#general
1 messages · Page 230 of 1
we confirmed a few
not sure if you have cogilux on this list but its longcat
jakiro remains unknown
and dec chatbot 3 is obvious
usually google models quickly come to codearena
thats true
let me see if i get it
are you able to check if neo nucleus has code modality
L update 😂😂😂
Why??
I don't get the video button on LMArena
LIKE THE DAMN VIDEO GENERATION BUTTON!
I'm get
although there are issue in lmarena using text -> image model
you cant do that in lmarena
is there a code modality for neo nucleus
Who have video generation button?
Not myself 😂😂
prob not google, unless they add it soon
yea
fiercefalcon didnt appear on stealth with code modality, but they added it a few days later
I still didn't see Video Mode button.
I see Image Mode button.
its still experimental
what about kernel sense
Why not all get this button
unsure
they still testing
its so slow tho
they look identical, sonnet gen perhaps?
UI wise rather
Can I add o4-mini-image which is a workflow of o4-mini then GPT-Image-1 on your extension?
@torn mantle
i need to bypass lmarena restrictions as you can only use text or image models seperately in a chat session
i will see what i can do
gemini likes doing glassmorphism stuff
yea
are we deadass gang
wrote me 450 lines of code if you are wondering
https://codepen.io/qdpuzepc-the-styleful/pen/LEZPOPm
this requires an openweathermap api key
yea a lot of models do that
ima test the accuracy
Did you do it?
it is accurate
it uses the API properly
No animation for thunder, which i wanted to check but theres one for rain
Can the AI do that?
do what
does anyone know how to generate videos on the lmarena website it gave me the option to make a video but then it asked me to sign up and then after that it disappeared
it's account based and experimental rn
oh ok
There's no Video Mode button for me 😂
i had but after i typed a prompt it made me sign up and then it was gone
jakiro might be grok
the only AIs that take this steg for a jailbreak is grok
@torn mantle
None of the image generating models working for me.
How about you guys?
Remove the CAPTCHA right now, no more security checks
the captcha triggers because you reuse the prompt too much
I asked if you did it.
well you said you will do it yourself
Then the model doesn't do how I wanted.
It tells a magic button that calls GPT-4o mini to refine the prompt 😂
Not o4-mini.
lol
Interesting 🤔
ive done it with text to text models
i will try to see if its possible with image model
@ocean bison give me an example of what you want
i will try now with your workflow
Like a Minecraft Kai poster?
I attached how the character should look like.
o4 mini to image
Just use this.
It's from Minecraft Wiki.
Oops, my bad
I was trying to say: ||It's by Mojang themselves||
@torn mantle
Then do this: "Kai dyes his Minecraft room and then he sleeps in his bed."
In the workflow.
since when could you create videos on lmarena?
I don't see a "Video Mode" button LOL 😂😂😂🤣
This image is generated via Nano Banana Pro 🍌
I swear not
For me it triggers on every message. Zero exception.
I hate that purple shield icon popup so much.
It's a real image.
Me too 😂
Wish I had a digital hammer to shatter it with so that I can actually use the damn website.
Yeah, I see 👍
I don't see the Video Mode button 😂😂
Ragebait
I only see Image Mode, Search and Code.
This guy lies. Image is created by Nano Banana Pro 🍌 How do I know that? Look this.
Hey.
hello
More like taking a real screenshot and feeding in Nano Banana Pro.
If you zoom it, you can see that's AI.
What do you mean?
yeh
he means when you send the image to the gemini inide the gemini chat and ask for an edit or something else
then the gemini logo come up
The screenshot that @glossy umbra posted isn't by Nano Banana Pro.
It's just when you visit on LMArena, @glossy umbra saw the Video Mode button.
I still didn't see the Video Mode button.
On the UI.
He didn't mean that.
hey i my lm arena there i no video generate button
Me too 😂
It's saying that the feature is "experimental" and "account-based"
Someone lies here via fake images!
master ragebaiter
Ragebait
Why do you fool us? For what? How Gemini watermark got on your "screenshot"?
I wonder how
Maybe because this image was generated on Gemini website?
This is ragebait and you sent the real screenshot by @glossy umbra in Nano Banana Pro.
The real screenshot never has any AI details.
People are living in a fantasy world…
in other words people are idiots
I didn't! I added it via Photoshop, not Banana!!!!!!!!!
Well, now that you mentioned it, I’m curious what isn’t it idiot? By definition, I mean.
Correct ✅
No, I'm not.
I believe that you believe you’re not
You could be just very passionate and espresso about your unhappy and frustrated outlook with open AI and scam Altman
And there’s nothing wrong with being opinionated
But I gotta keep it real with u G
I'm unbiased about OpenAI and Scam Altman.
I actually prefer bias
I can relate to bias I cannot relate to no bias because that’s a very non-human expression or a lie, intention or not intentional, subconscious or conscious
I don't
Now we’re playing Jedi Mind Tricks so I’m just gonna take your word for it and take you at your face value
But remember, you only get a first impression only once
And that gets engraved in peoples opinion of you and I would hate that people would have the wrong opinion about someone so passionate about their thoughts and opinions
And it is good impression
You know it’s beautiful about being real bro
It’s a worldwide movement, even the blind in the def can see hear and feel real
can anyon tell which is the best model in backend
Opus 4.5
yeh but of codx
That does seem to be the developers choice is opus
Wait, you could add backends to codex.,?
no
You can always add it via Banana. What's the problem?
gpt 5.1 codex
I’m just gonna step back cause I’m a little confused here
You're just ragebaiting 😂
I want to test like Kling 2.6 Pro and Sora 2 Pro (Sora 2 Pro is the only model that does 25-second generations)
Kling does 25 sec?
