Leonardo AI Changes Everything - So much more than MidJourney

[@max-imize]

This new tool has me SHOCKED!!! It is SO much fun. The combined desert image i made in this video is probably in my top 5 AI experiences of last month! If you want to start in Leonardo AI follow this video: kzbin.info/www/bejne/g4jQfJ-DmKpgbsk

[@henkvanholten3305]

Very Nice your video's Im working for 40 years now with ICT but i have the feeling we are on to something that will change everything we know, and this is just the beginning, but we must not be afraid of it

[@henkvanholten3305]

I am currently taking a Photoshop course, but with the new AI capabilities that are now available, Photoshop suddenly seems outdated. While you can still do remarkable things with Photoshop, the new AI capabilities are truly exceptional and seem otherworldly. I hope that the major course providers will follow this trend and offer new courses that focus on the future.

[@winmartinezk1]

Is it effective on character consistency? And creating character with your own image, i use midjourney i can say it need a lot to go in that direction

[@HansJoachimNolte]

There are plenty of people outside talking about AI Changes. But what i like much when you send your videos: i can feel your emotions. you are HAPPY and one can SEE and FEEL it. that was the reason i "abonniert" your channel. very thankful from my side. keep up doing your work!

[@max-imize]

Thanks so much for the kind words! Really means a lot. Vielen dank! And i hope you enjoy the ai future that is coming the next months!

[@sullivan3004]

The images do seem to be better quality then DALLE but as people stated they did have this technique "Outpainting" since a while back. As for "Inpainting" i think Playground AI got that some weeks back and would probably recommend them before Leonardo as they have a whole suit of AI tools with collaboration baked into the same interface.

[@JoopLiefaard]

Wonderful. Next time I open Leonardo I'll try.

[@davemorgan6013]

Yes, the Canvas tool is probably my favorite feature of Leonardo.

[@max-imize]

Yeah same! I feel like i just started out with ai art again :)

[@adamquek1095]

Great demo. Thanks!

[@max-imize]

Thank you! It is honestly so much fun. And can't put it down 😂

[@kirahokuten357]

Is there anyway if I upload a masked subject and use leonardo AI to add background(including blending stuff) on it?

[@max-imize]

What you can try is upload the image to leonardo, mask the background next to the subject and write a prompt with what it should fill it.

[@kirahokuten357]

@@max-imize I see, okay I will try that. Thank you :D

[@yoyoyo441]

Started getting into ai art! Thanks so much for these vids

[@max-imize]

Glad to hear it! Have fun :)

[@unitedstarsutopia]

Although another alternative if Leonardo decides to lock the canvas feature in premium if you are doing it free (also works better for humanoid character painting) in future you can do inpainting by downloading the picture, then extend the canvas background in photoshop, then upload it for inpainting and mask the blank space then add the prompt by telling ai to add features to it that you want, then it will extend respectfully, then upscale it for cut -out for photoshop and merge it with the original carefully. Then we've put the canvas future crisis to bed if it ever arises.

[@OolTube02]

If you create a PNG file with transparent parts in Photoshop then the outpainting tool considers everything that's transparent masked and meant to be inpainted. That way you can choose which parts the tool is supposed to replace with much greater accuracy than hand-painting with the mask tool.

[@unitedstarsutopia]

@@OolTube02 that's even better

[@dreamagination-infinite]

Thanks for the great demo. This is SO cool! I'm nerding out here! 😄🎉

[@max-imize]

Haha glad you like it as well! Honestly so hard to stop with this ai canvas

[@dreamagination-infinite]

​@@max-imize Yup, 😊While during regular prompt to image generation I am looking twice at how many tokens it's going to be, during the canvas runs I more like "Meh, just do it. Have to. Can't. Stop. Now!" 😂 Just spent 16 tokens in probably less than 5 minutes! 😅 I noticed that while it does well on the x-axis it has difficulty on the y-axis. Like generating new portions of the sky. I tried different things and each time it would give me another extra image on top, it wouldn't fit at all. Different color different everything. 🤔Well, this stuff is still new. But as fast as things move it'll probably be by miles better at the end of the week.

[@max-imize]

@@dreamagination-infinite yeaah it will get better faster! You could always try masking a tiny bit of the sky to see how it changes

[@jenstornell]

What if you use a photo or something that does not have a prompt. Would that work? Adding a custom prompt for it.

