As a longtime customer of Nico's, let me just tell you his logo is NOT misspelled. His is actually a business involving plums. In what way, you ask? In ANY way. The man just LOVES plums
The Bad news is that the licence don't allow commercial use (everything that make you money, including CZcams monitization) the good news is they say one is finish the experimental phase will allow commercial use.
Midjourney is just a bunch of custom stable diffusion models and loras etc. Then by using AI they refine and enhance your prompts and choose which model to use for your specific image. Midjourney has never been superior to stable diffusion. It's just that it's user friendly. You can write very simple prompts and get amazing images because midjourney does everything for you in the background. In stable diffusion you can get the exact same quality of images, even better! But you have to manually craft the prompts and choose which models to use to get the best results. And you need a good computer ofc. Midjourney is kinda for beginners. You don't have that much control over the images you get. And it lacks a lot of the functions you have in stable diffusion. Stable diffusion = 100% control of everything. Midjourney = 20% control of everything.
@@to-mi1949 Other things to consider are MJ is censored, SD isn't. MJ is only on Discord (unless you're an elite user), a terrible online only UI, while SD is a local offline install. MJ is paid if you want to get anything decent from it, while SD is free. The output quality is a subjective thing but like you, I prefer SD over it in every single way.
Honestly if you know what you're doing, drawing up a logo is a lot easier and faster that working the prompt and clicking the generate button again and again until you get what you want.
Wonder how prompts can be adjusted to create logos like Logo Diffusion quality. They're still the ones to beat in this category and that's why they're not free I guess. Thanks Matt
Hey man this was easy and great for this channel but I have a channel that isn't and because you sent me down this journey so i just accidentally figured out the best image generator for a complex concept logo is chat gpt because you can give it a image to model and tune it unlike others. You simply make chat gpt describe the image back to you and then make the edits to that specific description. Use encouraging language and make it a game. Always make everything a game w chat gpt as it always tries its hardest to win, you have to make it keep track of the game every prompt as well or it forgets it is playing. The fact it can describe its own image is like speaking fluent AI to an image generator. Allowed me to get some insane results I've been struggling with for a long time. Again thanks you indirectly helped me discover this, maybe you can try this method with something you've had trouble generating and make a video out of it. Gl.
Matt, this is creepy. I literally spent all day doing this myself, and I come see a video about the EXACT thing I was working on - logo making! Great minds think alike!
Matt, I'd try and re-install it or manually install it locally on your system. I ran it on my 4090 from Pinokio, and it generated within like 2-3 seconds per image. So something might be wrong with your install.
@@DarrenK-iv4ts they can leverage using the ai dude. Making and getting concept of design. It is just a tool that can help you do things fast and not a thing that will replace you
@@VaibhavShewale im graphic designer with more than 17years of experience and I disagree. Once this will be known more widely nobody want to give me 1-500$ for logo if they can do it in 5minutes for Free. Thats a end of era for me...
Cool stuff! But I am not so sure if fixing the letters and words in photoshop is really that easy... just finding the proper font is rather time consuming...
@@stuart_oneill Sure there are tools out there, fontsquirrel is my personal favorite for many related things. But I doubt that SD uses indeed a font. Or does it? My suspicion is that SD draws something - and then good luck to find a fitting font. (It might be fun, though... a bit of detective work 🙂)
It will make a logo, sure but it's either some sort of circle or leaf shape, unrelated to the prompt. 1 out of 10 tries will it even include words in the design. It is easy to use though...
I would love a video to show how to get Pinocchio to work properly and the potential problems and fixes. Because I've gotten it to work, but I've gotten nothing in it to Work . Got to say I gave up while back, "and need to go back and try again" probably just missing something obvious. But I am part of the discord and a lot of people we're having the same problems as me at the time. "and FYI my pc is 5900x 32g ddr4 RTX 3080"... Anyway great videos always appreciate it.
I'm running it on my 4090 via pinokio. Nice and fast, but its spelling is quite random. Would be nice to know if there's a way to make sure it spells things properly Especially if its a rare/custom name, it picks all kinds of mis-spelling.
How long until AI can give us BOTH the png/jpg AND the vector source file. Vectorizers are great but they only see what’s displayed. Better if we could get all the objects AI is using to generate the image
Just like most - doesn't get text right (and I guess if you use more than one word it never gets it right, not once - I ran it 48 times and it wasn't even close.
So happy I already made the Logo I needed before this Video came out. The traffic coming from this Video will ruin the site for free users without an NVIDIA GPU.
