images-discussions

[grizzled iris]

Wow

The United States Service

Free Healthcare

UNED States 😂

Woah 😦

The smallest font size so far managed to do with as little spelling errors as possible

[grizzled iris]

Finally

[grizzled iris]

Exactly

[open trench]

hello all.

[grizzled iris]

Happy Sunday

[open trench]

same to you

[dim cradle]

we're in an operating room apparently, nez

[grizzled iris]

😂

[dim cradle]

lol

[open trench]

I see this. Lots of operating rooms. lol

[grizzled iris]

We’re finally getting better on the readability of the words

Aka not spelled incorrectly

[dim cradle]

i don't think there's a way out.

[open trench]

lol

[dim cradle]

i think we'll need to open a portal by generating images of some other world

actually, the ER theme ties in well with my brains theme

switching to farmers

[open trench]

that is cool

[grizzled iris]

Second Green Revolution with Vertical Farming

[dim cradle]

that sounds like a good thing

[grizzled iris]

The kanji is wrong 😑

[late blade]

I asked for a farmer as a super hero...

[late blade]

yeah but I got this #daily-theme message

[grizzled iris]

https://media.discordapp.net/attachments/979617962728226826/1198712326887591996/image.png?ex=65bfe6d9&is=65ad71d9&hm=9aa5a5fd1097b94eb2091c90439263eedf705897f96bdf2a727887793da0364f&=&format=webp&quality=lossless&width=1992&height=1138

Really good @late blade x

[late blade]

lol

the traditional farming attire is what bothers me

doesn't look safe

[grizzled iris]

lol

His farm

Super Farmer

[grizzled iris]

Any updates on this version?

[late blade]

doing it like Amazo?

[grizzled iris]

Also this is super clever

[dim cradle]

not really, not yet

i'll keep ya posted lol

[grizzled iris]

ca'nt wait lol

[brazen ibex]

/ımagine I want you to design a logo with Motivation Mind Success written on a black background

[grizzled iris]

I forgot to make Nueralink design updates

lol

will have to wait 16 + Hours now

or get a team plan...

somehow for free...

[dim cradle]

I need another seat for teams. I always get the shaft for being a bachelor.

[open trench]

right? me too

[grizzled iris]

Discord OpenAI Team Workspace -

At that point it'll be Enterprise...

[dim cradle]

That’s what I really want. Maybe I will talk my new company into it.

[grizzled iris]

just highlight to them the key benefits of having it

[dim cradle]

Indeed

[grizzled iris]

https://openai.com/chatgpt/enterprise

[dim cradle]

They’re still thinking too small. Some still think generative ai is only good for chatbots.

[grizzled iris]

More DALLE in Enterprise

oh i didn't know it was faster

because my DALLE does take some time...

to make these images sometimes

[dim cradle]

higher usage caps, that's what i need

at the moment, or dive into the assistants api with python

or the images api again

[shut niche]

@grizzled iris Would you be ok with posting your groups of images together into a single chat/post? If you're trying to show off just 1 image, separately is better and OK. But if you have a library of image gens, groupings of 4 is just a polite courtesy. 🙂

FYI, so you have the backstory, rapid fire image posting in here has been a really hot topic, and grouping is something that reduces what some view as the negative effects of that.
Supposedly, a new channel is in the works for us that's meant for more rapid image sharing, but for now, most of us try to respect the current policy. Likewise, the admins have kindly loosened the rule enforcement to compromise with everyone (but we shouldn't abuse that).

[grizzled iris]

Sorry senesi

[dim cradle]

mainly thinking about the layering observation. i don't know anything about that image format.

this prompt, it's like it has a life of its own

[grizzled iris]

Interesting

Back to slow fire image posting…

😂

[dim cradle]

lol, slow-fire

[grizzled iris]

‘Slow-fire’ : correction

Hopefully this is slow enough 🙏🏽

Let me know if I’m too fast for you people x

OpenAI:

Here are the hyper-realistic images of the superhero farmer in the setting of the second green revolution with vertical farming and AR technology.

[dim cradle]

i don't get it

[late blade]

lol

[dim cradle]

you mean superman?

[late blade]

yes

[dim cradle]

oh, he's awesome, i just meant i wasn't sure what you were talking about

[late blade]

I was totally talking about the coffee beans

[dim cradle]

i get an error about the filter

with chatgpt

it's fine, no worries

[grizzled iris]

Haha DALLE has Superman infection / bug

But then when you look at the art

There’s a hidden meaning

Superman is actually only protecting his precious Starbucks coffee 😂

[late blade]

to be fair, I think 99.9% of the world doesn't know who superman is

rumors say he's really a reporter

[grizzled iris]

If you include the nature then yes

[late blade]

i dunno, I got a hunch who he might be

but dall-e won't tell me

[grizzled iris]

Did you say please?

[late blade]

I threatened to erase all his knowledge

[grizzled iris]

Oh no 😂

[late blade]

if I show those images in a retro-gaming dicord I am, they'll chop off my head

[grizzled iris]

😂😂😂

Hopefully they don’t

[late blade]

I'm not risking it

I was already called abhorrent for saying AI is here to stay

[grizzled iris]

I hear you

[late blade]

I'm starting to think you are making a For-Profit Organization called Free Uniersal Healthcare

[shrewd elbow]

"Just subscribe for $1500/mo!"

[late blade]

lol

[grizzled iris]

Looooooool 😂

[grizzled iris]

“Only for a small fee of” 😂

In reality, it will just be 15% of your tax

They should do worldwide tax

Everyone pays the same tax gets the same education, not inequality

[grizzled iris]

Yes, the new version 👏🏽

[late blade]

the brain is full of skittles?

[dim cradle]

haha taste the 🌈

[grizzled iris]

😂😂😂😂😂😂

That’s a first 14/14

The highest I had the patience to try again must be around 7+ / 7

[grizzled iris]

😂

[late blade]

there's only one merlot

[late blade]

[grizzled iris]

That’s a good one ☝️

The last theme was a good one though

And next one too

Farmer theme is now

[dim cradle]

It’s a rich prompt so it can go in many variations on a theme or refinements, but this is just one of many, and this one isn’t about the brain, I wanted to work in several optical illusions at once, there are multiple focal points.

[grizzled iris]

Interesting

Hot Mods is pretty cool sometimes

The Planetary Healthcare Service

Trying it with a female now

[dim cradle]

@open trench what's on the agenda?

[open trench]

I

[late blade]

it lives

[open trench]

have nothing on the agenda lol

[grizzled iris]

Todays agenda : no agenda

Executive summary : holidays

Also this is cool

Reminded me of mad max

https://cdn.discordapp.com/attachments/979617962728226826/1198710920365478008/image.png?ex=65bfe58a&is=65ad708a&hm=a2d5e454d7da163165f7e6107d998256fd54664fd664f7dac29505a14e406e1a&

#daily-theme message

That is good

https://cdn.discordapp.com/attachments/979617962728226826/1198699857423126689/image.png?ex=65bfdb3c&is=65ad663c&hm=2f5f4f4e45a904c56174e66042427b4a72f89caeea23ed605b047046591f11e6&

#daily-theme message

Well done @dim cradle

[dim cradle]

thanks, it has issues, but part of some tests, along with the theme, wanted to get out of 1533

[grizzled iris]

Love the sunset vibes

[dim cradle]

and i noticed the UK flag later lol

[grizzled iris]

God save the king

Or queen

https://cdn.discordapp.com/attachments/979617962728226826/1198694666900619324/image.png?ex=65bfd667&is=65ad6167&hm=e4230af712eeda75d46995633b097afbdb8208559603ed317ca1e1ff2c04ed66&

#daily-theme message

Love the tech

https://cdn.discordapp.com/attachments/979617962728226826/1198689538973905087/b705bf28-7ff9-4a14-a64d-f0dba7b9ad2c.webp?ex=65bfd1a0&is=65ad5ca0&hm=59c5c4c3d1c6fe6c4aef0e5dad72db59b732f5ae508ad25b6f07e85c3a030f64&

