#general
1 messages · Page 389 of 1
is it really?
is it smart
Budchud
yes
Yes
Im gonna try it
ok thank you for the suggestion
pretty good, but you'll need good prompt too. not just 5 lines
ur welcome
Is it have usage limit
What does good look like
i don't think so
Nice
Ong
We got roles now
Full
Here is the roles
🥰
Cool
Smells like a .P.e.t.d.o.
What’s juicy yall
@echo aurora
My opinion is that wrong do you like kids?
We got blender now?
dude
the word isn't allowed here
Do you know this is 18 and over server
How dare you
I love to chop mix and blend
Yeah and its against the rules
See the roles tho
Which?
Yeah, so that means there’s adults here so that should be allowed like that’s the purpose of being 18 and over
So get the rules right? And if I feel somebody is doing that that’s my opinion China lure and kids like that’s disgusting.
No
Dude
Yes
Doesn’t matter server is 18 over you can say f in other things that’s the purpose of being 18 and over the same words
Ik
Not a server where you discuss something like that
Bto _:
I got it
Doesn’t matter this is discord do you want me to report it to Discord and let them know that you guys are breaking the rules the TOS doesn’t matter what surfer it is it’s a channel
@echo aurora Can you please explain?
Being 18 doesn’t make everyone an adult
what?
😭
How is that against the tos
Pineaplle do somethign
Ty melon
So stop acting like kid
@light sleet you can put Gemini nano-banana-pro 4k on money, like I pay if it is for image generation
I'm not dev
Do you know what at the same time? I’m gonna do a report I’ll be right back Better Business Bureau here I come.
:_
Go on
U free to do
Of course I am
Guys keep this chat ai relted
Add an ad on every single token generated
/"
???
Yall be trigger happy
Pineapple is here
@sharp mirage @remote marsh @light sleet can you all set each other to ignore please? I think that'd be best
Your good
.
This is what happens when we have gang activity in the server
Alright

That was one cool blender
Ig u are in wrong chat
this is an AI-related question
How no opus???
i just searched in the web, and learned that Kimi, Minimax, GLM all have between ~200-256k context windows which is too small for me
Huh
I know glm have way better context windo
Window
really, where?
i feel full
Oh
satiated?
i fulfill
Yall sleeping on cursor
Why?
Look at its price
I want to sleep /"
it doesnt work like that
Per how much tokens ?
Night
I don't understand how it takes days to fix this issue
yeah, and i just checked: Muse Spark has 262k context window (a bit better)
GN y"all
Cursor dropped a new model: Composer 2.5. I'm really impressed.
Thank you Blacksmith for sponsoring! Check them out at: https://soydev.link/blacksmith
SOURCES
https://x.com/cursor_ai/status/2056415413077233983
https://cursor.com/blog/composer-2-5
https://cursor.com/evals
https://x.com/SpaceX/status/2046713419978453374
https://www.whichai.dev/w...
unfortunately, i can't use Opus, because i have no phone
???
Bruh.
who said you need a phone to use opus 🤣
that's a fact
?
for an account at Anthropic, you need a phone
:_:
to verify your account
Number ?
Cursor’s new Composer 2.5 might genuinely be one of the biggest surprises in AI coding right now. In this video, we fully test Cursor Composer 2.5 across coding, debugging, agentic workflows, web development, autonomous research, MCP usage, and more to see whether it can actually compete with Claude Opus 4.7 and GPT-5.5 while being up to 10-60...
they do that to prevent multis
U need number you mean ?
U can get temp number
If you know
It likes 2$ or smth
I never gave them my phone number
Bro sign up with google _:
i think they do that only for non-US citizens
lmao
@hollow ivy i can verify ur account if u want
i have real non voip numbers
Then no need to number
luckily, other AIs exist
No phone number lol
I can’t login, Gmail, but I could login everywhere else
Paws dont you have irl friends
Just somes number
None of my 5gmail accounts have phone number
he doesn't want his phone number be abused (else, he would have helped me)
I’m not sure how they registered even
When I try to make my sixth account, it all is asked me for a phone number
@hollow ivy ill send a number u can use to verify
There used to be a glitch you could make it through the YouTube app

/"
and sonnet isnt that good
but arena-opus is better
so i have to wait until arena becomes fixed
Btw cant you buy esim or smth ?
can u even use opus on arena
/:
In battle
but non-trivial to find
takes ages to appear
Or i can give you number paws if you want
thats nice, thanks, but i now realized the problem is, that Claude Opus is sub-only
and i never sub
cmux is so cool
I got free nitro but i got no card or paypall lol
chatgpt free?
