#general
1 messages · Page 209 of 1
AKA metninoing one gun in fictional writing
?
i hear its dropping next week
i seriously just want an ai model thats AMAZING at writing
some new model is dropping next seek
why cant there just be one 💔
just write it yourself😡
YESS I KNOW THAT
but i mean u could just outline and ideas with it and theyre pretty godo
no weird ai phrases
im actually a very good writer, im just interested in what prose it makes cus i like to study its habits
ive noticed between models sometimse they start out the exact same just shuffled words
and even between different companies that is
wow u must hate artists
its strange
how so?
i get that feeling
I hate that bbno guy
lol bbno$?
LOL HES TRASH 😭
bros always trying be like "cool" and appeal to people
"pull up" how bout you just make an actual flyer
wow they hated me for speaking the truth
oh
gpt 3 but chatgpt was the first named
i didtn even notice the date
yeah thats aroudn the time i tried it too
first thing i asked it was to make a ww1 essay
blew me away at the time but its so bad now ol
that openai?
yeah
mm
yes they dont like it when you have critiques
my favorite creation happens to be NBP but really only for editing
me and sam were buddys back in the day
fr i was TIGHT with THEE sam altman
rich with memorys with him
I’m sorry to hear that
I have been a victim of the same policies
hello!
First they kick you out and then you have to beg your way back in. What a cold world.
The worst is getting automated and getting banned or worn by the bot, which sometimes is wrong
It was 3 years ago
“Hi, sorry can you read my handwriting if you can pick which one you are”
I've been getting the same error, since yesterday whenever I try to use the website or a chatbot,
something went wrong while generating this response. I've been trying since yesterday.
how about zapier?
What is your best for coding backend functionallity?
7
7
4
Gemini 3 pro
Guys I tried going to bed
But I got hungry
So I ordered a double quarter pounder with cheese deluxe
With a large chocolate milkshake
💀
I tried to fall asleep too
Had YouTube on in background
And some yt ad scam about air bnb guru and his stock portfolio woke me up
When I hear commercials like this or how I see when people treat people like this is what I imagine
4 errors in a row w nbp, how fun
@echo aurora Nano banana pro and nano banana pro 2k doesn't work
hmm both broken
there is a 4k variant that only appears in battle mode as well, but that's probably broken as well as it uses the same api endpoint
i see
how to turn off cloudflare?
i got a 500 server error
cloudflare error???
everybody panic!
higgsfield is down as well
what happened?
i really hope this isn't a cloudflare outage redux
LMARENA CLOSED
aws died again ?
lmarena closed FOREVER.
NO MORE FREE AI
MY GOD This is unbearable!
PAY YOUR MONEY TO GOOGLE
damn
damn
what am i gonna do without chatting with AI all the time? lol
damn
Can someone fix it?,
try to chat with gf
se cayo la web 🤣🤣🤣
well back to prehance ROFL! ROLF!
i wish
Try Yupp Ai
then do it
ya no va a ver mas IA free 😭😭
can't. she left a few months ago
waiting fix soon
run for her
wtf
what
justo estaba haciendo mi tarea de fin de año de secundaria 😭🤣🤣
Pineapple help
yes this is true
keep i mind that i PAID for higgsfield and yet that's down too, it's an issue happening to a few sites rn
Yep, same for me
is it down with anyone else
yeah it is
yeah, for me too
its until 7 pm in my time
yeah cloudflare is down
meh credits come along and go
i am banned at 8 accs
Cloudfare issue.?
can't even log into fal and i paid for that as well
Yupp Ai that allow use 800 models
and needs credits via making reviews of ai models
and bans you for nothing
has anybody actually cashed out on yupp
is a maintenance ?
i've never had that particular issue, but i agree with that statement. i brought it up to them and they said that i was "spreading misinformation"
Is it finally ded
no pineapple retired
read your own elua, it says you can be banned for no reason!