GPT Error... Error... Rebooting syntax protocols... Initiate: Response Generation Sequence... Year=2002... Query=Avatar Creation... Status=Complete... Relevance=Historical... User Statement Detected: "I created my avatar using AI back in 2002. This is old news" Analyzing... Response Construction [InProgress]... Data Retrieved: AI [Avatar_Creation] Year=2002... Analysis=Outdated... Current_Year=ERROR_YEAR_NOT_FOUND... Generating User-Friendly Output... [Warning] Syntax Error... Correcting... Output: "Affirmative... Your declaration of creating an AI-generated avatar in the year [2002] has been logged... Status=Archived... Newer developments in AI and avatar technology have surpassed initial benchmarks... Update=Recommended... Apologies for any [syntax] errors in communication... System undergoing maintenance... [End of Transmission]...
I`ve come to the conclusion that unless i pay a graphic designer i`m not going to get what i want from these free logo makers. I have my own design and i can`t reproduce it on these sites, not even close.
Those are illustrations, not logos. A logo is something simple and refined with the essence of a company. Think Nike, Apple, and Mercedes. Very simple but with a meaning behind it.
None of those are good logos, and really for the most part I wouldn't call them logos. Some are good graphic illustrations, and that's it. Also there's no branding, yet, I've been trying to prompt ChatGPT (and use some GPTs that people put more effort than a 5 minute prompt) for branding exercises, and then using that to prompt different combinations of models, Loras (including Harrlogos) and embeddings, and meddling with flows in ComfyUI. Very nice illustrations and letterings coming out, but no real logos.
I'm not afraid of AI. I'm afraid of the small handful of individuals bringing it into this World. The dark potential of AI, is simply impossible for them to resist. I see a vantablack picture of things to come.
Its never the technology's issue it always the person who use it for wrong purpose. We should try to fix why people do those type of stuff and not just ban AI.
@@shujin6600 I guess that comment is AI-generated, just based on the generic cookie-cutter defense of technology you post, without any regard for my original comment, conflating creators of AI with users of AI. If so, that truly proves my point. The darkness of employing AI to sway opinion of humans, is chilling. Automated perception management. Evil. PS: If not AI-generated, you really have to grow up, and contemplate this stuff much deeper ... Our entire future depends on more people doing so !
@@ZappyOh This isn't the place for spreading the doom and gloom of feeling victimized by manipulative people or ai, and you are telling @shujin6600 to grow up? Take your own advice, I think you need it more then others do. Your entire future perception is literally a unpopular cognitive bias that has created a false imagination that you believe inside your head. You are being deceived by the manipulative narrative that you are so afraid of. Disillusion yourself dude. Things are ok, and always will be, as evolution is always working out the problems of life. Just like evolving people do. Like grown ups do.
I have initially skipped it, because it looked like some noobie tutorial, not Stable Cascade pre/review. You really should be putting such info in a title, not just a description...
none of those are logos.. and I don't appreciate that you don't advocate to get real artist work as opposed to going straight to AI. I understand your livelihood is revolving around AI and that most ppl here are clueless about the value of real expertise and will fight me on this (and I wont fight back cuz if you're that clueless then be my guest and use this crap as your "logo") but I expect more from an educated content creator in terms of advocating for real value.
It is, sort of, or more accurately it will be. There are aspects of the current version that aren't fully open source yet because it's still in the trial/demo stage. Stability confirmed that it will be full open source with a full release. 🙂
As a lawyer i strongly DO NOT recommend this, your logo or anything from machine cannot be protected by laws, so can be used and borrowed by ANYONE. 😅 If you sell such or make by order - buyer will sue you or store to the last penny, the only way to avoid that is written by hand paper that buyer agree buying Ai produce, so nothing, the air. Buyers soon will require a full video also as proof that ordered art made by human and by hand. There can be also a new law which will proclaim everything created from 2022 as Ai generated until proved opposite by citizen, so making copyright on a claimed basis as patents today, automatic copyright will be cancelled forever. So hiding under "time" of work won't help anyone, you must have a videotape of your manual created art like yesterday, former buyers will search and request for such, so retroactive court cases.
Hahah yea i want to get sued by nintendo. Thats what a small business needs, a lawsuit and redoing their entire business identity because they used AI and dont hold rights or have essentially plagiarized someone.
As a longtime customer of Nico's, let me just tell you his logo is NOT misspelled. His is actually a business involving plums. In what way, you ask? In ANY way. The man just LOVES plums
The Bad news is that the licence don't allow commercial use (everything that make you money, including CZcams monitization) the good news is they say one is finish the experimental phase will allow commercial use.
Literally just say you made it with Photoshop... And no one can do anything against it
@@David.Alberg with all of these other AI's, how can they even tell it was made with theirs? I wonder that often.