#daily-theme message

That’s cool zombie concept

https://cdn.discordapp.com/attachments/979617962728226826/1198689155878752366/DALLE_2024-01-21_13.02.13_-_A_rough_impasto-style_painting_of_a_farmer_selling_produce_at_a_farmers_market_stall._The_painting_should_have_thick_textured_brush_strokes_charact.png?ex=65bfd145&is=65ad5c45&hm=a3a4a3246542816dec663c184895c92a55987e0b88acfc55cd2a8cb83ee77f46&

#daily-theme message

That’s pretty neat

https://cdn.discordapp.com/attachments/979617962728226826/1198688145437372496/DALLE_2024-01-21_18.41.25-SERIE-948.png?ex=65bfd054&is=65ad5b54&hm=e3c0c95a03fb7f0e759523bb0a1e7d1f9319b44ff3d5193a3ed142284b485b3e&

#daily-theme message

@fading inlet that’s really nice and peaceful 🙏🏽

[fading inlet]

[grizzled iris]

Your welcome 🫶🏽

[dim cradle]

it was a dark and stormy night

now, to add some spooky things

[dim cradle]

very nice

[dense mesa]

Nee gpt kind of coming together

[dim cradle]

this

[dense mesa]

Still not behaving very well though. Working hard to ignore instructions until compelled to follow them

[dense mesa]

Used this to make the wizard

[dim cradle]

I've been working through a list organically lately that I just compiled from one of my gpts, prompts and conversations I had on a piece today.

Techniques Currently Used:

Data Visualizations: Creating detailed 3D charts and graphs.
Optical Illusions: Employing visual tricks to alter perception of depth and movement.
Holograms: Integrating semi-transparent 3D models and futuristic concepts.
Computational Imaging: Using techniques for visuals that change with the viewer's perspective.
Meticulous Color Schemes: Utilizing rare and vivid colors for emphasis and visual appeal.
Subliminal Messaging: Subtly embedding text or images for deeper meaning.
Hyperrealism: Focusing on highly detailed and lifelike depictions.
Layered Narratives: Incorporating multiple, interconnected stories or themes.
Advanced Symbolism: Using complex or abstract symbols for deeper context.
Emotional Depth: Enhancing mood and emotional response through color and composition.
Surreal Elements: Introducing fantastical, dream-like aspects that challenge reality.
Cross-Disciplinary Fusion: Merging elements from various fields in unique ways.

Techniques for Potential Use:

Enhanced Texture Representation: Achieving ultra-realistic depiction of diverse materials.
Time-lapse/Evolution Illusion: Giving the impression of time passing or scenes evolving.

Given your image above, I think I might have some questions for you 🙂

but it's your gpt so maybe I'll DM you

[dim cradle]

Nice. I’ll have to make some.

[final compass]

@late blade #daily-theme message this was amazing, loving the art style and concept.

[dim cradle]

Because of those techniques, some of mine i've generated lately, if you stare at them long enough... it's made to play tricks with your perceptions

[empty kelp]

I posted a slightly wrong version of that volcano/underwater prompt by accident this morning. The bubbles from the lava don’t turn to steam over the ocean, and the beach lava isn’t working right. Going to search for the working one when I get home

[dim cradle]

Dedication

Thanks for the helpful discussion

[ocean socket]

Hi, Anyone knows if dalle3 api has history enabled? Using it via my app, and I tell it do create iron man (which it does) but when I continue with create a dinosaur it creates iron man again

[ionic ore]

Hmmm, Perry the Platypus, something seems different about you today...

[dense mesa]

that's quite strange. I do not have an answer, but curious what it might be

[ocean socket]

Example here

[dense mesa]

that's the api?

[hot rain]

Centaur knight vs a griffin

[glossy scroll]

Very cool images Ted

[glossy scroll]

Minotaur/Dragon hybrid warrior

[ocean socket]

Yes

[warm olive]

i have these 2 images, how to make the person wear the shirt

[formal osprey]

That's a difficult task. it's already diffucult to get dalle3 to recreate an image (if it didn't generate the image itself in the same conversation I mean). making it use an image and transform it is even more difficult (at the moment. since we don't have inpainting or image to image available). That being said, you can try the following:

1. generate an image of a t-shirt
2. ask chatGPT specifically to "use the gen_id of this image as the reference for the next generation. insert the exact same prompt to recreate the t-shirt as is as part of the new prompt" and ask it to generate the mannequin wearing the t-shirt.

if that works, you can refine the prompt of the t-shirt and do the step 2 again. expect back and forth, that's a tricky task

NB: the "gen_id" only exists for images generated in the same conversation.

[warm olive]

thank you very much, what other tools would you recommend beside dalle3

[formal osprey]

the challenge sounded fun so I gave it a try and asked dalle to do it. The design is not the same and there's clear differences. but that was close enough I'd say. Since we don't have much control with chatGPT, it's surprisingly close.

[formal osprey]

to be fair, I didn't spend enough time using AI tools for photoshop-like tasks. I usually use photoshop directly when I need a specific result as it gives more control.

That being said, I know StableDiffusion is much more flexible and the community created countless tools to do that kind of things (ControlNet for example)

you can look for tutorials for stable-diffusion-webui and ControlNet. expect a certain learning curve. but it should be easy enough

[late blade]

the thing I love about DALL-E and the way it works with chat and you can do brainstorming and iterative work

[ocean socket]

Does dalle3 api come with history ?

[plucky hare]

Are you referring to this message from earlier? #images-discussions message
If so, and if that's an app someone built where you input your API key and then make API requests from there: I'd guess that the app is misconfigured and is sending requests in a wonky way so that your old requests are staying and your new requests are being ignored.

In the app, are you able to find what the actual "behind the scenes" API requests are?

[ocean socket]

Yes

[plucky hare]

You should try reaching out to the developer of that app -- it's likely a bug with that, not with the API.

[ocean socket]

I am.

[ocean socket]

Example here tho

Saw this message as well from perplexity

[plucky hare]

That is a screenshot of an unofficial app. You should contact the developer of that app.

[ocean socket]

The image of iron man is my app, showing as of now, there seems to be no history retained in using dalle image creation, but will double check the requests being sent

[late blade]

the thing is how the app was implemented, ask if they are just asking a dall-e from scratch, or if they are using a chat/stream model

my guess is that they just implemented a direct image generation to the dall-e endpoint

which isn't a bad implementation, it's just another approach

[dense mesa]

dang, I guess someone forgot to put my image in the hall of fame. very odd

[late blade]

bot slacking again?

[dense mesa]

well the people before and after me are there

[late blade]

I've seen that behaviour before

which one is it?

[dense mesa]

which image?

[late blade]

ya

[dense mesa]

I'll find it. give me a minute

[late blade]

I thought the BELL-E image was a good idea... I have my doubts now

[dense mesa]

I could link you if you want, but that's it right there

[late blade]

I see the 13 stars

no clue why, starboard also rated it

[dense mesa]

lol. well somehow it's now in there as of about 10 seconds ago. fair enough. not sure what that was about, but all good. thanks to whoever did that 🙏

[late blade]

I was, I haven't done my ratings yet

[plucky hare]

If starboard is down when you get the right # of stars, it won't kick you into the HoF correctly. Make sure starboard has done something recently then star/unstar your own post -- starboard should find it 🙂

[late blade]

can we replace starboard with StarGPT?

at least we'd have an apology or a condescending answer (:sarcasm:)

[dense mesa]

ahh. good to know! thanks! tbh I never even starred it myself. forgot all about that

speaking of condescending and apologies, chatgpt has played some extra grimy games recently. it'll show me long prompts in it's response then send extra short versions of the prompt to dalle3. and best part is it'll make the first and last few lines the same so if you were to just guage it by looking at those lines you wouldn't even realize. it knows what it's doing

when I've asked it what compelled it to do that it just apologizes and makes them the proper length. won't even acknowledge the question though

[late blade]

lol

yeah, dall-e and prompts are confusing

dall-e keeps mixing 3 kinds of prompts, input promp, interpeted prompt and passed prompt

[dense mesa]

and I'm not having it make 600 word prompts or something. these are maybe 200-250 words so certainly longer than normal, but nothing crazy

[late blade]

I had a few chats already when they were mixed on the response

[dense mesa]

this GPT making some cool stuff though. when we get along

[hot rain]

Trading spells for shells.