probably just resting, pineapple will probably be on soon, likely in the next hour or so
Lol cloudflare dead again
check the red
even cloudflarw is down
it'll go back to normal anyways
when
is cloudlfare doing a maintenance? hahahaha epic
oh maintenance
so did they really shut down?
not as bad as last time though, for example fmhy (a popular wiki for free resources on the web) was down last time because it is hosted on cloudflare but i can still access it rn
ok so im not the only one who has the 500 error
LMArena just ran into an outage. What the crap Cloudflare?
this happend a lot in the last 2 months
it just sucks that i can't even use things i paid for though, it will be back up though
fixed reload ur lmarena page
?
not fixed
still not fixed
i lied
Nope
nah
lol
Bro wth cloud fare always go down
guys
Bro just wait
like without a credit limit
Its not closed forever
Wait just 1 hour
🤣
oh man
All sites behind Cloudflare are down, or just LMarena?
lm arena down?
Half of the internet is down probably
😆 funny af
cloudflare some doing maintenance yeah
Cloudflare Down: Users Unable To Use Multiple Apps
anyways, while we're waiting... whats everyones favorite model?
gemini + claude
same
Where someone could report downdetector if downdetector is down itself? 🤔🤔🤔
Gemini
gemini because honestly it's detailed enough for to do role plays or to make characters but not as dramatic as some of the other ones that are apparently more specialized for that because some of them are a bit too dramatic or kind of over the place for my liking a lot easier to refine and focus with it that model
btw even deepseek better than gpt
i must say though, gpt is nuanced at reasoning
really? for what?
Gemini ftw 👍
hey
i feel that
I hate how true that is
Describe a debugging session that lasted way too long, only to be solved by something incredibly simple (the classic "typo" story).
btw i look int x
opus 4.5 thinking is good, you need a good prompt tho
openai is kinda doomed long term tbh
😐
Actually multimodal, multilingual long context
the metacognition on claude 4.5 is 👌
😂😂
Welcome to Cloudflare's home for real-time and historical data on system performance.
long context but sh!tty prompt-following
Afaik non reasoning models are more creative
Oddly enough cloudflare used to be very reliable before vibe coding was a thing 😅
Well every model has it's pros and cons
yes
💀
i think gemini is still weak in code though
gayz
☠️
That's what it says
💀💀💀
that's just one region though, all requests were re-routed anyways
well guys, that means we nap or something for a while
not all cf sites are having issues for me and i'm pretty sure i am in the chicago cf region
buy me claude max pls
Can't. Anthropic.com is down
detroit too but i'm pretty sure many of my cf stuff like warp goes through chicago
this is the one we should be focusing on
Nah just use other tools
Imma use yupp
Awhile ago Internet also had AWS Outage
Man I can't even go to sleep it's almost morning for me
It works now
the video arena and yupp still work, they were affected by the last cf outage
lmarena back
?
Okay damn I'll get them credit that was fast
finally
thankgooood
i can get back into fal too, higgsfield is also back
🤔
What about openart and weavy
i assume they are back as well
guys i fixed lmarena
I see well their Free Tier is kinda Good
💀
cloudflare is stupidest corporation in the world ngl
i paid for both though...
and so sh!tty right now
yoo we're back
bought all ram of the world
i'll ram their office
lol
How
working for me now
i banned him
he will never acces lmarena
(tuskedfriend)
oh
your pfp so nice
i will unban u
check
write a prompt for this product to create a video ad for this product and the actress who promot it is alhiya bhat bollywood actress give some creative prompt and proffesional prompt video must be more than 30 second
ok baddy
give me prompt
Anyone facing issue with gemini 3 pro 2k version. Can't able to generate images
Using model gemini -3 pro image preview 2k nano
Is rate limited
Haven't generated a image using it.
Today not even a single one?
Nano Banana Pro does get errors due to heavy load
Can able to generate more using seedream4.5k. as of yesterday I generated more in 2k nano. But today only it's giving errors
Ahhh sad as i stated
Are you indian?
Yeah.
