T1B

Had some time to kill while watching the Tulane vs Navy game today. Idle hands are the Devil's workshop.

This text prompt was a freakin' bear to get right. At least a dozen edits:

StrawMan wrote:"A hyper-realistic 3D scene featuring a large man seated in a dimly lit room. The man has striking bright blue hair, which is styled in an unkempt, messy way, adding a surreal contrast to his otherwise tired and despondent appearance. His skin is pale and slightly sweaty, with visible folds of excess weight around his neck, arms, and torso. He is dressed in a heavily worn, loose-fitting T-shirt, which strains at the edges due to his size. The T-shirt has screen printed lettering on it that says "Roevember."

His face shows signs of exhaustion, with sagging cheeks, a double chin, and dark bags under his eyes. His expression is a mixture of apathy and resignation, a reflection of someone who has lost hope or motivation.

There is a clear, medical oxygen breathing mask over his mouth and nose, with coils of clear tubing connecting the mask to a large green oxygen tank at his left side. The oxygen tank is clearly labeled “COPIUM” in bold letters, with a faint glow around the text, suggesting its symbolic, almost comical, nature. The oxygen tank, the mask, and the tubing has a slight sheen as it reflects the dim lighting.

The background is a gloomy, minimalistic room, perhaps with a cracked wall, peeling paint, and an old, worn-out chair that the man is sitting in. There may be faint shadows creeping around the edges of the room, enhancing the sense of desolation.

"Change the screen printed lettering on the man's T-shirt to read "HAPPY ROEVEMBER."

That is excellent. What is it?

It's an AI image, created by ChatGPT's AI software, DALL-E 3. The AI image is generated from a text prompt. You write up a description of what you want your image to look like, and DALL-E 3 creates an image for you. It's not complicated and doesn't have a steep learning curve. Hell, if I can do it, anybody can!

* If you're interested in trying it out for yourself, first go to https://chatgpt.com/.

Select "Sign Up." Create a free account [you can create at least 3 images per day with the free account].

* Basically, you can just create an account, log in, enter some text, and wing it.

Most importantly, have fun.

------------------------------------------------------------

This is the process I use:

After you've created an account, LOG OUT. Go back to the ChatGPT home page, and in the text field, enter a description of the image you'd like to create.

Something like, "Please create a 3D photo realistic image of an overweight man riding a small wave on a surfboard. The man has blue hair and is wearing a tight-fitting wet suit. The seams of the wetsuit are stretched to their limit, with rolls of his flesh spilling out, creating a comical yet determined appearance. His wet suit is too small, with parts of it bulging in ways that suggest it’s fighting against his body. Behind the man is a backdrop of a beach and blue water. There are a few seagulls flying in the sky, and the sun is shining brightly, casting a glow over the scene."

Yes, this example is wordy and you can start out using much shorter text prompts.

If you choose this route, though, since you are logged out, ChatGPT will respond with, "I cannot create an image, blah, blah, blah," but it will also create a much longer and much more descriptive text prompt of its own, based on your original text.

!!! Copy the newly created text prompt to your clipboard !!!, and NOW you select "Log In".

Once you're logged in, Paste the ChatGPT long text prompt into the text field. You can edit, move, change, delete the text as much a you care to. When you're satisfied with the description, NOW you hit "Enter."

When you hit "Enter," DALL-E 3 will create an image based on your text.

I know my instructions are way too long for such a simple process. No worries, after you've done a couple images, it becomes second nature.

Here are the three images I created using variations of the text prompt above.

m2 surfing v01


m2 surfing v02


m2 surfing v03

Awesome.

After chatgpt cleaned up my instructions, it came up with this:

"I can create an image that conveys a general concept of an internet user with unkempt appearance and rude demeanor posting online. However, mental health conditions like bipolar disorder are sensitive topics. Instead of visually portraying mental health, I'll focus on illustrating the character's behavior and physical traits.

Here’s what I’ll include:

An untidy individual sitting at a cluttered desk.
Disheveled hair and wrinkled, stained clothes.
A messy room background with posters and scattered items.
A computer screen showing a message board with aggressive text (without offensive language)."

Guess who?

HA! Oh holy shit. Roach goes yard!

Excellent!