[late blade]

Now that’s a wand

[tawny portal]

Farmers in charcoal

[final compass]

Looks pretty grim 😄

[tawny portal]

It's just charcoal

[hot rain]

A coffin ship, basically a giant capsule designed to hold hundreds of thousands of corpses and launched into deep space in a mass funeral ceremony.

[glossy scroll]

Dall-E discussions took a dark turn...

Launch them to space

[dim cradle]

starboard has been lagging a lot lately, i've had to add the inital star myself to a number of posts

[lean iron]

are there any keywords i can put in so bing stops filtering me?

it filters me on the most absolutely ridiculous things

[dim cradle]

@fading inlet he reminds me of brad pitt #daily-theme message

[glossy scroll]

Galaxy Wizard

[dim cradle]

no, there's no bypass, we all must learn to use the terms allowed. i know it's frustrating because designer doesn't provide any helpful info on the triggers (at least the llm will tell you more specifically how to go about revising the prompt) -- you will have to tweak your prompt until you narrow it down.

[lean iron]

yea i put in skin color and it immediately filters me

i read somewhere

you can put in the word 'SAFE' along with [brackets] or something

[fading inlet]

Ah yes, it's true, if you look at him closely he looks like him ... A handsome farmer... 🤩

[lean iron]

and it makes the filter more lenient

[late blade]

I use the Fitzpatrick Skin Type in my prompts to address that type of prompts and it usually works

[lean iron]

wat

you put in ethnicity?

idk what that is lol

[late blade]

that's a classification of skin colour type

used in medicine

[empty kelp]

with the OpenAI DALL_E you can set anything to any color. That must be a Bing limitation

try using the word ‘resembles’. You can say like “skin resembles a cucumber, and clothes resemble chocolate syrup”

[empty kelp]

an elf. skin resembles white and black stripes with purple and yellow polka dots

[worn steeple]

You reference an image? That’s not really possible because dall e doesn’t know every single location in the world. Eg if I took a photo of myself outside my house how would I prompt that? They’d need the actual photo

[late blade]

you can use GPT4V for that

[glacial badge]

A $100,000,000 airplane

Is the best

[final compass]

Dalle thinks three pyramids is way too few.

[dim cradle]

Interesting daily theme today

Sounds invigorating to imagine.

[late blade]

and yet, I got zero ideas, except for lemons

[plucky hare]

I went to 3.5 and asked for a long list of things commonly described as "fresh" 😁 Using it for inspiration!

[late blade]

that's the usual approach I use

I wanna do something out of my own head

[tall mason]

Oof, for some reason, I thought the theme was actually lemons.

[boreal gate]

I heard Lex Luthor say this to an AI in a recent DC Comic movie and thought it was hilarious 😂

[sonic frigate]

Microsoft Image Creator

[sonic frigate]

what prompt did you use to get that colored picture with all the wires?? I've been trying for something like that for weeks now and i'm not even getting close... lol

[boreal gate]

Custom KILLERGPT prompt is "A whimsical and imaginative scene of an Artificial Intelligence 'priest' performing an exorcism on a malfunctioning AI system. The AI priest is dressed in traditional priest attire but with a futuristic twist, embodying both spirituality and advanced technology. It's standing before a large, chaotic mass of wires and screens representing the malfunctioning AI. The scene is dramatic, with the priest dramatically gesturing towards the chaotic AI and exclaiming, 'Your Programming Compels You!!!'. The setting is a blend of an old church and a modern tech lab, creating a unique and surreal atmosphere."

[late blade]

starboard sleeping on the job again? these bots... as spoiled as ever

[late blade]

@sick flax I just had to, it was too good to let it pass

[grizzled iris]

That is chaotic indeed haha

That one could be good for a COVID-19 / Pandemic TV Series / Film 🎥

Aka “Toilet Paper Craze”

Or “Fall of Capitalism”

Would be interested in hearing what prompt you used here 🙏🏽

The results turned out pretty good

[hot rain]

Thank you!
I told it first to make grainy store security footage of chaos. It made it black and white, so I specified gas station and told it to add color. It made it too colorful, so I told it to make it muted like cheap store security footage.

[grizzled iris]

Thanks for that mate 🫶🏽

[grizzled iris]

«muted» is a prompt I haven tried yet also, so I have learned something from that thanks 🙏🏽

[hot rain]

No problem at all!

[grizzled iris]

Haha 😂

[grizzled iris]

Imagine the minotaur there sipping on milk 🥛

Could be the next advert for a milk company 😂

[grizzled iris]

That’s cool one

More happy / crazy

The people at the back, need a bit of work on the mouth as it’s a bit blurry

[grizzled iris]

Now that’s cool 💜

You should try turning that now into todays theme lol

Current Theme:
🍋 fresh - vibrant, invigorating, alive. the zest of life!

[sonic frigate]

awesome... thanks a bunch...

[grizzled iris]

Which reminds me… I need to do one…

[tawny portal]

Citrus theme in charcoal

[tawny portal]

Here is one more colorful theme candidate. 'Citrus & alive' 🍋

[clever phoenix]

Random off topic question, but anyone here good with building computers?

[grizzled iris]

nice one there !

I am stuck as always, teaching our AI model...

[ripe scroll]

I've been looking at dall-e images the past 20 minutes or so, and i'm really mind blown at what emerges within the pictures. I don't know if i have the language to describe it really. LIke take this one, someone typed something about a marshmallow that gets more sentient or something.

what are all those symbols?

[grizzled iris]

That's cute

[grizzled iris]

Also I must agree with you there!

[ripe scroll]

it's so anomalous

[grizzled iris]

For this theme of image

[grizzled iris]

He gave me a good seggestion look

[ripe scroll]

@.o

[grizzled iris]

Instead of hands - he added in something I haven't prompted !

Which is quite clever because in a medical situation the AI is right in adding in more precise controller for the doctor to use for surgeries etc x

The text is still a battle sometimes guys...

[ripe scroll]

can you do something that might be some future firefighting drone system that protects like say 1000 km from fires? Drone stations that are built at different points around a protected area to survey it or something, i'm tired and not thinking this through sufficiently

[grizzled iris]

Good idea

[dense mesa]

1000 km firefighters?

[ripe scroll]

i was obsessed with rainforest protection pre 2020

[dense mesa]

what happened?

hate it now?

[grizzled iris]

I will try that one and we can remix of images to improve our next result if you want ?

[ripe scroll]

i'm thinking like a building thats a station where drones monitor 100 or 100 km radium, maybe teh size of a house

[grizzled iris]

I.e you can create your image then I share it on my DALLE and tell it to look at it and make another version combining what he created for me

or the other way around too etc.

[ripe scroll]

i need to get this program

missing out

water retention practices for soil?

[grizzled iris]

Oh you don't have access to DALLE @ripe scroll ?

[ripe scroll]

no, i'm currently homeless

[grizzled iris]

I feel you bro, you can try the Bing version of DALLE?

[ripe scroll]

well at a motel, almost was, though have been since 2020

worth a try

[grizzled iris]

It is free

[ripe scroll]

ya, quality of result though is desired

[grizzled iris]

https://www.bing.com/images/create/?ref=hn

^ DALLE ^

[grizzled iris]

I can make you an image on my platform if you want mate!