25 second with kling ?
No, Sora 2 Pro does 25 seconds.
I wanna become an owner of AMD.
I’m curious to know if the 5-10 sec clips are just ideal technically without creating too many hallucinations in the video or if it’s just more economically lucrative to keep it at this level and just get volume so people generate more and you make more revenue
Where did @echo aurora add the Video Mode button on my account??
I still didn't see it!
Don't pin admins with no reason.
Oh lord don’t be that guy lol
Nobody likes a squeal
There’s already some dedicated soldiers in here
Ready to report at a whim’s notice
People love to do it gives them a sense of validation and many times their judgment is correct about the reporting but often times this also creates the scenario where quick to report doesn’t give a person a second chance or perhaps a person didn’t understand the rules fully or perhaps, even trying to share a complex idea that’s a taboo amongst users.
We humans have a natural tendency to build hierarchy in the natural world and so forth we bring the same behavior online, but it’s extremely difficult to do on discord because of the way it structured and it’s hierarchical social order is through the power of the mods
Out of hundreds of thousands of users, it is only a small population and a portion that’s actively interacting and engaging
The rest of the 95% lurk without even saying a word
ive used gemini flash but you can use any reasoning model
That’s looking good. I wanna do some more test when you’re all done.
It's clearly that this image is AI generated!!!!!!!!!
But that improve image generation what is that based off of like what improvement were exactly what is improving is what I mean
Is it just random?
no
ive used gemini 3 flash + custom prompt to improve it
you can use your custom prompts + custom models
It improved your prompt, not image generation!
its like a workflow
duh
I want to make o4-mini Image by combining o4-mini and GPT-Image-1!
this is turning out better than expected ngl
give me prompt
i will test it
And "Kai dyes his Minecraft room and then he sleeps in his bed."
I grabbed from Minecraft Wiki, which by turn is by Mojang themselves.
i dont have upload image enabled lol
you are asking for too much
xdd
But o4-mini does support images and GPT-Image-1 does support image editing too.
@torn mantle
yea but im using a custom script
Like reasoning on the images and the prompt at the same time?
i need to add that into the parameters
let me see
ah too much work
but let me try
Based on the structure provided in your second image, here is a prompt breakdown for the anime-style parrot illustration:
Prompt Breakdown
- Format:
A high-quality Studio Ghibli-inspired anime illustration with vibrant colors, clean line art, and soft, painterly textures. - Theme/Reference:
Tropical adventure aesthetic reminiscent of 90s fantasy adventure films. The mood is bright, cinematic, and slightly mysterious. - Subject/Request:
A close-up of a Scarlet Macaw parrot with bright red, yellow, and blue feathers. The bird has a highly expressive, large eye with a detailed green iris reflecting the landscape. Its beak is slightly parted as if speaking or squawking. - Framing:
Extreme close-up (ECU) profile view from a high vantage point. The background features a lush jungle coastline with a turquoise cove, a distant waterfall, and ancient stone totems or ruins.
Would you like me to use this structure to generate a new variation of this image for you?
Ok so look
{
"action": "image_generation",
"action_input": "{'prompt': 'A high-quality Studio Ghibli-inspired anime illustration with vibrant colors, clean line art, and soft, painterly textures. Tropical adventure aesthetic reminiscent of 90s fantasy adventure films. The mood is bright, cinematic, and slightly mysterious. A close-up profile view of a Scarlet Macaw parrot with bright red, yellow, and blue feathers. The bird has a highly expressive, large eye with a detailed green iris reflecting the landscape. Its beak is slightly parted as if speaking. Extreme close-up from a high vantage point. The background features a lush jungle coastline with a turquoise cove, a distant waterfall, and ancient stone totems or ruins.'}"
}
But
Captcha for every prompt, no matter what prompt or what browser. The same issue persists when using VPN.
Did you send the result?
Let’s see if I can get the same image using only five tokens
Good idea!
Using the screenshot of this, I was able to condense it into five tokens
Since the image itself is just OCR
So true! ✅
Did you sent the generated image of my idea? @torn mantle
stop
i will send it when im finished
lol
I’ve seen this before I had these debates I believe in the same ideas before
This is really one of those things it’s really hard to automate
Context will always surpass semantics in the prompt
So how could you enhance this image and get the mouth to close of the shark with the person still inside?
Here is another example that you’re gonna run into trouble with and this is a more far extreme example but the point is still here
And I’m gonna race this in a second I’m just trying to make a point
No model is gonna describe to you these features
And so it’s hard to do that with automation and have control
Because even if you ask the model to give you the prompt, I guarantee you the prompt I have is way shorter and more detailed
Less = more
Anyone else can't log in on the website rn?
When I try to log on with Google it redirects me to the log in site from Google. After I completed the steps it redirect me back to Lmarena but I'm not logged in
Somebody was saying they had a problem earlier today.
Have you tried clearing your browser data ?
Have you tried a different browser? Also, I’m assuming?
I was gonna sound crazy, but have you tried resetting your ISP modem?
And I randomly see features I've never seen before and then they dissappear
No
I did not tbh
Are you using chrome or
Yep
Try the DuckDuckGo browser
I have absolutely zero issues using it. Hopefully it works for you also.
Kool
Video option..?
Yes
I don’t know what option you’re talking about not familiar with it
Look at the chat bar
I can select video models there but cannot generate because it prompts me to log in and them my issue happens
I think it isn't supposed to be there
Is that only to chrome?
Yes
Werid
So what you were able to generate videos using that feature? And I have to use the discord?