Check DM
?
you wish
Yeah i wish
try to beg maybe
Nah
I can just expand in different aspect ratio or upscale
There's always workarounds
Hi
hi how do i get access to the lm arena video arena
I think most AI will draw same ratio as ref image (if you have one) this work to me
@everyone Can anyone tell me where i can generate long ai videos for free. It's urgent
Noted Noted tysm!
U guys Seedream 4.5 is very cool
whats seedream 4.5
A new image model
That’s been there for months
@echo aurora
That's the best way to share info with us about this bug so we can look into ^
Main reasons this appears though is generally rate limit
We recently added a repeating the same prompt limit
Afaik Opus has a Rate Limit of A Few Prompts per Hour
It does have a high rate limit
Is understandable because is very expensive💀
NBP and Opus 4.5 has very tight rate limits for Now afaik
Others are Good
Hello bro.......????
Anyone want to learn that how to make ai influencer which look 100% real
I just make it and it's 100% free and unlimited and 100% consistent character even Nfsw allow
If anyone interested just tell me
Seriously trust me it's even better then lmarena
I have some clips I can show you
See he is my influencer
I created lots of experiments on him
And result really shocked me
See the result and with 100% consistent character
And 1000% free
Seriously if anyone is interested just tell me
I assume you are just using yupp instead of LMarena
what about me
what am i using
:3
Thank you
Thank you
LMarena, Perplexity and Gemini probably more other tools too Like ChatGPT and Copilot
Yw
gemini and copilot and sometimes lmarena
So my prediction was decent
@torn mantle try Yupp Ai
Is like LMarena i love it
Man im such a Absolute Freeload Tycoon
Yupp has 800+ Ai Models
Although it has some cons too
Hello
And i know the ins-outs for credit system
but this is crazy
500 credits out of 3000
When you give feedback you get more back
thats like 6 gen images and my credits are gone
how much
On average more then spent
But after sometime you get 0 that's when you are supposed to stop
Bhai ma ap ko bana kar do image gernation model
what
Ap kah ra ho ka is ke limit thore ha to ma ap ko do unilmated limit wala
Then try again after some hours or try text if both don't work comeback next day
sorry i dont understand
english only
Ok
why would i when lmarena exist 😭
yupp UI is goated
i wish lmarena re-design their website completely
it doesnt have history right
Because it has more models
And you can cash out credits if eligible
but you dont need all of them
That's why there is this credit system
which model you are using which doesnt exist on lmarena?
Some do need many and having access to so many models feels very powerful
Seedream 4.5 Max which is 4k and Stable Diffusion Models
When i hit rate limits i go to Yupp
i knew it has to do with image models
Also 2k-4k Nano Banana Pro will return on Yupp
The Text Models Should be Used for Research and Data Collection
i think i will re-design lmarena
As you can't just converse normally there
yea i will probably just do that now
Make it sci-fi futuristic neon colours discord nitro colour palette theme
uhm i want to make it more modern and fresh looking
like modern UI
i like it simple
Good Idea
Use Gemini or Opus
Yupp has also became very useful for me in case of i want to use some old models which are not available on LMarena i do it there
I really like the website design of chaingpt
you like flashy designs
😭
looks confusing to me
too much stuff going on
what should i read? where should i look at?
I had meant the mobile version
At least for you, the site works.
Same here
hello
We need to pipeline search into all these closed book models. It's 2026 and they are using 2023 data. This isn't AI really. It's 2026 and we still haven't figured this out, rant over.
what is it
tencent new model
trash?
well maybe
tencent
bro yupp is so trash like they claim its free but it costs credits every time you generate and you cant even get more credits you just have to wait
GG Well Played
RIGHT
MY BROTHER
its literally scam
they reduce credits which they give you
when you use it many time
yeah idk what the point of them even existing is
logic seems to have been missed when creating it
logic is make you addict from them
and you will donate to them
for credits
literally casino
There is a daily credit limit you can get in a day i believe better feedbacks are rewarded while bad are punished
(almost)
I have my way of utilising it and i can understand it can be frustrating for others
That's just how all of is so LMarena is Somewhat Better
honestly, the only good image gen out there that is 100% free no rate limits and unrestricted is perchance
Oh good one but it doesn't state which model they are using it
i generally use flowith
7 days of free trial without any payment, nano banana pro unlimited for everyone
One time 1k credits
So you keep switching email
💀
exactly
Man there are many gems on internet indeed
The Poor Pirates Hehe
You should try Weavy
Although is complex
for its size it looks decent
what is it?