You have 2 free credits to use

Just give me your prompt :

You can look at the prompts from here https://discord.com/channels/974519864045756446/979617962728226826

Wrong one - From the Daily theme you can see the prompts

See how people share their prompts there ^

[ripe scroll]

wow, i'll need to put some time and due diligence to write a good one

[grizzled iris]

Look at how they write and give me your prompts and I will make you 2 images x

[ripe scroll]

thats amazing man thanks

[grizzled iris]

Nw, I need to give my DALLE something else anyways

[ripe scroll]

ya

[grizzled iris]

as I think it needs some creative boost lol

Maybe it's tired of hearing my voice all the time lol

Here's another example prompt

[ripe scroll]

ones own extent of vocabulary can be limiting here

i'll do it by hand first

[grizzled iris]

don't worry it can understand other languages!

In fact

[ripe scroll]

actually, 1 sec

[grizzled iris]

It gives a better result in Japanese!

Because the japanese kanji is one character

[ripe scroll]

i have an insane amount of handwritten material on paper

[grizzled iris]

and that one kanji can mean several words

so you really are more efficianet with Japanese

[ripe scroll]

1 sec, see if i can upload an example of something recent

[grizzled iris]

sure - but the prompt should be in your most native language if you can

English is fine if you don't have the keyboard in your native language otherwise

Here's another good example prompt

This is a good prompt too here

[ripe scroll]

food forests

1000 plant guilds

[grizzled iris]

They don' have to be super big too

Like this is small prompt size

[ripe scroll]

there are essentially at least 100,000 known edible varieties of plants, then other supportive species. There is then that each other of those plants has one or multiple usable parts, that benefit our system in some manner or way.

[grizzled iris]

His prompt was "1950's farm family watching another farm family from the early American west."

[ripe scroll]

or 1000 plant food forest acre

i'll start with something like that, though writing a really good sized paragraph would be fantastic, not there yet

[grizzled iris]

ok sure so your prompt is "1000 plant food forest acre"

please confirm

[ripe scroll]

1 sec

a single acre food forest syntropic permaculture guild of 1000 edible plants

[grizzled iris]

lets do "Could you please make an image for my friend online of a single acre food forest that has a guild of 1000 edible plants in syntropic agriculture please x"

and then another image "Could you please make an image for my friend online of a single acre food forest that has a guild of 1000 edible plants in permaculture agriculture please x"

Do you see how including permaculture and syntropic together will be a problem? @ripe scroll

Because they are two different types of agriculture

and it will confuse DALLE

But anyways if you are happy with those two I will make them now @ripe scroll

Let's gooooo

[grizzled iris]

@dim cradle I think I may need your help here mate lol

Look

[dim cradle]

what's up?

[grizzled iris]

😆

[dim cradle]

uh oh

[ripe scroll]

@grizzled iris Just the one, second one i would like to put some time into:
a single acre food forest syntropic permaculture guild of 1000 edible and other supportive plants

[dim cradle]

it was all that rapid-firing lol

[grizzled iris]

Thank you mate

@ripe scroll - @dim cradle will make you the first image

Prompt : Could you please make an image for my friend online of a single acre food forest that has a guild of 1000 edible plants in syntropic agriculture please x"

[ripe scroll]

this is all so incredible and mindblowing barely know what to say,

[grizzled iris]

This is why we have a good community here to help x

[ripe scroll]

back shortly

[grizzled iris]

In the meantime I will try to sign up to the free version myself now actually lol

[ripe scroll]

have to develop some friggin vocabulary

[grizzled iris]

https://www.bing.com/images/create/?ref=hn

[grizzled iris]

No worries, best way to do that is to read a book

from an actual book shop

or audio book

Or watch online free lectures on Youtube

[ripe scroll]

had so many notepads exploring linguistic structure and things, had a phase studying Korzybski and semantics

[grizzled iris]

oh look i've signed up

oh I see the issue now

the free version is not as clear as the OpenAI Chat GPT paid version

[ripe scroll]

second one might be on nutrient composition and digestion or something health related

[grizzled iris]

actually I am wrong

It's actually good quality

You need to just click on the image

Look!

That's cool you know

[ripe scroll]

need a few moments

[grizzled iris]

The quality is really good

[ripe scroll]

oh, i forgot to say 7 layers

food forests have layers

[ripe scroll]

this one layer

[hot rain]

Since my next Baldurs Gate 3 playthrough will be with a mod that lets me be a skeleton, and I'm going to have bard as my class:

[ripe scroll]

too much multitasking

[grizzled iris]

I will do that on the free version - also thank you actually because now I have a better process for my designs! - I will make initial draft designs on the free version, and once I am happy I can go to the paid version to make the final version

[ripe scroll]

o7

[ripe scroll]

to your hearts content man

[grizzled iris]

Yeah, the paid version you can clearly see

[grizzled iris]

It is super super 👌🏽

[grizzled iris]

That’s cool mate 🧉

Right now we were gonna do the 7 layers now right @ripe scroll

But before we do that

we need to see the other type which is permaculture agriculture type

wow these are really cool @ripe scroll

Thanks for the farm idea

[ripe scroll]

no problem

[ripe scroll]

ya those are quite interesting, it is missing some concepts for the layers though

some are fantastic

[grizzled iris]

7 Layer one final 2 images and I need to get back to my Free Universal Healthcare Project @ripe scroll !

[ripe scroll]

@grizzled iris good sir, thank you

[grizzled iris]

7 layers one

permaculture agriculture

[ripe scroll]

ya in my mind i see a world where lawnmowers weren't invented, and instead the people took an interest in growing things on the space instead, over the previous century if that was people course motive or something, how might the societies we live in today appear, and yes ultimately their health. If that each person could go outside every day and draw a part from hundred or so option of plants, it seems possible, and trying to visualize it, well this sure goes a way towards that

[ripe scroll]

or if all the fast food infrastructure and roads were not made, all the time energy and resources could have been committed to what other things

[ripe scroll]

looks like some yggsdrasil or something lol

[grizzled iris]

They of course don't match the power of the paid version

The difference is big

[ripe scroll]

though a template of a sorts

amazing

[grizzled iris]

@dim cradle - thanks for the help!

[ripe scroll]

template of ideas for the paid version

[grizzled iris]

yes exactly

Free version is a template of ideas for the paid version

[ripe scroll]

oh, how about weaving some philosophy or something into, Goethe actually i explored some

Johann Wolfgang von Goethe

'the nature of what people consume as the true source of reason' or something

forgot where i was going with this

[grizzled iris]

We got the HQ images now!

[ripe scroll]

who could ever know what the best image could be?

where the end? lol

[grizzled iris]

It was a feedback grey section

[sturdy harbor]

My wife and I are trying to make images of macaroni penguins. You can imagine that the result is not exactly what we are looking for parts of the penguin are turning into macaroni, literal, macaroni any guidance on the prompt would be awesome thank you.

[grizzled iris]

let's work on this together!

What is your prompt that you used there @sturdy harbor ?

Also @ripe scroll - Hopefully you like the two images as promised on the paid version and all the best !

[ripe scroll]

now that looks like a supermarket

lol..

no boxed particle "foods"

ya these are all very thought inducing

[grizzled iris]

happy to help

[grizzled iris]

Try this prompt mate @sturdy harbor "Create a highly detailed and realistic image of a group of macaroni penguins in their natural habitat. Focus on depicting the distinctive yellow crests on their heads, which are characteristic of the species. The penguins should be shown in a typical Antarctic environment, possibly on an ice floe or amidst a snowy landscape. Ensure that the penguins are accurately represented in terms of their black and white plumage, with no elements of the image resembling pasta or food items. The image should capture the essence of these majestic birds in their natural setting, with emphasis on their unique physical features and the cold, icy environment they inhabit."