[@max-imize]

If you add a prompt that would create something similar as the photo it could work!

[@axl_ai_music]

It's very well integrated but I don't see anything we didn't have before with automatic1111 and outpaint/inpaint or with other Stable Diffusion UIs

[@RasmusSchultz]

Yeah, I don't recall the name of that other SD interface, but this feature definitely isn't new. What was it called...? 🤔

[@RasmusSchultz]

Ah yes, here it is kzbin.info/www/bejne/nnq8c5l6qbCUq9U This feature looks practically identical to that

[@RasmusSchultz]

Invoke AI was MIT licensed, so they probably just stole the code 🤫

[@jjfiasson9167]

@@RasmusSchultz Leo founded here - we didn’t use Invoke - we built the canvas from scratch 🤓

[@kcoziar]

Leonardo is better in every way other than the art it produces

[@max-imize]

Exactly! Soon it will get there with the nice models they have

[@Bartetmedia]

I agree, the interface is good but the images are not that great.

[@aicnednetni]

Does that include hands?

[@armondtanz]

Like going to buy the best god damn oven on the market... its got speed, acurate pricise dials and many settings.... The only drawback is it burns everything :(

[@travelwell6049]

I really like images shown in the video. I don’t understand what you think is good if not the finished pictures?

[@boomieboo]

This is incredible. If you think about it this can be the basis for doing the same in animation. From one illustration AI fills in the blanks to get to the next. No doubt in my mind AI will become huge for animators.

[@a-51labsstudio90]

Except for the in-betweeners.

[@gameswownow7213]

do you receive a certain amount of credit that renews montly or?

[@max-imize]

Everyday 250 new credits! So quite a lot

[@gameswownow7213]

@@max-imize thank you man

[@freehawk78]

I’m wondering if/when they will give the public API

[@max-imize]

Most likely within some weeks!

[@KrisztinaPeterR]

Hi, Anyone knows how to turn real image to an AI prompt? I would like to use the AI canvas, extend a real image which I do not have prompt. Is there any converter?

[@Eren_Yeager_is_the_GOAT]

i think describing the image yourself is the best thing you can do

[@Mystefier]

Very cool. I realized I wasn't able to do the things you showed, but found out it's because you can't get the extending the image ability using "brave". It works in "chrome" however.

[@max-imize]

Ah i didn't know it doesnt work in brave. Thanks for the info!

[@sorartificial]

Leonardo is amazing, I have been waiting for the email from them after submitting the form

[@max-imize]

They have a bit of a brake on new users atm to handle the amount of usage :( hopefully soon they allow more people in!

[@sorartificial]

@@max-imize I hope so.

[@ChibiKeruchan]

does it works with photos?

[@stoparchitectsoffear]

Is there an ai art to create graphic novels in.

[@antonetta7704]

This is exactly what I'm breaking, I tried photoshop yesterday and content aware fill in photoshop, it's awful. Do I have to wait to get access to Leonado AI?

[@max-imize]

It's pretty fast to get access to Leonardo! I explain how to do it in two videos ago:)

[@antonetta7704]

@@max-imize Thank you very much, I signed up yesterday, but I'm still waiting

[@GHowd7]

Can you use it on computer or laptop only ?

[@max-imize]

Its all webbased do computer is fine

[@GHowd7]

@@max-imize mobile phone to ?

[@GabrielVeda]

I love your enthusiasm.

[@max-imize]

Thanks :)

[@benzyn]

Another brilliant video bro, thank you♥🤩

[@max-imize]

Thanks so much 🙌

[@shammikhan4064]

Hey, can I create a video there? I have seen some amazing videos that were created by AI, anyone know where I can generate nature videos?

[@bastianrivero]

Is it 250 tokens per day or per month?

[@max-imize]

Per day! So really quite a lot :)

[@OolTube02]

@@max-imize I never ran out so far...

[@nickg1895]

This is incredible. If you're not constantly looking for these new tools that change the game, you're gonna be left behind.

[@max-imize]

Honestly the future is for the people who know how to leverage ai tools!

[@wizards-themagicalconcert5048]

Sadly it seems not to work with images uploaded from your hardrive ! I tried it and it just generates trash !

[@max-imize]

Just need to really get the prompt right then. But yeah it is hard to get the correct one because of different models

[@wizards-themagicalconcert5048]

@@max-imize Any suggestions on the prompts ? It's a family picture,so there are no prompts to begin with :( I want to extend it a bit ! Appreciate your kind reply,Thanks !