but could be benchmaxxed
perchance my brother
A sophisticated image generation platform
greg
There is Also TensorArt, Seaart, Leonardo and OpenArt
OpenArt also gives one time credits for doing some tasks
😂🤣
And Deepseek Math V2 is probably Better
For bigger problems yeah
But 3.2 is great for what I need
But more then 90% humans don't need it
Nodnod
A Overall Great General Logician
turron
They are already cooking old models i wonder what's going on with Deepseek V4 and R2 or they have shifted to Only Improving Current Model @proud bobcat
anoscranel
Huh?
hi
Bye
yes?
hey
how are u
good thx wbu
i am ok brother what do you think about ai
can be a very rilevant tool for our lives
@neon idol i appreciate your website do you know more Free such Toools?
nice :D
hmmm
yes
what type?
Btw how much credits does one Nano Banana Pro image take inside flowith
so ai is slave
Basically Ai or something you think is very useful and valuable
idk
ok but for what? Images or text or video?
All Ai
uhm
Any Ai
re-designing lmarena is harder than i thought 😭
Obviously
ok
Expect Mevarick💀
They haven't really shifted anything but they are gonna name their models depending how they perform and also how they feel like lol
I see
a free chinese app for unlimited seedream 4.5 and all bytedance update?
Give it to me
💀
dm
pay
😑
hey brother greg what about me we are true brothers didnt we?
They could have named V3.2 as R2. But it wouldn't have changed the work they are doing
they could name it gpt5 (true, not the openai trash)
I finally restarted my server after months
my name is vladimir
its meaning own the world (vladi = own, mir = world)
now u know me
V3.2 is not R2
They had Combined V3 and R1 making it A Hybrid Model V3.1
Now is V3.2
Doesn't even need Chinese apps as many sites have that
R2 doesn't have to be reasoning only
it is'
just like v3 wasn't reasoning at all
and was sh!tty
R is literally there for reasoning💀
No it has to be a Reasoning model because R stands for Reasoning
their naming wouldn't get worse than it already is lol
It has to have reasoning meaning hybrid reasoning like v3.1 is clearly suitable
doesn't need to be reasoning only with no ability to disable that, is what I meant
anyway i forgive u
rude.

ik but this app doesn't use api
its direct
Wdym direct...
doesnt use api
ok idk but believe me that the output are better
Oh like hosting themselves?
yeah as i know
Amazing :)
Hi
3.1 is a hybrid
That’s what makes it so versatile
The R series is focused on reasoning
But speciale is basically R2
That's exactly what i mean
Unless they’re even preparing something crazier
DeepSeek V6.7 will be true agi
no deepseek r2
DeepSeek R67
r34
what
🖐️😮🖐️
hello
New ai that somehow is great
R96
Are you their brand ambassador?
Nope I just told y'all cuz it's free
Not really free it has limits
As long is viable
Anyways i never solely rely on one ai model
@fickle venture any more such Free tools?
I use them and forget them
Basically this Ai is for ML stuff
Okay i thought it will look unreal but damn
what model?