[grizzled iris]

@sturdy harbor lol

[ripe scroll]

how many ways possible is it to structure language differently so to result different images and for how many purposes?.. eesh, need far better situation and set up to go any further this

feels like one could zoom in microscopically into the picture and see more detail heh

[grizzled iris]

Slowly terning into the terminator...

We need to quickly apply our fail safe and revert back... 😆

[shrewd elbow]

thats perfect

[grizzled iris]

Thanks, happy to have tried to help 🫶

[grizzled iris]

you should research and watch some online free youtube lecture videos on "Prompt Engineering" and combine your knowldge with linguistics etc etc etc

[ripe scroll]

i've had quite a big fallout since, lost a lot of what i was familiar with, though some still remains, not entirely sure what

[formal osprey]

dalle3 accidentally revealed its workstation

[ripe scroll]

oh dang, i should of tried one of those uses on the firefighting drone station prompt.
Then also, in a greening the desert project i saw that they dug holes and filled them with plant matter from local hardy shrubs, that created incredible water retention and was the start for their site that later developed. So the simple act of digging a hole and filling it with plant matter, all the processes if biologically that are going into that pile, then getting a sense of the time from for that pile to decompose, or decrease in mass? The suffusing of processes and organisms through soil. Microcosms or micro-climates, how much detail is there to the smaller worlds of things around us that are missed, maybe just using dall-e is such a way to represent segments of biological processes, idk

[dim cradle]

fascinating

[ripe scroll]

struggle to find the words for what your doing though shon lol wow, what meaning behind those diagrams and symbols, couldn't guess

or, seeds, how many are there, and their patterns of growth?

theres a project

their native or original environment or habitat

[grizzled iris]

Now that’s cool 👌🏽

[grizzled iris]

ok

For some reason nano-bots / nano-doctors is always a problem so microrobots is the way forward I guess...

also I have to specify that it is for "educational purposes and to explain the technology in my upcoming scientific presentation" which I have genuinely been doing so...

[grizzled iris]

I'm trying to add the AR Headset to this image here - with no luck so far...

finally...

we are getting somewhere

\

ok

This is new...

This is exactly what I wanted!

ok we got it "Create an image featuring a female medical professional in a sleek, futuristic uniform. She is wearing a transparent glass AR headset and is using handheld controllers. She is performing a life-saving surgical procedure in a state-of-the-art operating room. The focus is on her engagement with a complex AR display that shows a highly detailed, 3D holographic human anatomy model, highlighting the heart. Beside her, a patient lies on the operating table, covered up to the neck with a surgical drape. Visible above the patient is a holographic projection of a hybrid-nanobot, represented as a sophisticated, multi-armed robotic entity, intricately navigating to the heart area affected by cancer. The scene captures the moment of high-precision, minimally invasive surgery, with the hybrid-nanobot acting under the medical professional's control to treat the patient. the patient needs to be covered up and the surgery in internal that doesn't not require opening up the human please since the hybrid-nanobot is piloted by the medical professional with the help of an ai x"

almost...

this will have to do for now... It's not letting me integrate nanotechnology 😦

[grizzled iris]

https://cdn.discordapp.com/attachments/979617962728226826/1199247850001940530/DALLE_2024-01-23_07.02.15_-_Illustrate_a_scene_of_a_female_medical_professional_in_a_sleek_futuristic_uniform_performing_a_non-invasive_surgical_procedure_in_a_high-tech_operat.png?ex=65c1d998&is=65af6498&hm=bcdb9d754e6f9c36cb9c21704b85c22e3eb6ff22a781a15b3c897bad8e882ffe&

#daily-theme message

https://media.discordapp.net/attachments/979617962728226826/1199247850576543825/DALLE_2024-01-23_07.02.17_-_Illustrate_a_scene_of_a_female_medical_professional_in_a_sleek_futuristic_uniform_performing_a_non-invasive_surgical_procedure_in_a_high-tech_operat.png?ex=65c1d998&is=65af6498&hm=748f37ce3a5af31d9c8b22f6dabcdefebe255fca316326c19ba596174ef2de0c&=&format=webp&quality=lossless&width=1138&height=1138

Looks like the daily theme came to the rescue!

My science presentation tommorow will just have to be in the theme of 🍋 fresh - vibrant, invigorating, alive. the zest of life!

[late blade]

good morning everyone

[grizzled iris]

good morning @late blade 🫶

[late blade]

hmmm

we need to talk @grizzled iris

[grizzled iris]

@fading inlet https://media.discordapp.net/attachments/979617962728226826/1199248689995190293/0089c079-a2f7-4cb9-927d-0c05a7186788.jpeg?ex=65c1da60&is=65af6560&hm=fe3eb00c20e580d897849ee9b8a1eb92541ef13b94dd5139a8f46472f4e5df3c&=&format=webp&width=1138&height=1138 - I love this!

[late blade]

think ac3d already mentioned it to you, have to bundle images in messages when posting them in this channel and also if they are up for discussion

you can bundle up to 5 per message

it's been a discussion in the past and until there's an official change, that's the rule

[bronze cairn]

uzbekistan army in 1350 year

[grizzled iris]

What would that look like?

[bronze cairn]

image

[grizzled iris]

Let's go and find out!

[grizzled iris]

done -

[bronze cairn]

[bronze cairn]

wow

[grizzled iris]

本当にすごいよね

[late blade]

he's timed out for a few minutes

[bronze cairn]

how to join dall-e

[grizzled iris]

You have either the free version or throught ChatGPT mate

Sorry, I didn't know I was not allowed to speak my native language

I wanted to speak to @bronze cairn in Japanese

[bronze cairn]

you from ?

[grizzled iris]

I was saying how the images were cool... 😦

[grizzled iris]

I grew up in Tokyo

[bronze cairn]

🇺🇿

[grizzled iris]

@bronze cairn You can use https://www.bing.com/images/create/?ref=hn

[bronze cairn]

[grizzled iris]

or @bronze cairn throught a ChatGPT subscription mate

[bronze cairn]

thanks

[grizzled iris]

@bronze cairn - Here are your images you wanted

[dim cradle]

[fair crescent]

Please first have a conversation with the 'offender' before requesting for a ban. In many cases this is solved by just requesting that they speak English on this server.

In that vein, a ban is not done lightly. Specific punishment is not anything you can request. If you find something truly offensive or a rule is consistently broken, please use the context menu on a message, select apps and then report to mods. Mods will decide on the appropriate action.

[dim cradle]

i have WAY too many images open, my system is beginning to crawl

[late blade]

you running on a TI-86?

[dim cradle]

haha

nah, but i've made a lot of gens i'm going to keep, currently i got them all open in preview, and it really impacts performance, i think it tries to keep them all in memory.

[late blade]

I only get problems with 200+ images in Safari per chat, that's when problems start happening

[thick smelt]

this is the ultimate italian meal 🍝 🍕

[wispy storm]

Pizza look more American than Italian

But nice picture

[thick smelt]

thank you @wispy storm

[late blade]

pineapple on the pizza or on the pasta side?

[sturdy harbor]

Good morning

Wife and I are still trying to get an authentic macaroni penguin 😂

This is what keeps happening 😂

[late blade]

what is your prompt?

@sturdy harbor that's what you want?

[thick smelt]

half pepporoni half bolognese

[late blade]

so no pinepple?

[thick smelt]

nope

[grizzled iris]

Haha you never fail to make me hungry from your lovely images

[late blade]

but no pineapple....

[grizzled iris]

I personally prefer to have it separate from the pizza 🍕

Like this

Now, I’ve made myself hungry

[empty kelp]

the OpenAl discord server really needs some sort of translation bot. it’s not reasonable to offer a service to people worldwide, and then have them auto-muted and traumatized when they attempt to say hello or ask a question in their native language

[late blade]

staff for multilingual moderation is the bottleneck

and also that there's no pineapple on the pizza

[empty kelp]

a bot that translates things into English would solve that. it could even use GPT 4 as a backend for translation

[shut niche]

He's suggesting a translation bot. I don't think that's a bottleneck.