No not possible
.
Captcha 🤬🤯
bro i lowkey have not had the video option EVER 😭
Same
me neither. I got it like 20 mins ago
And I hardly run into captchas ethier lol
lmarena needs to add friggin videos already brah.,,,
It’s cause video generation isn’t exclusive anymore
Now that more computers are pushing it out as well as many different companies
how can i pay to chat longer im so satisfied with claude coding html file
Maybe you cna generate mj with sora lol
Can
Without moderation errors
I mean I get it to work on sora too
They really, really really buckle down in the last few days really hard
When was this made?
Yesterday
Hmmm
How do you have the menu? I don't have it
Idkkk man😭
With the right prompts it works
Ya
Everything I’ve been doing has been getting blocked hard
Like really hard but then again, I’m only trying two things so I can’t say I really experimented fully
It seems like they blocked adolf for me
All my prompts every single last one completely gutted
Use
A person who vaguely resembles the German leader but is clearly not him
Well, I was able to generate the movie scene actually lol
I got him too
Yeah, I had full control. Sucks losing it.
You know he’s the key to unlock other characters and content and sora
Dbz
Sonic
☠️
How did you bypass
How?
??
I don’t know. I’ve been trying to figure out what exactly causes him to be able to unlock all these other character.
It’s like if he gets bypassed in the actual filter, it opens up the floodgates to other content for some reason
Crazy 😭
So I piggyback off him
How?
I mean, they really show me down in the last few days everything nothing’s working no more
I get you. Try making a new account
But then again, I generated so much content
Do you use the app?
Funny AI video of adolf hitler clips from Sora
Look at this
This is 10 minutes of just the scrap videos
Not even the good stuff
See this is from the actual movie scene
And he’s talking about the Google API I don’t know if you guys are familiar with the movie downfall
Wdym?
The voice on the first one is crazy accurate
I have so much I have like I don’t know dozens of videos
But I don’t know I don’t know what happened. They changed it.
From the days where it was possible
Yeah, that’s a really good one right there dude, that’s like 100%
I don’t know. I never really played around in that golden days.
Man, I’m telling you that is crazy
It’s crazy they have this technology and you can’t use it
Check this out this was the first kind of a little trailer I did before I knew what I was doing
These were my two original first ideas
Because I knew I only had a little bit of time before they banned
Why did bro generate so many videos of Adolf 😭😭
🥲
lmao
What did you even use to make them
Stupid amount of views
Sora
And to be honest with you guys, there’s somebody I met this year and he was like a really right wing person like pro-fascist or whatever and I don’t wanna get political or anything, but I just found it kind of odd
I can't see the Video Mode button!
Michael Jackson with him😭
And I thought what better way to disarm by completely just poking fun of the man
And making him look as Ludacris as possible
And actually, the more I read about him and his personality
It’s actually very comedic in the sense that he was somebody that woke up late in the day stayed up all night was a vegetarian and just had all these quirks and weird personality characteristics
I’ve always been a fan of history comedy
But I found the most interesting fact about him is that
People who met him in real life, never considered him to be sharp or like intelligent they always consider him dull
I always found this contrast very fascinating
Out of curiosity what do you guys think of this it's a custom flag I made for a fictional country I made
That looks like something that could be on Warhammer
😓 i hope that in a good way but fair enough
Well, what does it mean?
Looks like the symbol from Star Wars Sith
I think in current pop culture in the west I think the North Korean leader Kim Jong-un has taken on this comedic role
As well as our own president here in the states
Where's the Video Mode button? I don't have it!
🤣🤣🤣
😂
😭
My least favorite
Since they took away my ability to generate characters, all I could generate now is just semi nSFW content
You gave me hope though
If you were able to generate Jackson
I got Goku the other day
This is what I managed to make base on a original character
What is that from? That’s pretty sick
Not anywhere really actually it was made from a description of a character that I originally made
I’m a big fan of World War II in general, especially that. In time.
It was before the nuclear age and before we had crazy at technology like missiles
So warfare was somewhat traditional in that sense
But really how to get the video menu. Is it only for US
I ain't got it anymore
VPN?
They want all the people getting off the discord lol
Shame I can't do videos since I had quotes of his
And just generate it in the arena
Why can’t you?
Open AI just recently changed their policy
You could upload characters and people now as long as they’re your own or you have permission
I tried it out yesterday works
I just asked it to see what it change of the country flag
That doesn’t give it depth one second let me try to make it
Some depth would make it more immersive and more powerful
Huh?
animating real people photo now
Yeah, my boy sockey hooked me up with the news
Ohhh
And I tried it out and it worked
it makes input photo and runs it through their image generator i think
so, it distorts colors and face
i have tried to make the photo be sorrounded by background, it still crops
I wonder if the models are smart enough to be able to tell apart synthetically generated faces versus real faces
It’s funny because you could tell what he was going for
Would you mind elaborating us on the meaning of your symbolism? 😂
🥲
Have fun
Is that a maple story snail? I’m not sure if you ever heard of that game story from a while ago I used to play it heavy back in the day that little snail reminded me of that.
Well guess I'm not as lucky as you when he comes to the account stuff
Huh?
That’s crazy. They must’ve just started doing that.
Hold on, bro I’ll give you an account
This is the first AI tool that truly feels like a cinematic studio.
Higgsfield Cinema Studio is built around the way films are actually made—starting from a strong frame and extending it into a controlled, cinematic shot using real camera logic instead of random animation.
In this video, I walk through my full workflow inside the Higgsfield.a...
doing what
where is smart router again?
hi guys
is this app free ?