[@max-imize]

@@wizards-themagicalconcert5048 real images will be harder. Try portrait, 4k, realistic, dslr. Something like that:)

[@wizards-themagicalconcert5048]

@@max-imize Thanks mate, I will give it a try !

[@unitedstarsutopia]

Am having trouble extending a character's cropped body

[@criminaldesires]

This is so crazy, I'm crying from excitement 😭

[@max-imize]

It is soooo much fun 🙌

[@DeusExMachina50]

That is wild!

[@max-imize]

Go try it out! It makes you feel like a child again :D

[@Mindset-132-Hz]

You should cover the open-source versions like automatic 1111, there's way more tools and more things you can do then Leonardo ai since its open-sourced, tons of advancements for training and all sorts of stuff come out almost every few weeks.

[@max-imize]

Yeah it's on my list to cover soon. Unfortunately it's too heavy for most people to run on the pc. Might cover rundiffision as an easier option. What do you really like as opensource version?

[@thedevo01]

Isn't this using the Dall-e API?

[@alias234]

You can do the same things without any restrictions if you take the time to install invoke ai

[@bambinoesu]

Yh but Leonardo is web based. And invoke ai you need an expensive graphic card for it to run without lag

[@sekonameo3025]

what about license ? I didnt find any info on the app. Can we use the generated art for commercial purposes ?

[@max-imize]

Yeah you can sell it everywhere you want!

[@DarKxRaideR]

Isn't this just like InvokeAi but paid? Sure it does not need you to have the hardware on your own, but it is also not something that hasn't been seen before?

[@max-imize]

Leonardo AI is based on SD yeah. But this tool is just so easy to use for people !

[@CopterPirates]

Looks exactly like InvokeAI but as a paid/hosted version to me, too.

[@nickichandan7761]

please make a video about "consistent characters with different poses. Leonardo ai"

[@bySterling]

THIS IS NEXT LEVEL!! Wowww 🎉⚡️🥳⚡️👏🏻💯

[@max-imize]

Yeeeaaah so much fun!

[@popvisly]

So addictive 😅 need more credits got the shakes 😂

[@max-imize]

Never enough haha

[@mahnazmomeni8305]

Hi, I can't use Leonardo ai, it lead me to discord and then nothing, please help me

[@max-imize]

Hi! Watch this video: kzbin.info/www/bejne/g4jQfJ-DmKpgbsk . It shows where in discord you should send a message and fill out a form.

[@mahnazmomeni8305]

@@max-imize thanks alot💜

[@bastianrivero]

Fantastic

[@bambinoesu]

Honestly I've generated some amazing art work with Leonardo especially if you write some additional prompt

[@max-imize]

Have to try it out!

[@dynamomixx]

Can you do a review of "Rundiffusion"

[@max-imize]

I was planning to! Did you try it out already?

[@dynamomixx]

@@max-imize nope..I'm waiting for you 🤣

[@RaendyLeBeau]

GREAT STUFF! 😀 but i have a Q: if possible to create - For example, for a children's book always the same character (main characters) together in different scenes and with other characters ?😙

[@Bartetmedia]

That's the problem I have, their models are trained with only a couple of characters... not good!

[@stoparchitectsoffear]

yes, we need to keep the main character in multiple scenes. Is there an ai that does this?

[@jjfiasson9167]

@@stoparchitectsoffear Leo founder here - we’re working on a solution for that.

[@wa7smail]

Try in Midjourney "multiple poses and expression". Or search it in KZbin there are Tutorial that explain it step by step.

[@ShawnFumo]

I believe Leo lets you train your own models, so you could use that functionality to put the character in different situations. And if you don't already have some different images of that character, you probably can use controlnet to get some different poses to start with and then use those in training.

[@pn4960]

But the canvas already existed with invoke ai. I mean Leonardo nailed it but let’s give credit to whom it’s due

[@CopterPirates]

Exactly, InvokeAI has had this exact canvas feature (for free!) for months. I would not even say Leonardo "nailed it", from this video it looks exactly the same to me. So much, as if this Leonardo is simply InvokeAI with another layout/skin ...

[@PeterSolerom]

why is it telling me i need to get invited? how do I get invited into it?