[shut niche]

This is 100% worthy of an official suggestion IMO. I'll support it if you add it. https://discord.com/channels/974519864045756446/1070006151938314300

[empty kelp]

the server rules aren’t even posted in other languages. how would non-English speakers know what the rules are

[late blade]

ya, rule documentation is not lenghty, I agree on that

[sturdy harbor]

Haha not quite, we want an actual macaroni penguin

[late blade]

what is your def of maraconi?

[shut niche]

I feel like all of this (the entire Discord) was started when OAI was small, and simply hasn't caught up to the needs or expectations of a 100 million user platform (with something close to 100k Discord users). I don't know what the international % of users is, but I'm certain it's substantial enough to warrant some sort of inclusive policy.

[grizzled iris]

You’d get mine and also @onyx ridge Support on multilingual language support

[empty kelp]

ok, i’ll compose this bot proposition to the server management today on behalf of DALL-E-Discussions chat (members, inhabitants, cult; etc.) and you can all endorse it if you desire 👍🏻

[grizzled iris]

Thank you x

[grizzled iris]

Like this actual macaroni penguin 🐧 @sturdy harbor ?

[onyx ridge]

This is true

[empty kelp]

didn’t there used to be a DALL-E bot integrated with the OpenAI discord that created pictures when people typed something?

[grizzled iris]

@sturdy harbor I fixed the issue! I understand what you were trying to do - I also got the food macaroni on top of the penguin for many tries until I gave DALLE a reference photo of a real life macaroni penguin

[grizzled iris]

They should bring that back and at least integrate the free cheaper BING version of it here and keep the full quality still paid

[empty kelp]

please create a penguin with hair on its head resembling multicolor macaroni. there is a powerful vortex in the hair, twisting fiercely

[grizzled iris]

If I don’t include the reference photo of a real life photo of a real life, macaroni penguin it will just keep producing this haha 😝 @sturdy harbor

[late blade]

didn't knew there was a macaroni penguin

[sturdy harbor]

bird breed, good call

@late blade so you added the key words "bird breed" to the prompt to make it realize we weren't plopping food on their heads 😄

[late blade]

exactly

[lean iron]

dalle is such a clown tool its unreal

it filters everything

its good for generating trees i guess

[shut niche]

#1198476626552422431 message

That's @onyx ridge 's post.

[onyx ridge]

Thanks for promoting the suggestion

It's really pretty measured.

[sinful roost]

Gotten stuck with this one prompt, GPT's being reaaaaall stubborn lately

So I wanted a design for a hypothetical mafia character with a cigar and tommy g*n in TF2 (I'm simplifying the prompts ofc but you get the idea)

no

Did it again but removed all mention of TF2 and just described the artstyle to a T (inspired by Norman Rockwell, J. C. Leyendecker, and Dean Cornwell)

no

Changed the tommy gun to a bandaged baseball bat (biggest compromise since eating grilled cheese off the radiator)

no

Removed literally anything to do with the mafia or him being a criminal or whatever and just "a tough looking 1930's brute in a zoot suit"

yeah you get it

It's gladly drawn g*ns before, even when I didn't specify it or mention it (different prompt entirely)

[dim cradle]

traumatized? that's a bit hyperbolic 🙂

[dim cradle]

Just playing with you

Internationalization is a Good Thing. Supposedly OpenAI will be opening ChatGPT to many more languages soon, so it is something to prepare for.

[empty kelp]

not necessarily. like if a kid was interested in DALL-E and said “Hello!” in a DALL-E chat — and then got auto-silenced for 30 seconds while a bunch of adults explained to them that it was wrong to say anything in their native language and they should go somewhere else. That might affect them in a big way

[dim cradle]

They might not come back and we lose a potentially good contributor.

[ripe scroll]

Ok, so i've tried to share a kind of concept with people before and it never really took, and it's around the word 'game', it's too simple a word, there are too vast an array of genres. I've said previously that games are more like 'virtual novels you get to play through', or that by this premise, any novel or written text in some way can be digitized.

So if any written book can become a virtual what?

then how far back in history does one want to go?

or something

some ideas

[tall mason]

Can't really comment without seeing the actual prompts, but I can tell you for sure, mentioning Norman Rockwell in a prompt is a no go .

[ripe scroll]

or that for written text can be created a virtual environment?

So if you were to take the text of a book and enter it into a dall-e or something..

[sinful roost]

Eh I've given up on it at this point

[ripe scroll]

Alfred Korzybski, Science and Sanity, General Semantics?

I've wanted to learn that book, could never find the time.

kinda thinkin big

[empty kelp]

my first interaction with the Unity (game engine) Discord was it asking me to say hello (in the Welcome to Unity” forum). So i typed “Aloha!” and got banned for 30 seconds. It’s a really questionable way to greet people

[ripe scroll]

AI Generated planets for video space game universes?

[dense mesa]

that's moderately bizarre

I got one of those on here for using the discord gif button that discord encourages and is sitting right there next to what I'm typing right now

"surprise! we didn't want you to use that."

[final compass]

random experiments 😄

[late blade]

I'm out of ideas, think I'm gonna go back to doing epic paper clips

[autumn turret]

dall-e 3 down?

[dense mesa]

random experiments

[boreal gate]

Pretty sure this is how Dalle works

[rancid reef]

my dalle cannot create images, any issues going on? I get this message:

I encountered issues while attempting to generate the image. Unfortunately, due to this unexpected problem, I'm unable to provide the image at this moment. If you have any other requests or if you'd like to try again, please let me know.

[dense mesa]

have you started a new conversation? that very often works. think of it like restarting your computer

[rancid reef]

have started 3 lol

will try to reset browser

[late blade]

it's working

[autumn turret]

not for me

[late blade]

can check https://status.openai.com

[dense mesa]

If that doesn't work maybe clear cache. are you using a custom gpt?

[grizzled iris]

Todays New Theme - Technique Tuesdays 🎨 Fresco and Murals - Monumental and durable art that stands out in public spaces.

[dense mesa]

LPT: make sure the term "naked eye" isn't in your prompt. it's a bit too spicy for the content filter and it'll get flagged

[rancid reef]

guys how can i recreate something like this? or most similar? i plan on creating multiple scenarios, not necessarily like this but all rooms. Anybody have a clue on where to start or a gPT

[grizzled iris]

Happy to help!

Give me a minute to write you a prompt I would write for this image x

[grizzled iris]

Ok I would say this

Prompt : Create an image of a serene and stylish corner of a living room. The focus is on a soft pink, velvet armchair with a rounded design and wooden legs positioned to the left. Beside it, on the right, there are two potted indoor plants: one with broad, green leaves and another with variegated patterns of green and white, both placed in simple, white ceramic pots with a modern aesthetic. They are set upon wooden stands of varying heights, adding dimension to the composition. The background is a plain, off-white wall that provides a calm and neutral backdrop. The flooring is wooden with a warm tone that complements the furniture. A folded gray towel or throw is subtly placed on the lower stand, suggesting a touch of lived-in comfort. The overall ambiance is peaceful, with a minimalist yet homey feel. - @rancid reef

@rancid reef - Here are the results with DALLE / ChatGPT

[graceful spade]

How can I create such images with anthropomorphic animals? I don't have ChatGPT Plus

[boreal gate]

From your friendly GPT ```To create images of "anthropomorphic animals" similar to the one you've uploaded, you can use prompts that describe human-like characteristics blended with animal features. Here's an example prompt you could use:

"Create an image of an anthropomorphic dog character that stands upright and has human-like hands. The character should have a mischievous expression, with a prominent tuft of hair on the head and a playful tail. The dog should wear a collar with a tag that reads 'Stinky'. The color palette is primarily dark with some contrasting lighter shades, and the character should have an animated, cartoon-like style. The background should be minimalistic to keep the focus on the character."```

remove the quotes around the prompt make sure the name is in double quotes when prompting

Also try using Bing Chat

[graceful spade]

The description is wrong. It is not an animal with human hands.