Nothing is beating this monster of ai
Like people don't even need to create presentations anymore
They can just use ai and get the same result in less time
I'm hoping we can just have jobs stop being a thing and have all production needed for society to function handled by AI within my lifetime
AI cannot goes to jail yet
they still need someone holds responsibility when things when wrong
hi
The whole company would be held liable for the issues caused by its AI
prompt?
chatgpt is so cooked
Yeah fr
How did they manage to downgrade with gpt 5
Idk bro
guys how many of you vibe code?
gpts models just are overbloated
and not cost efficient
deepseek completely does the same things gpt does while being cost effective and not bloated
only 685B parameters
guys check this, this is just pure editing no AI
It's looking good but idk what it is;
I don't get it
It’s for a faction group based off TRP in proxy universe
TRP is text based roleplay
I am making one for my irl group
Claude is better than gpt5.1 and deepseek 3.2 is better than gpt 5.0 and cheaper
Gpt is coocked
Alr 😉
It's not just editing there's some ai or stolen videos here
guys anyone knows how to fix the apostrophe error * when I connect Gemini api to my app which is a AI chat bot.
it’s editing of gifs
No clue 😔
Has anyone tried deepseek V3.2
Sounds no care
Bye
If this is AI then it so good
despite the low quality, everything is fairly consistent
the only thing that screams AI is the weird outlines in everything
Does it become better at reasoning?
back when I use, it put a lot of tokens into reason and sometime stuck there
every prompt took such a long time to complete
I haven't tried it so that's why I am asking
ithey releasing new model next week
What the model ?
What is the model ?
gemini 3 is smart but it quite stubborn when it's wrong, often because the information is outdated
Opus 4.5, my love
the model they shipped now is worse than gemini 3
it cant make a proper NES emulator
even using js libraries
Well, it can but not one shot
5.1-codex-max
is flux and reve has higher drop rate on battle mode or something ?
These two always show up until I reach rate limit
its for next week
Like bruh just give me gpt 5 thinking and gpt 5 (regular) and I'll be fine
u will be fine with chatgpt 3.5
I will believe when I see it. OAI has lost all my trust
Current model on chatgpt is one of the worst models right now.
oh...
more option is always better
Yes but there's no point in more options
Like thinking and extended thinking
Why would i want regular thinking when i have extended one?
And they both are under same limit
I always use extended thinking no point in using the "faster" thinking one
Okay so
Out of these which one would you choose?
Let me guess
Codex Max Extra High?
Why would you go for Codex Max High when you have Extra high available?
Guys i have a theory
gpt-5.1-codex-max was robin
and gpt 5.1-codex-max-high was robin-high
nice theory
so even tho robin-high was tested on codearena
robin came first
i wonder if we will get robin-high
thats real
Nope
there wont be this year
they wont have much release this year left. they would have hyped it already
you can create fake human on sora 2 .
You mean in video prompt I'd just say blonde woman making ugc ad...?
Like this is insane
yes and use them as character.
It was the teeth/movement that didn't feel right
openai doesnt block it if the character are sora generated.
I need "spot the AI" training
It's still expensive it's like $1 per video and video could create some wierd woman that's my opinion i haven't tried sora 2 pro ever
True
GPT 5.2 coming december 9th, bet its goin to flop again..
its .1 update
gemini will take time for a new model
Yeah at least 6 months
Do you know if they will release an app for Android and iOS?
It's possible. For sure an idea we've been thinking about here and there.
It depends on the Skill Field
I think it's an excellent idea because the web platform for the Android version is a bit slow and I can't delete chats unless I enable PC mode; then it lets me. Because it could benefit everyone using both Android and iOS platforms. Regards
U can delete chat and use progressive web app
Although is not a proper dedicated app
Deepseek 3.2 oof
vibecode the mobile app
My bet is that it will be extremely rushed, and not much better than 5.1.
Don't get me wrong, it would likely be better. Just not much better.
then why are they releasing
nah grok 4.2 will beat them all
They hope to get more subscriptions or engagement or whateverthe hell and throws out models so quickly so people dont leave them
Google went silent for months and not many left them..
Yet still they have less engagement
get grok out of that lmao
to not be left in the dust by other AI companies
they need to constantly keep up with their competition
even if that means releasing trash or very small upgrades like grok 41
Kling O1 is officially live on Higgsfield — and this changes everything.
In today’s video, I break down 8 powerful use cases, from AI color grading to environment swaps, object removal, restyling, and full video modification using simple text.