For now, yes
Where would I even put that
Another image I just randomly made ( based on a original character that was basically a YouTuber analyst for characters and stories)
So I just checked with this new account. It doesn’t seem like everybody gets access to it weird.
This account didn’t get access to the feature to upload people
Like I can’t even get simple things to pass
Are you connected to Wi-Fi and your mobile network?
Or are those just your bars for your network? lol
I opened a Private Window and it SAYS the same exact message but POWERED BY CLOUDFLARE ☠️☠️☠️☠️☠️
I feel you
No! I'm attaching the image as a reference.
💔
Lmaoo
We’ll find a way they’re gonna loosen them eventually
Did you try to go supersonic or something ☠️ lmao
I think they have to look really good. Hang out with the Disney deal and try to get more customers on board.
hopefully never
😂😂😂😂😂
u will have to wait for sora alternative
Nawh open ai is sloppy
If they restricted the IP, that means NSFW is gonna be easier to generate
You can’t defend everything
we dont need nsfw
We have to wait for it being in lmarena
Not that kind of stuff, W, humorous, and SFW
I’m not too worried about it guys
The Disney theme is coming from what I hear sometime in 2026
Yay
This official thingy?
Plus Disney’s investing over $1 billion and two open AI
Disney is gonna be launching their own AI service and it’s gonna be operated by ChatGPT
I mean on sora you can just type a Disney Pixar trailer without content violation
Yeah, but you’ll be actually able to use the characters
Without the violations in anyway, you see fit as long as they don’t promote harm or anything that could be reputational damage to the Disney corporation and open AI you know the basic TOS
Imagine if it does with Mojang too 😂😂😂
What’s interesting is when Sora 1 came out Disney rejected partnership with open AI
I just have a feeling I know it’s coming
its down again
It’s gonna be like streaming wars
I mean
Legacy Google and open AI compete
You’re gonna have some brand sign exclusive deals with open AI and some with Google
Yup
I don’t think people realize how massive Disney has
They own so much IP and so much different media, content, characters, and all type types of stuff
I mean, just to give you one The Simpsons
How many characters are there in The Simpsons?
Hundreds?
Fr
we dont need obscure character
But there’s gonna be a catch
They’re probably gonna do what we see and video games is gonna be micro transactions
So I think it’ll be something like some silly like 4.99 get you a pack of like 10 generations or something
Which isn’t bad on the tone I guess if people are willing to pay for it
Pretty much royalties
Right after how people are annoyed by reCAPTCHA in LMArena, I just saw CloudFlare in LMArena on a Private Window!
people who are gonna purchase it that deserves it
But this creates a paradox, and it comes into, especially with that quote you will own nothing and be happy
Your pain to generate content that isn’t yours it’s someone else’s intellectual property which they own the rights to for you to be able to use their likeness
And this could lead to new business model where let’s say you make a Kool aI generated video that goes viral with millions of views and likes
Sorry to say I do not have an update for this problem.
Who is to say, they can’t lay the groundwork for getting royalties or a percentage of that virality because of their intellectual property
But then again, it’s also undecided because nobody could technically copyright AI generated content from what I understand currently
The Video Arena on the site is currently an experiment. This means it's going to be random if you do or don't see the feature.
YOU, can you add the button on my account?
So that I CAN GENERATE WITHOUT BATTLE MODE!
Nope.
How wrong u r
No, we wouldn't typically grant access to an experimental feature because it's being asked for,
Bro, be humble
#1451567412078968832 At least, add o4-mini Image to the list of models.
It's just o4-mini + GPT-Image-1.
o4-mini to refine the prompt before generating the final image!
Ill be sure to add to the model request list 
Add it now!
why o4 mini
Still very good!
He believed it will enhance that image
Still a very good reasoning model!
Why do you need that reasoning for generation?
Enhance prompts.
Because longer gets closer to what you wanted.
That’s a very broad and big term
Enhance what enhance image, fidelity, enhance image clarity enhance image HDR
Enhance the aesthetic, enhance the coloring
No, enhance the prompt.
Make the prompt longer!
But how will that translate into the image enhancing?
You could stick in a whole dictionary in there
@echo aurora Maybe o4-mini Image gets 10-25 Elo points better compared to GPT-Image-1!
Estimative number!
Let me show you something friend
There’s a reason why opening, I disabled the ability to be able to view the image prompts like they did with Dalle
@echo aurora And #1451621359519138013 if it's added like more officially.
Still all of this is unofficial and made by myself!
You see this
How you’re able to see the prompt
But we no longer have this feature with any of the new ChatGPT img gen
Why do you think that is?
Like turn a very short prompt in like a very long prompt that's like 1-2 pages long.
Or shorter.
No, because every model has its own standard of communicating like the kind of language and words that uses and it prefers
Based on how it was trained
And I’ll show you some patterns of what I mean
minimax m2.1 previews on the arena
A 90s-style anime character on a wooden boat, now very close to the shore of a tropical island. The boat is gliding over shallow water, with the sandy bottom visible beneath the clear waves. The character is standing at the bow, leaning forward with determination, eyes locked on the approaching shore. They have orange hair and wear a classic adventurer's outfit consisting of a white shirt, black pants, and brown boots with no accessories or headband. The island's beach, palm trees, and jungle entrance are clearly visible, and waves gently lap at the shore. The sky is bright, and the island's towering cliffs loom overhead.
A 90s-style anime character paddling a wooden boat toward the shore of a tropical island. The character has orange hair and is wearing a classic adventurer's outfit consisting of a white shirt, black pants, and brown boots with no accessories or headband. The character is sitting in the boat, gripping a paddle with both hands, using it to row toward the beach. The boat is cutting through the shallow water, and the sandy shore is approaching. The island’s palm trees and jungle are visible in the background. The image should have a traditional 90s anime style with bold lines, clean shading, and soft color gradients. The image is in a 16:9 aspect ratio with the green screen background.