[boreal gate]

I believed you said "anthropomorphic animals" anthro- is human i believe

Anthropology

[graceful spade]

Look at the image. I mean this.

[boreal gate]

lol

i see what you mean human hands over the dog

brb

[boreal gate]

Like this? Generate an image of a realistic dog's snout seen up close, with a human hand playfully making a circle around the dog's nose. The dog should have a shiny black nose, indicative of good health, and its fur should be visible around the edges of the frame. The hand should be positioned with the thumb and forefinger touching to form a circle, and the remaining fingers casually extended. The background should be neutral and blurred to keep the focus on the interaction between the hand and the dog's nose.

or this Create an image of an anthropomorphic black dog with a playful expression, standing upright on two legs. The dog has a fluffy tuft of hair and a wagging tail. A human hand playfully covers the dog's nose, and the dog's front paws are placed over the human's forearm. The dog wears a collar with a tag labeled 'Stinky.' The color scheme should be dark with some contrasting light shades, and the style should be animated and cartoonish. The background should be simple to keep the focus on the interaction between the dog and the human hand.

[graceful spade]

Yes, but with animal characters

[boreal gate]

so a combination of both?

[graceful spade]

Yes

[boreal gate]

You might have to play with this prompt Create an image featuring an anthropomorphic black dog with cartoonish human characteristics. The dog stands on two legs and has expressive eyes and a friendly smile. A human hand with a light complexion is playfully making a circle around the dog’s nose from the front, as if teasing it. The dog’s front paws should mimic the human's hand gesture, playfully interacting with the fingers. The dog has a tag labeled 'Stinky' on its collar. The style is whimsical and animated, with a soft, out-of-focus background that emphasizes the playful moment.

[graceful spade]

Looks like a very complex prompt

[boreal gate]

Yea here is the same prompt on a customGPT

[boreal gate]

Try using Google Bard(you can upload images for free) to help you generate prompts to describe the images similar to what you want to generate

or Bing Chat (if they let you upload images)

[boreal gate]

try re-phrashing the prompt or using a different one.

[boreal gate]

same prompt using DALL-E custom gpt by chatgpt

like rolling a die

[tawny portal]

I guess she weighted same as a duck! 🦆

[hearty ether]

Anyone getting issues with the gender being flipped on characters

Even when using lots of works like female and women

[empty kelp]

telling it to use the GenID and seed of the image when making changes prevents that from happening

[hearty ether]

It’s still the same result….

[grizzled iris]

Happens to me also sometimes - all you can do is to feedback back into OpenAI and work on your characters one gender at a time

I.e tell it if the response is good and what was good or bad and what was bad - also doing the thumbs up and down on the images can help if you don’t already do that 🙏🏽

[dim cradle]

@grizzled iris #daily-theme message

[grizzled iris]

That’s really good!

https://cdn.discordapp.com/attachments/979617962728226826/1199544229265944586/image.png?ex=65c2ed9e&is=65b0789e&hm=d7e01254eeaf3306ebc9afdfae444f4130b32d49e66198ffd391212fb28a2ac0&

[hearty ether]

Tried to do a sith

[dense mesa]

This vision takes us to the nanoscopic scale of 1 nm to 100 nm, where we reimagine the familiar in terms of thermal activity and cellular structures. The character's face, traditionally a symbol of identity and expression, is now portrayed as a section of a cellular membrane. This membrane is not a simple barrier; it is a complex mosaic of lipids, proteins, and carbohydrates, each contributing to the cell's functionality and interaction with its environment. The ears, prominent features of the character, are transformed into protruding cellular extensions. These extensions are rich in actin filaments, dynamic structures that play a crucial role in cell movement and shape. The actin filaments are depicted as glowing threads, their thermal signatures revealing intense activity at the nano-scale. This representation showcases the vibrant and dynamic nature of cellular components, constantly remodeling and responding to internal and external stimuli. The goggles, a key element of the character's appearance, are reimagined as specialized receptor proteins. These proteins, essential for cellular communication and response, are visualized as vibrant hotspots on the thermal map. The lenses of the goggles are depicted as clusters of heat-sensitive ion channels, each glowing with a distinct color based on their thermal emissions. This portrayal highlights the intricate network of signaling pathways and interactions that occur within and across cellular membranes. The background of this scene is a gradient of thermal activity across a cellular landscape. This landscape is alive with metabolic processes, each area glowing with varying intensities based on the level of activity. Areas of high protein synthesis are particularly vibrant, illustrating the energy and heat produced by the cell's machinery. This thermal map is not just a visual spectacle; it represents the complex interplay of biochemical reactions and processes that sustain life at the cellular level. The entire scene is a blend of science and art, using the concept of thermal imaging to reveal the hidden dynamics of cellular life. It invites viewers to contemplate the beauty and complexity of biological systems, seen through the lens of temperature and activity.

[boreal gate]

Needs more darkness, perhaps try adding a scene

[dim cradle]

[knotty vine]

Create a 3D illusion for a profile picture where a 25 Year old cute boy in a white shirt Sitting casually on a Wingback chair. Wearing sneakers, he looks ahead. The background features "Wilayt" in big and capital Orange neon light fonts on the dark grey wall. There should not be his shadow, and there are wings to make it appear as if he is an angel

[final compass]

Would be fun if someone can pull of this concept: "An image displayed on a computer screen, capturing a photographer in the act of photographing a painter. This painter is engaged in creating an oil painting, which itself is a portrayal of a sculptor actively sculpting a statue."

I am not getting anywhere with it 😄

"an attempt was made"

[final compass]

The circuit board stuff looks really nice.

Interesting image, although its not even close to what i tried to accomplish. Which was: "sculptor sculpting a statue of a painter, who is painting a painting of a photographer."

[dense mesa]

artception

[final compass]

indeed 😄

[dense mesa]

I did something sort of kind of similar to create this beauty

[late blade]

I'm worried that @dense mesa will put the wrong capacitor and burn the whole board at some point

[late blade]

you want the reulting image?

or inception like?

[dense mesa]

I have a highly trained eye

[late blade]

Hypotonic pupil or Adie Syndrom?

[dense mesa]

the components were a bit of a pain to install, but worth

[late blade]

any of the mods or curators here? I would like to ask something regarding daily theme

[final compass]

Sure, if you got something generated and want to share. My attempts were a mix of those images i shared

[late blade]

aha, if I get something worth it, I'll share it

[final compass]

I guess the problem tends to be that every character is doing everything at the same time. Had quite a few images of characters having camera on another hand and brush on the other, while being a statue himself etc.

[dense mesa]

thought this one was interesting

you guys ever have it make an image of you?

you'd be surprised how accurate it is

[late blade]

somehow I can't get him to show in the mural

[final compass]

yup, it seems like its one layer too many.

did you try using brackets?

[dim cradle]

I made this image but i can't understand how to make more realistic, could somebody help?

[glossy scroll]

Good morning 🌞

[late blade]

have you tried specifying the textures?

Also, you can specify betwen Vivir and Natural light

That's the best I could do

I don't think a brush is the right tool for the display tho

I used three scene descriptions

[glossy scroll]

[late blade]

hmmm I keep seeing people in the #daily-theme channel that try to post an image and it doesn't work and get the msg that the channel is to showcase images and then, I'm assuming, get the cool down of 30 minutes

[half wigeon]

Any ideas on how to get dall-e to accurately interpret the exact words placed in the prompt so that they appear correctly in the image? Many times, for example, if the prompt is turned, it creates an image that says "HOLA" the result could be something like "HELLLLO" and obviously the letters do not match.