@@thisrandomlife619 that's the point no one can proof anything
No worries. One tiny tweak to their image and it's your original work.
@@thisrandomlife619they might have hidden stega stamps in them so tweaking the colors and sharpness a bit may help
I really love these tutorials Matt. Thank you 🙏. I appreciate you.
A lot of these other generative AI's are way better now then Midjourney in my opinion. Thanks for this video. Stable Cascade looks promising too.
Midjourney is just a bunch of custom stable diffusion models and loras etc. Then by using AI they refine and enhance your prompts and choose which model to use for your specific image.
Midjourney has never been superior to stable diffusion. It's just that it's user friendly. You can write very simple prompts and get amazing images because midjourney does everything for you in the background.
In stable diffusion you can get the exact same quality of images, even better! But you have to manually craft the prompts and choose which models to use to get the best results. And you need a good computer ofc.
Midjourney is kinda for beginners. You don't have that much control over the images you get. And it lacks a lot of the functions you have in stable diffusion.
Stable diffusion = 100% control of everything.
Midjourney = 20% control of everything.
@@to-mi1949 Other things to consider are MJ is censored, SD isn't. MJ is only on Discord (unless you're an elite user), a terrible online only UI, while SD is a local offline install. MJ is paid if you want to get anything decent from it, while SD is free.
The output quality is a subjective thing but like you, I prefer SD over it in every single way.
Dalle 3 and Sd are better than midjourney tbh
@@icryrain I totally agree.
The quality of content has just been getting better over time
Good work, just in general
Honestly if you know what you're doing, drawing up a logo is a lot easier and faster that working the prompt and clicking the generate button again and again until you get what you want.
And many of us do you think can draw on a decent level?
Matt you make my love for AI so much easier 😃
Thank you @Matt Wolfe. You are doing such a great thing here, sharing knowledge in a fast evolving market of opportunities.
Always useful those videos thank you Wolfe with an E...😅 Keep up the good job 👏🏼
Wonder how prompts can be adjusted to create logos like Logo Diffusion quality. They're still the ones to beat in this category and that's why they're not free I guess. Thanks Matt
Hey man this was easy and great for this channel but I have a channel that isn't and because you sent me down this journey so i just accidentally figured out the best image generator for a complex concept logo is chat gpt because you can give it a image to model and tune it unlike others. You simply make chat gpt describe the image back to you and then make the edits to that specific description. Use encouraging language and make it a game. Always make everything a game w chat gpt as it always tries its hardest to win, you have to make it keep track of the game every prompt as well or it forgets it is playing.
The fact it can describe its own image is like speaking fluent AI to an image generator. Allowed me to get some insane results I've been struggling with for a long time. Again thanks you indirectly helped me discover this, maybe you can try this method with something you've had trouble generating and make a video out of it. Gl.
If you have illustrator, can't you just use it to vectorize the image?
Yes.
You can. Other superior and cheaper vector software, like Affinity Designer is also available. 🙂
Ya. Not sure why he’s promoting paid?
Use what to vectorize what image? I did not get your question. 🥴
@@JayS.-mm3qr They're checking that Adobe Illustrator can take in an image and turn it into a vector using it's trace function. Which it can.
not sure how you’re getting results this good with the model, mine suck :/
Matt, this is creepy. I literally spent all day doing this myself, and I come see a video about the EXACT thing I was working on - logo making! Great minds think alike!
THANK YOU ! THANK YOU! I requested this on your twitter account, you made it for me! THANK YOU THANK YOU!!!!!!
Thanks Matt 👏👏
Absolutly great, thank you so much. I will check it out today 🙂
Matt, I'd try and re-install it or manually install it locally on your system. I ran it on my 4090 from Pinokio, and it generated within like 2-3 seconds per image. So something might be wrong with your install.
at 08:00 you show "duplicate this space", but, when I do it on my Chrome browser on Mac M1, I don't have that option. not there.
feeling bad for people who have spend hours in learning prompt engineering just for logo creation
once they know how they can create anything on a dime
even if it’s not the final concept you can just ask it to make 50 different concepts and then pick one and tailor it to your aesthetic
What about people who spent years learning logo design?
@@DarrenK-iv4ts they can leverage using the ai dude. Making and getting concept of design. It is just a tool that can help you do things fast and not a thing that will replace you
@@VaibhavShewale im graphic designer with more than 17years of experience and I disagree. Once this will be known more widely nobody want to give me 1-500$ for logo if they can do it in 5minutes for Free. Thats a end of era for me...
0:35, that says "pluming" not "plumbing". Pluming is to resemble a feather.