[@max-imize]

Follow this guide: kzbin.info/www/bejne/g4jQfJ-DmKpgbsk

[@hellerart]

I would love to play a RTS Game like command and conquer oder settlers in this graphics.

[@max-imize]

That sounds amazing 🔥

[@Wanderer2035]

I think this format is way better than having to use discord to make the art. Discord is great, but it doesn’t feel like the right place for art generation

[@armondtanz]

Yea. Its got no room for doing all this. Could be its downfall in the end.

[@Kombo-Chapfika]

'Copy and paste' is the key to AI art.

[@rhbrolotek]

It's not a new way, it's outpainting and you could do it in Dall-E and also Stablr Diffusion UI if I'm not mistaken

[@max-imize]

Very true! How Leo brings it in such a nice gui and so easy to use is amazing though :)

[@rhbrolotek]

@@max-imize yes it really is, can't wait to see where everything goes with these improvements ! To be more exact, InvokeAI Toolkit for Stable Diffusion has something similar. But it's pretty complicated to use for a casual user ;) Would like to see something other than Midjourney trained for for cool web designs and UIX interfaces.

[@YatYas1833]

If you guys haven't picked up on it yet, the move is to create your own AI chatbot and add your own features and call it the next best thing. You're welcome.

[@evefavorse5939]

Very cool!

[@max-imize]

I think so too!

[@jamesgillis8122]

the terminator thing..... was fn hilarious

[@max-imize]

Glad you enjoyed :D

[@leandrostasi]

RunwalML is doing this called infinite image for a long time

[@Avantdor]

M enjoying Outpainting since Dall-e released it n then SD

[@ATLJB86]

One thing I haven’t seen you do is train models. Midjourney is good at making variations so you can build a model base in Midjourney then train Leonardo to make images in that style. Now you have the best of both worlds.

[@max-imize]

I trained one model here: kzbin.info/www/bejne/bpWtnqxnbZx_hKc but the results could be better! Great idea generating the images for the model in MJ!

[@souleimanabourich]

First seconds watching this video and i immediately know you're Dutch ahhaha nice accent bro

[@max-imize]

Thanks G!

[@305unleashedgraphix3]

Thanks !!

[@max-imize]

Enjoy! It's so much fun

[@305unleashedgraphix3]

@@max-imize looking forward to re editing and adding to all my previous works ! Especially the landscapes

[@hightechstuff2337]

is thia freee?

[@max-imize]

yes! 250 tokens a day :)

[@hightechstuff2337]

250 tokens means i can make 250 pictures per day?

[@max-imize]

@@hightechstuff2337 not exactly 250, maybe between 150-200

[@FusionDillusion]

Amazing :)

[@max-imize]

Thanks! 😄

[@Wanderer2035]

This is way better than mid journey 😂. I tried mid journey for hours theres no way I could make human faces or the rest of the human body that clean. Especially with the fingers it would mostly be deformed

[@timeless3d858]

Yeah this is awesome, although it's a dumbed down online interface of Automatic 1111

[@max-imize]

Yeah it's just super convenient for users :)

[@ashleybirch93]

Except the images are public aren't they

[@max-imize]

Not the ones you make in the ai canvas. If you are a business using Leonardo you can contact them to get private images.

[@QuantumFluxor]

Yeah, Leonardo tools is very cool. But Stable Diffusion is far away from Midjourney in quality. May be some blend of them will be game changer.

[@xvifim]

This looks like the openai dalle-2

[@divibs5371]

It's like DALL-E but better

[@max-imize]

Yeah you can't even compare dall-e vs leonardo imo

[@E-JFD]

The most exciting part about AI is how amazing it is in the beginning already, and how ultimately this is the worst iteration of AI we’ll ever see.. nowhere to go but forward!

[@max-imize]

Exactly! The AI future will be unimaginable!

[@eliascarlsson2624]

Midjourney still makes by far better images, the features of this ai tool was pretty cool tho

[@bambinoesu]

I've generated some images as good as midjourney with Leonardos different models and good prompt description

[@petsandpaws8906]

This is EXACTLY like Invoke AI.

[@max-imize]

True! Easy to use for everyone without downloading stuff though :)

[@pelikularecap]

is this free?

[@max-imize]

Yes for now!

[@Kakistocrator]

nice vid, cool tool, but ur mic is almost impossible to listen to

[@ArielTavori]

Lol "seemed like every day a new..." Ironic rhis video is released the week before latent couple/paint and ControlNet...