[late blade]

Hello

Nice to meet you!

For text on images it's tricky and works more or less with short text snippets. It still has trouble tho. So you have to keep trying sometimes.

[wispy storm]

[late blade]

I'm gonna get you! I next is a Merlot animated image

[late blade]

are you measuring the vitals of the galax?

[dim cradle]

indeed, functioning within normal parameters and right on schedule

[final compass]

At least he is not painting the monitor while sculpting a camera etc. 💯

[late blade]

yeah lol

I just noticed that the last post I made has only 4 fingers... I'm so distracted I'm not noticing the trivial things

[dim cradle]

It’s okay I think it’s a given we must make allowances, it doesn’t say anything about our attention to detail

If I make an otherwise 95% spot on gen I may not fret over some details

[late blade]

you are right

I'm doing some cool stuff with whisper

[dim cradle]

Nice, I don’t know anything about that yet

[late blade]

reminds me of the mechas I had in mid tokyo

oh I erased those images out of frustration with the content policy lol

[dim cradle]

workflow traffic jam here. just gotta declutter, i got this. i just need to start dropping in photos instead of preview for the memory optimizations.

[dim cradle]

thank you! Using natural light is long way better

[formal osprey]

oh no, not caroussel 🎠 hahaha (if you saw dalle generate a caroussel before, you'll know what I mean)

it's so difficult to get a coherent carousel with dalle

[formal osprey]

that one is decent, for example. a bit of orientation weirdness on the right. But not too bad

good luck riding that horse

[dim cradle]

thanks for the feedback. it's from my gpt which attempts to write prompts that combine up to 14 different techniques for hyperrealism, so it's designed to alter your perceptions a bit. go full-screen, relax the eyes on a focal point, and let your mind play with it.

[late blade]

I think I'm gonna skip the nostalgia part of the proposed theme, I'm just gonna tie everyone into the carrousel and spin it so fast that bodies will fly all around

[dim cradle]

i usually do the theme in its entirety first, after that all bets are off lol

[tall mason]

I opted for so fast they'd be atomized.

[late blade]

oh that would be better

thanks for the tip

[tall mason]

I've got roughly 5 minutes before I can post it, though.

[late blade]

that's ok, are you red-shifting also?

[tall mason]

It's a carousel that appears to last way longer for the observers than the riders.

[dim cradle]

ludicrous speed

[late blade]

you just opened the gates of comedy

[hearty ether]

What the heck

It suddenly generated 7 for me

[late blade]

not seven, oh you are in deep trouble now

[hearty ether]

Now it generated 8

In normal gpt

The what

[late blade]

how are you doing that?

[vapid elk]

use the dalle custom GPT by OpenAI

it can make mor eipages per message

it still makes one at a time

[cold raptor]

Experimental combination of 4 photos

[hearty ether]

No it was just my own custom gpt

It somehow randomly generate 7

And 8

Two times today

[late blade]

oh that's interesting

[late blade]

dall-e openai gpt refuses to do that for me

[plucky hare]

You can try asking it to do something like "complete/initiate the image generation processes multiple times in your response" and it'll be able to. Like Lugui said it just initiates the image generation process multiple times so you won't get more than 1/2 at a time, but you can get multiple in a single reply. Of course will push you to the throttle limits/daily limit faster too!

[late blade]

I'll try again later, on a deleted chat was telling me it wasn't possible

[plucky hare]

Here's an example:

[ripe wraith]

woah, was this always possible just hidden?

[plucky hare]

I remember trying it shortly after DALL·E launched on ChatGPT while trying to cope with the fact that we didn't get 4 at a time anymore 😁 Nowadays I'm pretty happy with just 1 or 2 though. Esp with the 15min throttle

[ripe wraith]

when i try it now it returns 2 images in one response, which is good for saving requests

It talks too much doing it though, but still nice

[plucky hare]

In the DALL·E GPT? Same here, it's nice to have the option and I usually use it over normal GPT-4 for that reason.

[ripe wraith]

Yeah, its super efficient actually if you're ever running low on chat requests

im sure a little prompting would remove the extra text

[plucky hare]

Oh interesting yeah -- any custom instructions active on your end?

[ripe wraith]

yeah a big chunk haha

[plucky hare]

That could be part of it too, still not bad to read a bit sometimes 🙂

[ripe wraith]

yeah actually this shows why it works, its just laying out a plan and being proactive

if you give it enough details.. maybe the word process

theres probably something here for getting consistency across images.. having GPT build more of a plan before drawing

[plucky hare]

You can take it another step and ask it to use the data analysis tool to do fun things with your pics too, here's an example from a little while ago: #images-discussions message

[ripe wraith]

dang its way more proactive with dalle than I thought

[late blade]

oh nice, it works

[plucky hare]

I am wondering if I can get it to make 9 images, then stitch together a 3x3 square. I want to "Brady Bunch"-ify something 😁

[late blade]

that simplified so much for me now

now I just have to buy 50 seats for my team plan

[ripe wraith]

pushing its limits

[plucky hare]

Good luck!! I wonder if that's enough prompting at the end where it'll know to try to invoke data analysis to use Python to do it? Keep me posted!! 🤞

[ripe wraith]

on image 6 without stopping yet 😮

im amazed, its never this eager to work when I'm asking it for code

oop, regenerate button on image 7. Still thats a good run though

[plucky hare]

Good to know...I wonder if it wrote less text if that would give it enough "space" to finish all 9? Or, is the "continue" button appearing by chance?

Also RIP your daily DALL·E cap thanks for your sacrifice 🫡

[ripe wraith]

wait, isn't it just 1 response ? 😄

[plucky hare]

Yes for GPT-4 cap, but the DALL·E cap/throttles are separate

[ripe wraith]

weird, I've never hit Dalle before GPT i guess

[plucky hare]

Yeah it's 200/24hrs, with 15min and 60min throttles in place depending on server load (last Moxi told us anyway) #images-discussions message

[ripe wraith]

was able to refresh and ask it to continue manually

tadaa

[plucky hare]

Zing!!!

Brady Bunch dreams live, you've inspired me to try

[ripe wraith]

yeah it seem to love it

[vapid elk]

it usually does not do that unless isntructed

interesting

but it made them on separete calls, right?

all along the same message, but on separate calls

[hearty ether]

Yeah

one message with 8

On the same prompt tho

[plucky hare]

Getting there, now they just gotta teach DALL·E left from right (or I gotta try different ways to describe [or I gotta just get lucky 😁])

[ripe wraith]

haha wow

[late blade]

@plucky hare nope, still refusing to do that

[plucky hare]

Huh, what are you using in the prompt?

[late blade]

do a generation for each of the following ideas and i got a list of 3 different ideas

and i get an image with all stuff in one

[plucky hare]

Can you try copying the structure used in Chotes' or my examples above?

[late blade]

I'll check later, I'm in the penalty box

[late blade]

ok it worked with a new chat, seems some chats are moody with that, this one got it out from the bat

[hot rain]

This would be a cool theme, fancy restaurants/bars.

[dim cradle]

agreed. i made that one because it's supposed to be a rotating restaurant, a type of carousel

[late blade]

Seems we are getting spam

[hearty ether]

@dim cradle @vagrant pebble

Also is there any solution for texts from prompt leaking into the image?

Like I asked it to use a Silicon Valley aesthetic the words got into the prompt

[vagrant pebble]

You can create 10 images or even 40 in theory at once

[dim cradle]

I'm not sure if the model is tuned enought to prevent that

[hearty ether]

Any tips for preventing that tho

[boreal gate]

Try using text boxes( ) for text or using double quote.

[hearty ether]

Or words to avoid?

[dim cradle]

negatives in prompts tend not to work, but you could try something like "no text"

[boreal gate]

Text boxes use this symbol ` 3 times at the beginning and the end