Former Community – We Moved
🆕 Coached community here, for posts & feedback public community here.
Infographics: sharing, previews, & feedack
Thanks for your trust Lucio.
Some images in this article seem to be AI-generated.
https://thepowermoves.com/an-apology-without-change-is-manipulation/
In those images, the fingers seem to be a bit weird.
Example of an Image that Looks AI-Generated
For example, in the image below, the man's left hand (on the right if viewed from our perspective) has six fingers.
I bring this up in case a graphics designer is using them to generate these images.
Lucio or someone else here may very well be already generating these images via AI.
If a graphics designer is doing it, he probably spends only a few minutes on each image or even a few seconds.
I thought some money and time could potentially be saved.
Generating a Similar Image with BlueWillow
I tried replicating the image above with the following prompt to the AI image generator, BlueWillow, which offers a few free test generations
https://www.bluewillow.ai/
A man with closed eyes and an annoyed facial expression. Both of his hands are raised up with palm facing forwards.
Typing the prompt into BlueWillow gives the following two images:
I got sort of lucky and got what I wanted from the AI model on the first try.
I gave the example of BlueWillow if you're interested in trying out a few free generations.
However, if these kinds of images appeal to you, I would not advise on a paid subscription to BlueWillow.
Proprietary AI Image Generators: MidJourney & OpenAI's DALL·E 3
I would go for MidJourney or OpenAI's ChatGPT subscription which includes DALLE-3.
MidJourney: https://www.midjourney.com/
ChatGPT: https://chat.openai.com/auth/login
DALL·E 3: https://openai.com/dall-e-3
The downside is that there's no free option for both.
MidJourney starts from USD 10 a month.
https://docs.midjourney.com/docs/plans
ChatGPT Plus costs USD 20 a month.
The upside of ChatGPT Plus is that you get all the other features as well (GPT-4, Advanced Data Analysis for Coding, Image Recognition, Plugins).
For example, you could ask GPT-4 (the chatbot) to refine a prompt to generate the images.
Then, you may not even need to think of an initial, elaborate prompt to describe the image you want.
OpenAI is gradually rolling out DALLE-3 to all plus subscribers so getting the subscription later could be more worth it.
If you get it now, it would depend on luck on whether you get DALLE-3.
Open-Sourced AI Image Generators
It's possible to use open-sourced AI models and software to generate your own images but it requires a bit more technical know-how.
https://github.com/lllyasviel/ControlNet
But sometimes people are kind enough to host state-of-the-art open-sourced, AI image generators for free!
Here is an example of Stable Diffusion XL Base 1.0 (the name of the AI image-generation model) hosted on Hugging Face:
https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0?text=A+man+with+closed+eyes+and+an+annoyed+facial+expression.+Both+of+his+hands+are+raised+up+with+palm+facing+forwards.
The field encircled in green is where you enter the image prompt.
The AI-generated image is encircled in red.
Feel free to try out what works for you!
I personally like the open-sourced model above.
Hello Matthew,
first of all, good to see you here and welcome back!
And for that image (and a few more), I'm aware of that and yes, you're absolutely right, it's AI-generated (hadn't noticed it was 6 fingers LOL).
They're from the Microsoft offering, and so far I find it superior to the others (including huggingface).
Quote from Matthew Whitewood on October 15, 2023, 4:40 pmI would go for MidJourney or OpenAI's ChatGPT subscription which includes DALLE-3.
MidJourney: https://www.midjourney.com/
ChatGPT: https://chat.openai.com/auth/login
DALL·E 3: https://openai.com/dall-e-3The downside is that there's no free option for both.
MidJourney starts from USD 10 a month.
https://docs.midjourney.com/docs/plansChatGPT Plus costs USD 20 a month.
The upside of ChatGPT Plus is that you get all the other features as well (GPT-4, Advanced Data Analysis for Coding, Image Recognition, Plugins).
For example, you could ask GPT-4 (the chatbot) to refine a prompt to generate the images.
Then, you may not even need to think of an initial, elaborate prompt to describe the image you want.OpenAI is gradually rolling out DALLE-3 to all plus subscribers so getting the subscription later could be more worth it.