[@max-imize]

Yeah now every video is outdated a week later :o . Crazy future is coming!

[@mttpul9843]

It's just what DALL-E was doing already though isn't it?

[@KetaFPV]

I'll just stick to InvokeAI, not limited to what Leonardo doesn't want you to create and has the exact same canvas.

[@max-imize]

True! Leonardo is much easier for most people though :). Do what feels best for you

[@OolTube02]

I can't find many limitations to what I want to create these days -- particularly with models like Deliberate 1.1 in the menu.

[@TerbrugZondolop]

Subscribed got nothing in mail. Doesn’t work.

[@artvizualvideos]

Look like a combination of plagroundai and dreamlike

[@djr3386]

This is why the universe is so expansive, God kept on adding galaxies 😂

[@max-imize]

😂😂 he couldn't get enough of ai tools

[@Blackskillpower]

Last week i use only leonardo and i have so much fun with it also their discord community is awesome and helpful sometimes i have to force myself to use midjourney otherwise i pay for nothing 🤣

[@max-imize]

Haha me too! Community there is great indeed:) but MJ still can create better thing often:( AI canvas is really amazing though

[@Blackskillpower]

@@max-imize that's true but i feel i have more freedom on leonardo even if i switch between models it stay consistent in midjourney i usually get something random but beautiful

[@ShawnFumo]

@@Blackskillpower There is definitely more options in some of these tools, especially with ControlNet getting integrated (for exact posing). But keep in mind you can have a bit more control in MidJourney via input images as well as using Remix feature to guide an image in a different direction (changing the prompt as you select variations). You even can add new input images during a remix to more subtlety combine elements from another image.

[@fratta9915]

I'm afraid it's not free though

[@TheReferrer72]

Oh look they took Invoke and put it online.

[@jjfiasson9167]

No, we built it all from scratch.

[@marcus_ohreallyus]

It would be great if they would verify my email. It's been a week and they still haven't

[@max-imize]

Yeah i think they had to put a temporary stop on new users to be able to handle the traffic..

[@CharlieFleed]

It is mindblowing but at this stage it still feels like a huge waste of time. It just doesn't do what you ask and when it does it looks awful. You can see too clearly it is just regurgitating other images.

[@oldrichstibor2845]

this is not new. dall-e has been able to do this for awhile now.

[@Bartetmedia]

Their Interface is good but the quality of the images are not that great, and too much of the same character.

[@LloydLadera]

As a traditional artist : welcome new AI artists. Bring your AI brushes and take a seat beside the other artists and let's keep making cool art together.

[@mindysommers]

Nah.

[@marcobelletz4734]

it's the same thing almost all other AI can do...

[@mr_pip_]

To be honest, I don't like this Leonardo policy very much. I've been waiting for my invite link for what feels like an eternity now, and I find this elitist posturing completely inappropriate: especially since you can already achieve quite wonderful results with Stable Diffusin, which now allows targeted gestures, outpainting and even video production.

[@max-imize]

I think unfortunately Leonardo userbase grew waaaay to fast since they are still a small company. That's why they need to put a stop on new users. But i don't think they are elitist tbh. Just bringing very easy UI to the people :)

[@Andonchoto]

Dalle is crap. MJ is expensive. Leonardo does not have Hires fix. SD for LIFE!

[@quintenmantez6934]

Try openai or lexica!

[@puamana2]

Slow down. We are beginners.

[@Ilyak1986]

Tokens, meaning it's a pay to play app? Sorry, but...I'm not opening my wallet up indefinitely.

[@max-imize]

You have 250 tokens each day. So really enough to have some fun

[@Wirmish]

MidJourney is super bad at hands drawing. Or things held in the hands of the characters. Can't draw a Pegasus. Can't draw a crow/raven. Will wait 1 or 2 years before try it again.

[@max-imize]

Innovation is going so fast! will be much better before you know it :)

[@iloveqq2011]

MJ's UI is terrible!

[@max-imize]

Yeah that is just classic discord. I released a recent video with 3 free art generators you don't need disc for :)

[@ShawnFumo]

They do have a beta version of an interactive GUI to make images, but it isn't rolled out to all users yet. They recently started a re-vamp of the web UI in general, so am crossing fingers they'll get the creation flow out of beta soon.

[@Rai2M]

Stop saying "art".