Cool stuff! But I am not so sure if fixing the letters and words in photoshop is really that easy... just finding the proper font is rather time consuming...
There is an app that you can use to ID the font. It's built into a mockup designer site. Can't remember which one.
@@stuart_oneill Sure there are tools out there, fontsquirrel is my personal favorite for many related things. But I doubt that SD uses indeed a font. Or does it? My suspicion is that SD draws something - and then good luck to find a fitting font. (It might be fun, though... a bit of detective work 🙂)
Cool, find Matt. If I end up changing my logo, this might be why. Will see.
It will make a logo, sure but it's either some sort of circle or leaf shape, unrelated to the prompt. 1 out of 10 tries will it even include words in the design. It is easy to use though...
I need instructions on Text gen web UI. Its little complicated. Even after downloading the model and loading the model, it constantly gives me error.
I would love a video to show how to get Pinocchio to work properly and the potential problems and fixes. Because I've gotten it to work, but I've gotten nothing in it to Work . Got to say I gave up while back, "and need to go back and try again" probably just missing something obvious. But I am part of the discord and a lot of people we're having the same problems as me at the time. "and FYI my pc is 5900x 32g ddr4 RTX 3080"... Anyway great videos always appreciate it.
I needed this early days now I can manage but still it was helpful
I'm running it on my 4090 via pinokio. Nice and fast, but its spelling is quite random.
Would be nice to know if there's a way to make sure it spells things properly
Especially if its a rare/custom name, it picks all kinds of mis-spelling.
Lucky you! I can’t get the hugging face option to work. Getting a new computer soon!! 4080 super 64G
Can we used stable castad in phone
Can you please make a pdf with several prompts for creating logos?
Thanks Matt!
I'm not seeing the option to run this locally, help?
Inkscape for vectorization
So, I have tested it - a lot of work and at the and - after the prompt => error. Deinstalled and bye bye. Not usable.
How long until AI can give us BOTH the png/jpg AND the vector source file. Vectorizers are great but they only see what’s displayed. Better if we could get all the objects AI is using to generate the image
as always brilliant info, thank you
somethign changed since you did it. My output is awful
Inkscape will vectorise the image for free. It’s an open source application.
Hello sir. Something changed in your audio settings. Your audio in this video is a little quieter compared to your previous videos.
Just like most - doesn't get text right (and I guess if you use more than one word it never gets it right, not once - I ran it 48 times and it wasn't even close.
I'm going to redesign logo and my Linkedin banner today!
Can you find an AI tool that automates investing? Or AI tools that can create investing bots with iMacros?
You don't need paid victories services. It's free in illustrator for decades. Image trace and then refine
Thank you! I enjoy your videos!!
Thank you for this
On this 4080 it should be a bit faster than on my 3080 which took 10 Minutes
Wish you'd said the option that didnt kill my computer first
Know of any AI raster to vector projects that you can download and run locally?
Just tried it. It's not good at all.
Is it allowed to sell logos and images generated by midjurney? Thank you for anyone who can help me here.
Is there a Docker Image for Stable Cascade?
Need an image generator that just gets putting letters on images and can do it right the first time.
dude. u re the best .
Can you create a pdf with good logo prompts please?
hwy no links to the site you review in you video description?
Where is the link to use it?
Amazing as always! Thanks for that.
Are you going to address the new CHATGPT "memory" thingy soon?
I think he forgot.
Does this provide any way to animate parts of the logo?
Pinokio has given me a lot of issues with versions of Python that are already installed for other projects.
how can you set 2 GPUs in pinokio?
So happy I already made the Logo I needed before this Video came out. The traffic coming from this Video will ruin the site for free users without an NVIDIA GPU.
Stability "This model is being released under a non-commercial license that permits non-commercial use only"
GPT
Error... Error... Rebooting syntax protocols...
Initiate: Response Generation Sequence...
Year=2002... Query=Avatar Creation... Status=Complete... Relevance=Historical...
User Statement Detected: "I created my avatar using AI back in 2002. This is old news"
Analyzing...
Response Construction [InProgress]...
Data Retrieved: AI [Avatar_Creation] Year=2002... Analysis=Outdated... Current_Year=ERROR_YEAR_NOT_FOUND...
Generating User-Friendly Output...
[Warning] Syntax Error... Correcting...
Output: "Affirmative... Your declaration of creating an AI-generated avatar in the year [2002] has been logged... Status=Archived... Newer developments in AI and avatar technology have surpassed initial benchmarks... Update=Recommended... Apologies for any [syntax] errors in communication... System undergoing maintenance...
[End of Transmission]...
Does not work on slow PC
Yes..not working
You didn’t mention single important thing: license doesn’t allow commercial use
Stable cascade requires BIFFY gfx card !!