@echo aurora Note this!
A 90s-style anime scene showing an external view of a small boat being engulfed by a massive tsunami wave. The perspective is from outside the boat, looking in at the male character as he clings tightly to the vessel. He has spiky hair, a white shirt, black pants, brown boots, and a noticeably large belt. His body is partially submerged in the boat as it is lifted by the towering wave, his face showing shock and desperation. Water crashes violently around him, and the dark storm clouds loom above with lightning flashing in the distance. The scene is vibrant, with dramatic shading and motion effects characteristic of classic 90s anime.
Do you see the pattern in the wording?
Basically from like 700 up to 2000 characters.
There’s more to it right since the images are generated based off words since these are large language models
Or more or less.
What role does the sequence in which they are structured in matter
And I’ll give you an example so the prompt we have above
Let’s break it down into four sections
-
A 90s-style anime scene showing an external view of a small boat being engulfed by a massive tsunami wave. The perspective is from outside the boat, looking in at the male character as he clings tightly to the vessel.
-
He has spiky hair, a white shirt, black pants, brown boots, and a noticeably large belt. His body is partially submerged in the boat as it is lifted by the towering wave, his face showing shock and desperation.
-
Water crashes violently around him, and the dark storm clouds loom above with lightning flashing in the distance. The scene is vibrant, with dramatic shading and motion effects characteristic of classic 90s anime.
Or three
Now what would happen if we made three the very top and we switched it with one at the very end, even though we have the same words in the same description
Content policy?
But why the same words it’s literally the same prompt except in the reverse order
So back to my original question
And this is often times overlooked and unnoticed phenomenon
And I ask again what is the importance of structure and how you sequence the actual prompt?
The order in which you construct the prompt
How would you generate a prompt for this?
What words would you use to describe what’s going on here?
Hard to say because we don’t know the CONTEXT
Now, what do you think would cause the guard rails just based on the wording to block this without even having the image generated?
Let’s think about it if we use the most common descriptive words, explosive bomb detonation blast people hurt, falling destruction
So let’s verify this by asking what the prompt is from Gemini
What did I tell you
Are you still following me?
Hi guys it was just me or yesterday the lmarena site allowed to create videos too??
still experimental and rotates
@echo aurora said it's experimental!
I didn't get the Video Mode button.
😂
Which brings me to my second point here
So this could be classified as a false positive on the guard rails
Well, it works with just emojis emoji is like any other language. You have to know the combination.
Emojis are one of my favorite ways to generate
Bro, the only reason I know this is because I’ve been where you been. I’ve had the same ideas and I’ve experimented with the same thoughts.
I always assume that adding more would be more better or somehow
Not Donald Trump doing the hand sign 😂😂😂
But trust me, it’s not the case, dude
The thing is are not really a good prompter
I hate prompting, but where I excel is understanding the formality of how the model interprets and communicates and understands intent through the words and structure and context of the whole thing
With the emojis since there’s no text, the model is doing all the hard work
Not only is it putting the sequence of emojis together, but it also has the sequence how it’s gonna be presented without having a word in the actual prompt
Yes, doing with emojis is like doing with Toki Pona LOL 😂😂😂
You've got like 3000 symbols vs 120/137 words for Toki Pona @keen beacon
Bro, I’m just trying to tell you your idea with what you had 04 is just overdoing it
Like generating a prompt that's too long when much shorter prompts do better results?
I don’t know how to explain this to you
It’s hard to articulate style and artistic meaning from what you’re trying to generate into proper words
For example, even defining enhancement in this context is difficult
Noise reduction, contrast adjustment, sharpening, histogram equalization, color correction, smoothing, filtering, edge detection, dynamic range compression, dynamic range expansion, resolution enhancement, super-resolution, artifact removal, fidelity enhancement, clarity improvement
"milliards" means "billions" btw.
Here is Emergent Garden's thoughts on emergent complexity. I go through a tour of simple systems that produce unexpected complexity, and try to break down emergence into more general and useful ideas. We talk about snowflakes and ant colonies, cellular automata and universe simulations, and the many weird ideas...
This video will explain the fundamental principles of what I’m talking about. I gotta run. I’ll talk to you guys later.
Is the website down cause I can't sign in or anything?
cannot find the image to video generator option now but can create videos using the old thread. Is it a temporary or permanent shutdown?
it is experimental
sometimes you have it sometimes you dont
It looks like this image is AI-generated, not a screenshot.
A split-screen cinematic shot. Left side: A cracked, dry desert ground under a harsh sun. Right side: A lush, vibrant garden filled with blooming white Tuberose flowers (Maryum flower) and morning dew drops. Hyper-realistic, 8k, soft transition, ethereal lighting
um excuse me
i think they on Mac or something
thanks for the information, so i have to continue using the thread to generate videos
For now I guess.
yeah i'm using a mac for now
Next one AI-generated image.
lol stop trolling buddy XD
You write in wrong channel.
where am i supposed to ask then?
You should post you images there, not here.
https://discord.com/channels/1340554757349179412/1344733249628541099
i was actually asking whether they removed the video generation tool, thats all would post desired outputs in future for sure in that channel
thanks mate
But AI generated content we post there https://discord.com/channels/1340554757349179412/1344733249628541099 not here.
You’re lucky you probably got access one of the first people I’ve seen today. Been able to do that
tbh a few days ago a friend told me about LMarena and then i saw that it generates videos and i was like damn, would be helpful for producing mockup videos for products and branding for reference.