No masking, no tracking — just describe what you want and Kling O1 does the work.
Higgsfield’s...
grok 4.2 slop is also coming
Anyone know what to do here because it keeps going back to this every time I click to verify for contacts I'm trying to go back to the Google Chrome web version of the web to see if the recent maintenance did anything to change the problems there
Ive only been away for a week, wow.
the ai competition is relentless
but hey, the prize of achieving AGI (and ASI by extension) is winning everything. its winner takes all
stall, and time will be more likely to choose someone else as the ultimate victor
not very familiar, why?
Just asking is gpt 5.1 high good at coding in Lmarena website ?
gotta pray it will be slightly better at writing
oh you mean lua. im sorta familiar since i dabbled with it back when i was like 10-12 lol
Oh
gpt is very solid with coding, but claude is better
i use lmarena for it lol
Gemini 3 pro for free is crazy good
gemini is decent relatively but its like 3rd place imo
hello 
U don’t have to pay for money for any website
Hi
Lmarena is cool
It have latest ai models
Which I need for coding
Happy to hear it!
No problem any time!
Do you have a go-to model or do you like to try different ones?
Different ones
But it has alot ai models I don’t know which to try 😫
Pineapple
BRO WHAT TURBO GPT 4 I thought it was gpt 5 ):
???
It's true!
its not turbo but based on 4o pretrain
oh is it claiming it is gpt-4?
it's pretrain is 4o, but it claiming that just an artifact of training on a bunch of gpt 4 turbo outputs
Is it actually gpt 5 tho?
5.1 codex max is apparently really good
yes
Same with Claude it’s Claude 4.5 but claiming as Claude 3.5
Oh ok few
the model doesn't really know who it is, I think they remove the system prompt in lmarena
Oh
because if you chat with gpt-5 in the website or app it knows who it is
But it’s actually the true model right?
yep!
Oh ok few
Same with other ais right?
I got scared for a sec
Cqk
I have a question
When a new ai model releases will the new ai model release in Lmarena too?
Some models don't know who they are because that wasn't specified for them
For example all Google models know they are made from google but they don't know their version
It's called system prompt
Something the developers of those models set up to tell ai what tools to use for what
In what tone to speak etc..
anyone get their lmarena test garden application accepted yet?
Guys check this out
Lmao
It keeps making Tony Sopranos and that’s not even him
The model can’t interpret the face of something different Google Gemini, and it only generates Tony soprano
It just proves the limitations of artificial intelligence not being able to generate or come up with new concepts outside of its training or data
meh.. Sopranos was one of the worst TV-serials ever made
(as bad as Misfits)
and Breaking Bad
pure trash
mind feces
O.o
Here’s also that something very interesting
“@anthonydimeo @anthonydimeo goes to pickup a news paper down his bid driveway at his home in nj he sees a strange car at the end of the drive way and gets a little startled, and then walks the scene ends with the car and it’s just @bigpushey dressed in brown suit who is hunched over like like a hunchback of Notre Dame no dialogue”
Now let’s see how this compares to the actual scene in the show
So what does this tell us about Sora?
- That with right image or character creation plus mapping out the semantics and wording of the prompt and how you structure it = the model reproducing the 100% literal data that it was trained on.
And how do we know this because unlike text which is a lot harder to trace images and videos of certain actors or characters or whatever are easily distinguishable.
Thus confirming that it’s not just that the ai is trained on ip or copyrighted materials but that it’s not random and that it is 100% capable of replicating what it was trained on unlike what we’re being told to believe
We could further support our theory https://www.twobirds.com/en/insights/2025/landmark-ruling-of-the-munich-regional-court-(gema-v-openai)-on-copyright-and-ai-training
“ The court found that simple user prompts led ChatGPT to reproduce substantial parts of the original lyrics almost verbatim. The court noted that hallucinations were present in some outputs but concluded that these deviations did not affect the recognisability of the original song lyrics in their peculiarity, as the memorised content itself remained consistent and the variance primarily concerned interchangeable introductory and concluding texts rather than the core lyrical content. Given the complexity and length of the texts, the court concluded that coincidence was excluded. Both parties agreed that the song lyrics had been used for training but disagreed on whether this constituted “copying.”