If you get it now, it would depend on luck on whether you get DALLE-3.
Thank you for this, it's helpful!
I think it makes sense to test out that DALLE-3 once ready and if it's good, then a subscription would probably make sense.
first of all, good to see you here and welcome back!
Appreciate the warm welcome, and good to see you again too!
I realised that the BlueWillow images disappeared.
They're from the Microsoft offering, and so far I find it superior to the others (including huggingface).
Yeah the huggingface example is probably not the best.
I'm trying to find a way to adapt open-sourced models to the level of commercial models.
I think Microsoft is ahead among the big tech companies due to its partnership with OpenAI.
Microsoft invested USD 10 billion and owns 49% of OpenAI.
So far Microsoft looks like it's plugging in OpenAI's models all over its products.
I think it makes sense to test out that DALLE-3 once ready and if it's good, then a subscription would probably make sense.
OpenAI says that it will release fully to all plus subscribers this month of October.
So unless the gradual roll-out backfires for whatever reason, it will be accessible to all plus subscribers by November.
Here's a snapshot (I got access to DALL·E 3):
I used MidJourney for two months in May/June when I needed to make lots of images.
It was the leading image generator.
But, since OpenAI included DALL·E 3, I may not need MidJourney anymore in the future.
Some people did a direct comparison of DALL·E 3 versus MidJourney.
https://goldpenguin.org/blog/dalle3-vs-midjourney/
In short, I'm happy with DALL·E 3 and the whole ChatGPT Plus subscription at the moment.
So I'll be sticking to that.
Hi all,
- Goal of this infographic:
Convey the concept of Power University built upon thousands of books/studies.
- Where:
It may go in the "resources" page
- Question for you:
Any preference for these?
The designer asked me if I prefer the university to be frontal, like in the first picture. Or sideways, like the other two.
Any preference?
Any other idea welcome.
1:
2:
3:
Hi Lucio. My preference is for the first one. I like the cleaner design and the eagles flying high convey better the courses' main goal. In general, I think nowadays scientific expertise is more associated with published papers than books (which I associate more with literature), especially in "hard" sciences. The frontal orientation conveys better that the viewer is about to enter that university. The rotated versions give me the impression of inaccessibility, and as if I were just looking at a nice landscape.
The other two are beyond doubt exquisite, but would make me think that you teach philosophy or esoteric stuff.
Quote from Lucio Buffalmano on December 9, 2023, 4:51 pmHi all,
- Goal of this infographic:
Convey the concept of Power University built upon thousands of books/studies.
- Where:
It may go in the "resources" page
- Question for you:
Any preference for these?
The designer asked me if I prefer the university to be frontal, like in the first picture. Or sideways, like the other two.
Any preference?Any other idea welcome.
1:
2:
3:
Cheers.
Personally I like more the 3rd picture.
The first one looks very artificial (seeing so many eagles and paper sheets in a single photo feels weird for me for some reason) and the second one looks like a colored drawing to me.
Third one looks more clean, the background and seeing the building sideways from afar adds a nice touch.
Thats how I see it.
Thank you so much Kavalier and Power Duck, golden feedback!
To add some more details:
- It's initial AI examples so far: so probably very far from the finished product
- PU building logo will be the "building", so these buildings are at most just inspiration for now
- Background/eagles: the eagles in #1 is an added prompt I added later. The background in #2 and #3 may be ignored, but I'm thinking about the eagles
Definitely the first one for me (although there are some things I'd probably change to it).
The second one has too many gears for my taste which gives me the feeling of an engineering class—and with the soccer field, it ends up coming across as too off-topic.
The third one, once again, has a feeling of space (even NASA), science, and technology for me, and I don't get the impression of "social and power skills" by looking at it.
By the way, I wonder how it'd look to give the eagles in the first image small graduation caps, so it implies, "The students became high-flying eagles once they graduated Power University."
Hello Lucio,
the first one! It’s a bomb. I did not read yet other comments.
Quick take on this: the only thing to tweak in my opinion is the eagles. I think there are too many of them. I think it’s good to have less and more different sizes: big one medium very small to show perspective.
Might comment more later.