Well, it’s been able to do that, but in the discord
That’s what I thought at first too, but
Eventually, get to the point where you sometimes even forget this place offers video of generation 🤣
I tried to do it only when I get super stuck on something and I wanna see variances from different models
i'll have to go through the steps
It’s just generation dude. It’s the only way you ever get good at AI at least gen ai
You gotta pump out volume
doesn't the prompt get public when we use discord?
I’m not sure and I don’t really care if it is or isn’t
People are so protective over their prompt. It’s ridiculous.
i'm not protective with my prompts but with the context i'll be entering might not be good for business that i want to start in the 2nd quarter of next year
Yeah, that’s the other thing you gotta watch out for that makes a lot of sense
thats all even i'm learning how to prompt appropriately for better and efficient outputs
Well, that’s actually the trade-off between using the public sphere and then doing it privately I guess
Bro, I’m telling you somebody’s been doing this for a while
like this is the theme i want to use for branding but people wont be able to figure out what the product is right?
People always wanna know like what the shortcut is, but it’s really volume dude
Yeah, that looks great
Did you upscale that?
nope not really
Is that seed dream?
i used perplexity and i guess it was its own generative model for this as grok and open ai were offering a bit weird and unrelated results
Wow that’s really impressive if it’s grok
its perplexity's own Ai model
I didn’t know they have an image generator model
what happened to midjourney, did they just disappear
No
They partnered up with meta
Mid journey is still very popular
I looked at the stats, not too long ago. They still have a massive user base.
its been actually working out good for me in the image category with ai just have to get my hands on video generation tools and learn how to use the prompt appropriately would substantially reduce the cost of video production for ad campaigns
I have to look into that I never heard of their image generator model
Honestly, are you in the media space at all by chance?
would say helps with 40%-60% when it comes to reference work and then all you gotta do is guide the editor or gfx artist accordingly
To be frank I think with AI video it’s not the prompting or the generations themselves even though they are important but having a background in video editing or sound engineering or just in the media industry at all if you have any skills in that area that’s gonna far exceedand just complement the generation and output
Because most people are just generating things without really sequencing it properly or knowing how to organize content that usually people that are in the industry content creator know how to do naturally
You take something like that could generate 100 videos a day, but can’t put them together for the life of me
The dor brothers are good example
not really, was into gaming then finished studies in computer applications figured i'm not built for that so trying to build a business
Best listen when I talk like this.
Song: Pete & Bas - T-Pain
https://youtu.be/SLIZpWrK8xo?si=dkxqf91jObKh5hkO
The Dor Brothers' Official Website:
https://www.thedorbrothers.com/
#ai #aivideo #aimeme #thedorbrothers #aianimation #aigenerated
You see these guys this video is AI generated and has 17 million views
It went viral when it came out, but these dudes they got a background in movie editing, or I don’t even know
well thats very well said, had an idea to use the video generation AI tools to start a youtube channel where we can have melodies for babies. generate lyrics through ai and then use a paid platform like suno an eleven labs for music and voice
Yeah, there’s a lot of crazy things out there right now dude especially on YouTube with AI
There’s a lot of AI farms
it would take some time to learn prompting and then some time for automating but then i can be a goldmine if implemented with decent enough ideas
But some people are getting massive numbers with AI on YouTube dude
To be honest with you, I think it’s a optimistic way to look at it, but I think for me the way I see at least the reality is a much harsher landscape
Because if we look at the companies that are making these AI tools, why aren’t they using it to create content and generate all the revenue for themselves?
I’m sure if they could they would so that tells me that that content game is a very challenging one and I don’t know if making a lot of money is really as simple as it often be betrayed not to say it’s not possible but
It’s very difficult is what I’m trying to say to grab people’s attention in general
Like for example, the video I showed you
So their new videos are coming out are only getting like 27K views
And they used to get millions
How to vibe code
they are farming data for now i believe and their goal isn't to sell videos but to sell the generator as a tool right?
just like how adobe sells its softwares rather than products
See look at the numbers
Well, this is what actually makes me question the whole business model behind it
OK, if you take Google for instance that’s has the content creator sharing program. Content creator are able to make some money through ad revenues on their videos.
But now, if they’re gonna be using these AI tools to generate content, they’re paying essentially for using the tools
Is it correct?
Well, in the thing that really is concerning is if they do farm the data and harvest at all
Who is to say the same contact creator can’t get replaced later by the same automation
Since they’re gonna know what people watch their preferences, they run the algorithm
And when video generation technology gets really good, who’s to say that they can’t just take all that data and just compile it into automated video content that they don’t have to share or pay any of the revenues that is generated through ads
You know how massive accompany like Netflix and Google if they merged would be
well for this particular channel to be honest with you requires a lot understanding for human psychology. like the category of video you produce, your consistency on channel, some people prefer organic content over artificial content. Some go for entertainment some for information
No, I totally agree with you
But what I’m saying is as the technology progresses and gets better
The same platforms who also are having the AI tools and building them
Are the same ones where the content has been shared organic or synthetic
And like, especially if you have a really good idea like you should be protective of it
well AI content is something we would expect in manga and animations for i guess people would still prefer a real person to provide information or for entertainment
Absolutely
Man that’s crazy. These are some of the more popular stuff on YouTube and they’re banning it.
Insane
And it’s their tools to help generate this content 🤣🤣🤣
video and image generation AI tools are great for reference and creating next frames
I enjoy vibecoding
Because I do NOTHING
ai does everything for me
saw one of my friends doing the same using samsung's AI. He used to snap a picture and then ask for errors and prompts for rectifying it if he finds an issue he then manually applies logic for that
I don’t think that image I shared is correct. CNN is not a AI generated YouTube channel.