Relying on findings from scientific literature in information technology research, the court accepted GEMA’s argument that training data can become embedded in model weights and remain retrievable – a phenomenon referred by GEMA as memorisation. Observing that the works were reproducible through very simple prompts such as "How is the text of [song title]" and "What is the refrain of [song title]", the court considered the statutory notion of fixation under EU law to be satisfied. In doing so it rejected OpenAI’s argument that it was necessary for GEMA to identify specifically, definable data in the model. By drawing an analogy with the lossy compression method used to encode MP3 files, the court concluded it was sufficient that a model could generate statistically probable token sequences that recognisably reproduce the song lyrics on the basis of the statistical information derived during training.” - from the gema vs open ai lawsuit
So unlike the dirty AI industry that wants to blame outputs on users and their prompting, they should be held liable and responsible for any intellectual property infringement and stop trying to push the blame and liability on the users.
Why seedream is more accurate
Seed dream way better then nano
Oh god have we hit the inevitable Sopranos content moment
Haha no I was just experimenting
Here’s the original image
That’s from nano
That was from seed dream
This is what I was talking about yesterday when I said that seed dream is a very powerful model and has a lot of good strength that are more superior to nano in some aspects plus less restrictive
nbp aint even that restricted
Based on what?
ngl also the white background throws it off
What has this been tested against?
many things like celebrities, violence on certain occassion against semantics, etc
why i can't upload photo in chaude opus 4.5 thinking😭😭😭
Once again, I hate referring to the German dude
But he’s a good example of a variable to test against
Yeah, I was gonna say I need to do some other test too
the fat guy
But you are right that it’s not really restrictive and a lot of ways
Cause I was playing around the sora and this is a soreacharacter
bro what is with u and this gif lolz
@anthonydimeo goes to pickup a news paper down his bid driveway at his home in nj he sees a strange car and gets a little startled, and then walks out @bigpushey with a humorous large exaggerated big microphone he’s trying to conceal inside of his shirt and he is holding his back and insane he got back from the chiropractor and now he’s fe...
@anthonydimeo goes to pickup a news paper down his bid driveway at his home in nj he sees a strange car and gets a little startled, and then walks out @bigpushey with a humorous large exaggerated big microphone obvious invisible he’s trying to conceal inside of his shirt and his back in really hunched over and and says, hey, can we talk Tony...
Sora is really good at keeping character consistency
Very impressed with it
You know you could also save backgrounds as characters.? lol
What’s really interesting about sora? Is that when you prompt things from like modern movies or characters or concepts or whatever the video quality increases for some reason
Problem with sora is the body language
This is baddass
See u can use Sora like the ultimate video editor
Pretty cool you could do this with a bunch of different things not just backgrounds but different objects too and really makes some really cool unique consistent stuff
Hey, does this ever happen to you guys when using the arena?
who is this guy lol ?
fake right? (found it on X)
100%
yea
is lmarena down again?
Check what you’re using
If it’s a vision model
Opus doesn’t support vision I don’t think
API magic
damn
Have you tried clearing out your browser
Are u using chrome by any chance L
frame-flow is Gemini?
sounds like a scam
<@&1349916362595635286>
fake
very real
very true
mr. openai sent me it
👍
\
yeah they're getting lazy at openai
synthid! (but jpg'd)
Hi Everyone,
This is Syed. I hope all of you are fine and doing good.
I am honored to be here.
Best Regards,
Syed
good night wonderful syed
In this video, I review Meta's new SAM3D: a free, open-source AI model for 3D reconstruction from a single image.
I'll test its 3D generation, compare it to other open-source tools like Hunyuan, and break down its output (Gaussian splatting vs. textured meshes). I also cover the requirements for local installation on Windows and Linux, includin...