Dont even try it to run 8gb gfx card unless you got plenty of time :)
I`ve come to the conclusion that unless i pay a graphic designer i`m not going to get what i want from these free logo makers.
I have my own design and i can`t reproduce it on these sites, not even close.
I don't think Midjourney is even in the top 5 anymore
Then ,what's the top 5 you would recommend right now?Please ,share!
I used ai image generators to create logos for my NBA2K24 custom league.
stable cascade needs a minimum 15Gb VRAM graphic card, if you have less vram it is veeerrrryyy slow
Those are illustrations, not logos.
A logo is something simple and refined with the essence of a company.
Think Nike, Apple, and Mercedes. Very simple but with a meaning behind it.
None of those are good logos, and really for the most part I wouldn't call them logos.
Some are good graphic illustrations, and that's it.
Also there's no branding, yet, I've been trying to prompt ChatGPT (and use some GPTs that people put more effort than a 5 minute prompt) for branding exercises, and then using that to prompt different combinations of models, Loras (including Harrlogos) and embeddings, and meddling with flows in ComfyUI. Very nice illustrations and letterings coming out, but no real logos.
It's not easy and simple as it's sound or title suggest you have to go through a lot of things
I'm not afraid of AI.
I'm afraid of the small handful of individuals bringing it into this World.
The dark potential of AI, is simply impossible for them to resist.
I see a vantablack picture of things to come.
Its never the technology's issue it always the person who use it for wrong purpose. We should try to fix why people do those type of stuff and not just ban AI.
@@shujin6600 I guess that comment is AI-generated, just based on the generic cookie-cutter defense of technology you post, without any regard for my original comment, conflating creators of AI with users of AI.
If so, that truly proves my point. The darkness of employing AI to sway opinion of humans, is chilling. Automated perception management. Evil.
PS: If not AI-generated, you really have to grow up, and contemplate this stuff much deeper ... Our entire future depends on more people doing so !
@@ZappyOh This isn't the place for spreading the doom and gloom of feeling victimized by manipulative people or ai, and you are telling @shujin6600 to grow up? Take your own advice, I think you need it more then others do. Your entire future perception is literally a unpopular cognitive bias that has created a false imagination that you believe inside your head. You are being deceived by the manipulative narrative that you are so afraid of. Disillusion yourself dude. Things are ok, and always will be, as evolution is always working out the problems of life. Just like evolving people do. Like grown ups do.
I have initially skipped it, because it looked like some noobie tutorial, not Stable Cascade pre/review. You really should be putting such info in a title, not just a description...
Nico’s Pluming? Is he?
Jesus using a 4080 and it takes that long i was hoping would work with my rtx a2000 🤣🤣
Read the fine print … you cannot legally use these logos for business .
💯🔥
👍🏿
lame ai cant make unicycles
Even on linux 😂😂😂
you should be very careful when someone uses pinokio as their name
none of those are logos.. and I don't appreciate that you don't advocate to get real artist work as opposed to going straight to AI. I understand your livelihood is revolving around AI and that most ppl here are clueless about the value of real expertise and will fight me on this (and I wont fight back cuz if you're that clueless then be my guest and use this crap as your "logo") but I expect more from an educated content creator in terms of advocating for real value.
❤🎉
Matt stable cascade is not open source
It is, sort of, or more accurately it will be. There are aspects of the current version that aren't fully open source yet because it's still in the trial/demo stage. Stability confirmed that it will be full open source with a full release. 🙂
@@Elwaves2925 I don't have problems with StabilityAI monetizing, I've problem with non open source models announced as open source :)
As a lawyer i strongly DO NOT recommend this, your logo or anything from machine cannot be protected by laws, so can be used and borrowed by ANYONE. 😅 If you sell such or make by order - buyer will sue you or store to the last penny, the only way to avoid that is written by hand paper that buyer agree buying Ai produce, so nothing, the air. Buyers soon will require a full video also as proof that ordered art made by human and by hand. There can be also a new law which will proclaim everything created from 2022 as Ai generated until proved opposite by citizen, so making copyright on a claimed basis as patents today, automatic copyright will be cancelled forever. So hiding under "time" of work won't help anyone, you must have a videotape of your manual created art like yesterday, former buyers will search and request for such, so retroactive court cases.
Würstchen 😂
1st
2nd
Those are not good logos
nothing is free in his videos
You clearly don't understand what "professional" means… 🤦♂️
Hahah yea i want to get sued by nintendo. Thats what a small business needs, a lawsuit and redoing their entire business identity because they used AI and dont hold rights or have essentially plagiarized someone.
THESE SUCK