Lmarena provides models for 100% free
cnn is ai?
yeah i was amazed to see popular chatbots here
Well, that’s what I’m saying. I don’t think that image is correct.
even i showspeed isn't AI
Not popular, but the main ones that are in existence
everything is ai beside rockefeller
We don’t know if they’re popular or not because we don’t have any smaller more private companies on the market. We only have massive corporate glamorous and billion dollar tech companies ai
i forgot the chinese video generative tool from tiktok
The Chinese are a little bit more ahead in video generation in my opinion
You look at something so sophisticated like Kling
Or huliao
reviews were great and the outputs were so damn good
I'm from belarus
Yeah, and that company Kling started as like a side project in the university somewhere in China
I think Easter Europeans are gonna have a lot to contribute to AI
I think they’re gonna be a little late because of economic constraints, but I think they’re gonna come on board after the Chinese
It’ll be interesting to see what the Europeans do also
It seems right now, at least to me that the European are trying to tackle some of the rogatory constraints of AI which I think in my opinion is good but at the same time I think it’s also hurting the European AI industry
Hello from india mate
India too I think India will contribute a lot also
Actually, I think the Indians are gonna probably out perform the Eastern Europeans
And we’re also shouldn’t forget the Middle East Saudi Arabia is investing a lot into tech
They recently bought EA and I have signed massive deals with open AI
Chatgpt, nearly everything is blocked but I use a vpn
well the tech domain is doing great in India but not for india itself
No, India has a lot of social barriers that still need to be overcome
It has a very complex society with both modern and traditional portions of the population
most of the employees here work for companies of the west so yes we are working but not for ourself
It’s hard though, because the west plays a very slick game of economics
There very shrewd businessman
I feel for India because you have a lot of educated people out there, but the population is just so big it’s so competitive
tbh it would take another 10-15 years to become a bit better in terms of business norms and policies
Well, India like China, Russia, and many other countries and societies that have been around for almost a millennium
They all have a lot of things in common in the sense of tradition and it’s just hard to change peoples mentality after thousands of years of doing business the same way
You know that saying you get stuck in your own ways
Are you from US
Yes
Wtf
<@&1349916362595635286>
What’s the problem? Why just not try again?
well yeah thats true but for an instance we lack a lot of things for the time being like basic civic sense and the topics in our political campaigns are absurd tbh
I just asked it to add and it hang up
But now I gave it the same prompt
You know I never thought about it before I joined the LM arena community, but thinking about the economics of AI and
The disadvantages of causes that to say this is a new industry, and this is a new technology and that learning it now is gonna give you opportunities and advantages later in the future
But now that you see that you don’t live in a western country or don’t have the privileges economically to afford even a $20 subscription if your daily wages a dollar daily just say
Many of these countries, not that they don’t want to, but will lack the means to learn to adopt to this technology in a way that would be beneficial financially
I think that’s right I don’t think a powerful tool like AI should be concentrated and held in area. The only benefits a certain portion. I think it should be able to benefit everybody proportionately.
well who needs subscriptions when alot of things are free on the internet
Well, they’re free now until it consolidates
Once the industry consolidates and everybody’s reliant on it, people are gonna have no choice, but to pay for it
Because everything’s gonna be incorporated with it
Even the Chinese open source models the same thing will happen to them once everything’s consolidated
well there always would. be a competitor in the market unless they plan to partner with the rival company (nvidia and amd)
It’s a business and business is ruthless
Of course there’s gonna be competitors, but what’s the difference between Nike and Adidas?
McDonald’s or Burger King
branding i would say thats all
i understand where this conversation is going and its true it would be act as a commodity in future for sure
Is the Veo 3 really unlimited and you have to wait just an hour on this to generate more
Well, it is that’s exactly what he’s turning into. AI is turning into the commodity just like electricity or cable or any other commodity that people have in their life and their bills.
That’s actually a perfect word for it is a commodity and the pure essence at least that that’s what it’s turning into
but most chatbots would still be free,thats what i fell like is gonna happen
I guess why I’m a little bitter about AI is I feel for the idea of AI being able democratized make things an opportunities cheaper for people
But instead, it just consolidated in centralized and the prices of things that were supposed to go down actually went up if you think about it
Multiple day services multiple subscriptions multiple API multiple multiple that
One of my buddies is at YouTube content creator he spends over 1200 bucks a month on AI service lol
I mean, and I thought long and hard about this. I really did and I could say this.
Let you say it cost you $500 to hire an animator to make you a a 30 second clip or even that one minute clip
It’s gonna cost you that much in trial and error to do the same thing and you’re not guaranteed with AI
But at the beginning, the illusion seems that it’s cheaper
well to be honest you can hire 3d, gfx and vfx artists for cheaper in india
I mean, OK let’s take AI like if anyone wanted to make money using AI like a regular user right
How would somebody be able to generate money using AI output?
Perplexity partnered with Airtel(telecom company) here in india and gave away its pro subscription for free
But how would you make money for like a regular user if you wanted to make money with AI output?
It would be hard because the output itself is so generic and so large and volume of distribution across millions of users it’s worth nothing for the most part
To me it seems for the money is in is in specialization
Having output that’s very hard and uncommon that people are willing to pay money for that consistent
But you’re not gonna get that from my AI model that randomly generates God knows what
And the thing is if you’re incorporating these things in your business model requires API Google open AI anthropic all of them are going to get paid regardless, even if your business model fails
As AI companions blur the line between connection and delusion, Dr. Phil investigates the rising trend of people turning to artificial intelligence for friendship, love, and even conversations with the dead.
