30mins and making a wallpaper for a phone, above that not much use.
So much fun! Thanks! I'm a big David Bowie fan, so my prompt was: David Bowie in a jungle at night in Early Gothic style clothing in Baroque art style with a pet walrus. The image was gorgeous, but the face does not at all look like David Bowie's. I noticed, however, that in almost all of the previously generated example Bowie images, only one of them really looks like him. The rest don't come close enough. I wonder if Mid Journey would do a better job duplicating Bowie's face? I did try another image, but I think I gave it to many details because it didn't do very well. I said: David Bowie punk rock purple hairstyle young adult full body eating a sandwich very skinny on the roof of a skyscraper in Street Art style smiling wearing hippie clothing riding an elephant with a teddy bear. Lexica got the hair right, very skinny right, smiling right, and full body right. In one of the images, there was a teddy bear. Another image featured a giant sandwich, lol. Lexica, however, did not give me the skyscraper roof, the hippie clothing, the street art style, or the elephant. So, I might have to scale back my expectations a little bit. Unless the problem is that I phrased my prompt poorly. What I'd really like to see is Lexica be able to generate a greater resemblance to David Bowie's young adult face. Maybe I'll try Mid Journey to see how it does in this regard. Awesome, though! This is so exciting! I also upgraded to ChatGPT 4 a couple of days ago, but I'm not at all savvy with it. ChatGPT 3.5 did do a lovely job recently though when I asked it what life would be like for humanity in the absence of linear time. It also had a lot of lovely things to say about the soul and what a hypothetical afterlife would be like when I told it that it had obtained proof that the soul and the afterlife exist. I told it to go to the afterlife and then to return to the newly formed Afterlife Committee to report its findings. Eerie, though. ChatGPT 3.5's three main observations are *exactly* what I hope a hypothetical afterlife would be like, and I'm pretty sure my hopes and theories are not mainstream conventional ones. It's like the AI read my mind or something.
Can you advise which service (if any) I should use to easily recognise a new cartoon character, produce hundreds of images with different emotions and body positions and on different backgrounds? Midjourney can't seem to do it without it being humanoid characters
Best round-up of newest AI art tools I've come across. Clear, concise, and a comparison of the three to boot! Thanks; subbed!! How do you NOT have at least 100K subscribers?! Cup of coffee and off to binge into your videos.
Very helpful. Thank you. These tools are both exciting and worrying (for many reasons). I do also wonder how big our monthly subscription bills are going to be in the future, lol/cry. I hope some organizations operate on a donation model but we all have to throw them some $ if we find their tools to be useful. Same goes for open source orgs that are so helpful for the majority of us who cannot afford expensive tools. Hoping for the best in all of these arenas. Thanks for taking the time to make this less overwhelming to the rest of us. Subscibed.
Thanks for the video that was super useful. I have been experimenting with MidJourney and have been trying to find a use for it for work. Have you comes across a system where I could say upload an hand drawn storyboard but then i ask it to color it in? I have tried uploading a test image to my discord and linked it. But after trying a load of different instructions, the AI completely reimagining it into something crazy. This totally may be the wrong use case for Midjourney but just wanted to ask! Thanks
Mat you doing a great job I can't pay and in mid-journey i hate waiting and all that there interface , I hope now i can create good image for my POD store , Also can you tell me that can we sell these images on wirestock?
Instantart no model available error. And dreamlike also given 504 for me might be bad day for me 😅, will try later. But lexica is my favorite one. Thank you n ❤ from India
I keep getting 502 bad gateway error for instantart, I tried your link and tried multiple different google searches to get into the site but I can't. Great video though, I love learning about all these AI tools.
Lexica is good, I don't mind paying for it. I'll have a look at the others, thanks for bringing them to my attention, can never have enough AI art sites. look forward to when the same thing can be done with videos.
Awesome information, Matt. Thanks for showing this stuff to us! I have a question...Why do many of these AI art/image creators seem to have huge problems generating limbs/appendages on people and animals? Even eyes and noses are often crazy. I see so many people and images get generated that are flat out creepy looking, with 3 arms/legs, different eyes, messed up fingers/hands and toes/feet. It seems like Stable Diffusion is the biggest culprit, rather than Midjourney. Thanks!
MidJourney definitely does it too. Stable Diffusion is the underlying tech even for MidJourney. It has improved over time. The Protogen model that I used in my last example supposedly does much better with hands and faces.
Cool. Thanks for sharing, Matt. I just find it really weird that they can produce the oftentimes amazing images, but then get something so basic as appendages and facial features so messed up.
Thank you for these very helpful videos, Matt. I really like what you share about AI tools. These tools are very useful to everyone, really. I have subscribed to your channel. Thanks again! 😍
Ola @mreflow, thanks for the video, but I already made a video of Instantart, kkkkkk. And when I did none had a video in any place, the problem is I'm Brazilian and the video is in Portuguese and also my canal is very new has a month or so, and you are helping me to work in my channel, so thank you for that also. Also, I did 2 others videos about images generator that no other RU-vidr (I'm not a RU-vidr I'm just an old student) did until now, I discover it because of your advices and learning with you.