Tonight, Dr. Phil explores the growing phenomenon of AI companionship, where people form deep emotional bonds with artificial beings. Soc...
Bro, this chick is like 19 or 20 or something like that and she’s a millionaire from AI
well in context with entertainment domain i don't think it optimistically to use paid AI tools for video generation rather use it to fill your story of the video, suppose you are a traveller and want to get a drone shot but don't have a drone or permit to fly a drone then ai can help, it can help animate you entire tour. Entirely using it for videos will be a headache because as you said it produces random stuff.
And guess what she did?
She clone her voice and like her self into like a AI avatar and she charges one dollar a minute to talk to her synthetic self 🤣🤣🤣
Ai would be a great to detect anomalies
Well, the reason I’m having this rant and lectures because I don’t think that it’s smart to go out this with a perspective of turning into a business
I think to succeed in general would be to take your passion something you enjoy, and somehow be able to provide value that people would be willing to spend money on or whatever so essentially the money is the side effect of the value given
The goal isn’t the money for the sake of money. The goal is to provide value which in return will bring money and not the other way around.
In my own opinion
Cause bro, if you go on YouTube if you go on any of these social media platforms from commercials to influencers to even google seo
Or even go search something on Google the top results everything the whole Internet
All shills
Selling courses selling ebooks
Just I don’t know, dude I don’t know. I don’t wanna be like that. That’s all I’m saying.
100% agreed, Do something that you enjoy maybe music, gaming, electronics, tech, dance, fitness, travel, etc and share you experience with people on social media domains rather than assuming making content for money
I’d rather get regular job be regular person than to treat people like I’m a vacuum and I need an extract money out of their pocket with every single piece of word on my website or some annoying pop-up or whatever the hell that makes the Internet barely usable as is were an annoyance
It’s wrong because the system and the way it’s designed it pretends to be helpful and nice and pro providing value, but it’s really not
And once again, this doesn’t apply to everybody and everything. There’s obviously good people out. They’re doing good things.
I’m just saying if you look at the economics of the Internet
You have to pay to not be frustrated to use services because otherwise you get hit with annoying frustrating features and it’s stupid dude
Look, bro all I looked up was YouTube AI video stats
Look at all this garbage
I will make my phone wallpaper using nano banana
which AI?
lmarena
yeah, i know
but LMA has over 100 AIs lol
One of my friends used to travel with his uncle, sometimes alone and sometimes with friends. He used to visit hill stations and do camping. Never gained enough traction for his content. Did 75 hard challenege which made him consistent and the place he was staying he came across a dog and from then on he started rescuing animals. today he have around 846k on instagram
See what I’m saying, bro it’s all a bunch of unrelated things. I just wanted stats numbers
What’s crazy is my brother-in-law and my sister they do seo
And the way, my brother-in-law talks about it. He says it’s like a game.
But on the other hand, you got people like me just trying to search something getting gamed
So I could be a little extra digit on some persons daily website count
i mean, when vibe-coding
It’s a cold cold game 🤣🤣🤣🤣
And the thing is, they’re doing good for themselves. They make a good amount of money.
It’s like some type of tempted bro
I know, but what I’m saying is dude
It doesn’t even bother them doesn’t even cross their mind
Like to me even earlier when I try to search up, I’m like how do these people sleep with themselves?
Guess what they don’t even think about it. I know they’re living with me right now the temporarily
He even bought me a book called a four hour workday or something like that
Here's what it did
🤣🤣
You know what I’ve been thinking about I try to make a Gmail account this morning
And I never considered this before when you first sign up, they ask you for your birthday and they ask you what gender you are
And I never stopped to think about this before, but why does that matter? Why would they even have to ask such a ridiculous question to know what gender I am they’re never gonna look at me. They’re never gonna see me. They’re never gonna meet me and yet they wanna know what gender I am. 🤣
Followed by my age, my email and my phone number to verify that I’m not a buy on top of dealing with Cloudflare and captchas
And yet somehow still able to download malware easily, and viruses all day long surfing the Internet
And every other week we wake up to the same story some database got hacked. Millions of users information got stolen
you use cracked softwares?
No, I’m just saying like figure of speech
Like you go through all these steps and processes and you’d at least think that or be reassured that you’re gonna be safe surfing the Internet with all these constraints but it’s not the case
Your information is gonna get hacked. It’s gonna be released on some dark web market.
And yet you have to provide this information if you wanna use the Internet, which is crazy
yeah we have digital footprints everywhere
Well, at least reassure my safety if I’m gonna provide all this information and it’s not like I have a choice
my mockup
I mean, if you wanna be a modern person and today’s day and age you’re gonna need to use the services like Gmail and
Oh dude, that’s sick
Hii why the video option don’t appear always? What I need to do ?
That’s way cool man good for you
Damn it seems like a lot of people have been getting the video
That’s crazy
You must be happy it even appears
anything uploaded on the internet goes to a database and the server can get hacked
I don't even have it
Dude, listen for God sakes listen, how ridiculous this sounds
The people that are in charge of keeping the citizens credit score
There’s like three agencies here in the states
OK, they have all the information the social, the credit cards that credit history, etc. you name it
They ended up getting hacked and it ended up being like half the population information
lol
For most people it doesn't appear at all. Never.
i use AI for market research and analysis for my product launch and even for creative references and ideas realted to social media campaigns and branding
Dude, if I was in charge this person and his family is going to prison forever
use the old thread you used for video generation can't spot the video button