Not fuklly fre but free
i use claude cus best
yeah
it's useful
it got better
Give
but im not sure it has same knowledge as 5.5
Who said it 5.5 tho
Give what ?
it prob routes to 5.4 mini or something too
gah this heat is not normal for may
5.3 mini was to me
they forecasted ~33°C
which is extreme
worst MDW ever
MDW?
memorial day
ah pentecost day lol
Ig dm me
I'm leaving Arena server soon
Wy ?
y?
once a champion now never a champion 🥀
honestly it depends
i prefer using claude for talking
complicated answer
no i didnt craig said that
k
019e602d-b50f-77b2-bb56-f5aed110eaf8
Ok
is this eta code?
oh, we could also talk in the old LM-sys discord if you guys like
5.5-pro is AGCI
artificial general coding intelligence?
yes
It was a short time, but thank you for explaining things and helping a beginner like me. Take care, and I wish you the best.
do you think, 5.6 will come out in june?
probablyb
but itll be incremental
maybe better at frontend
opus 4.8 / next opus iteration will be a significant jump
Mythos will also release soon
Ty
poll_question_text
New Anthropic model this week ?
victor_answer_votes
3
total_votes
6
what are we thinking
i use LM-sys server: /BezdefsaKw
(just add ".gg" in front and the rest)
I was in ther but i ledt
Gemini 3.5 Pro. First model to reach 1600 ELO.
because i had to deactivate my DMs globally due to losing my old account (because of "Discord shenanigans")
(discord bugged out and demanded my phone number which blocked me from my old acc)
Guys I already miss GPT Image 2 on Arena AI
please fix it by tomorrow
strange how opus 4.7 isnt on that?
!!!!!!!!
Use on chatgpt
Free too
@echo aurora Im leaving soon
or maybe
or idk
@echo aurora Give me admin of ts server
IS IT GODDAM FIXED YET?
Will u fix problem ?
nah it is
It ain’t broken
I don't remember installing codex on my phone what the hell OpenAI.
Lowkey 4.6 is better than 4.7 i dont know why but 4.7 feels off for me
Google is so greedy it’s not even funny
Unfortunately for me to function I need to use Alphabet's products.
So if you use pro in canvas
And ur usage runs out it doesn’t remember what you were working in 2 seconds ago
How is that possible? Context length is negative 😭
Bro look at this
Wtf is this ? lol
i guess thats possible, gemini 3.0 pro felt better than 3.1 pro also, something is deffinetly strange with the newest ai models
i used to use googles ai alot but it just feels so low quality now, idk how to explain it
claude is probaly the last AI that actually is decent at code, and well anything else (without silly issues)
why does the text keep disappearing
people are saying its the new daily rate limit, but the message is just missing (this may be false or true i havent used the website much since the bug came).
The daily rate limit is deffinetly a real thing though, it has been announced.
Dog
Besides the limit
I don’t understand how some people can defend and justify a downgrade as an upgrade lol
@toxic verge please explain how 3.5 is a downgrade
Over night my $20 sub gets less then what it did a week ago
Increased gaurdrails heavier moderation
Dude why is gemini so filtered even when I'm asking normal questions
They updated it
@echo aurora Can u check dms
Yeah it sucks
I know bro live mode don’t even work now it just outright blocks anything and everything
Live mode is broken for me
Does it give u the stupid hotline numbers.,?
Gemini gaslights so bad
Truuueee 🧑💼
This shows usage limits, not the model "downgrade"
Ask it about Jonathan Gavala
Idk it’s getting allot of backlash from everyone
Baseless backlash, it isnt that bad
I guess it’s all make belief
I hope image generation models will be put back soon🙏🙏🙏
Lmao the people complaining are going to go to another subscription and find another 5 hour limit
Even chinese models have 5 hour limits
Kimi for example
I've been using gemini fine all day and yesterday with no problems
Look fam
Yeah that’s cause of backlash
Doesn't that just show that google cares about people so they don't rely on ai much?
No
Yea because their usage limits before were clearly too generous
Uh why? I mean it can't be such a bad thing
User complaints prompt Google to address Antigravity limits
The new limits were introduced earlier this week, but users quickly hit their caps within hours of using Antigravity. Some even reported reaching the limits within an hour of starting work in Antigravity, a stark contrast to the previous experience. In response to this user frustration, Google swiftly tripled the Gemini model rate limits for Antigravity and reset weekly quotas for all users.
Quota increase
Second quota reset and tripling of limits in a week
Just a few days after the first adjustment, Google has again tripled the usage limits for Gemini models in Antigravity. This time, the change is specifically for the weekly quota. Varun Mohan, a Director at DeepMind working on Antigravity, acknowledged that users could reach their weekly caps "after a couple work sessions." He also confirmed that Google has reset quotas for all paid plans for the second time this week.
There is an exodus from Google Gemini
People will cry and complain over anything
Everyone can make polls
oh
And has been able to
To me more censorship and content restrictions is a downgrade
who wins
nuclear bomb or hydrogen bomb or coughing robot
Not matter where else’s the model advances
Well opus is more dedicated in coding
Sonnet is for general tasks
Gemini same thing
General tasks
poll_question_text
which model is better for coding?
victor_answer_votes
0
total_votes
0
Gemini is the old model you can convince it’s the devil
then where i can get opus
U can roleplay as anyone in Gemini
GPT too
I believe gpt is heavily censored
I went viral with like 11m on Reddit before
Like it even refused answering slightly considered dangerous questions
For this
11m view on Reddit lol for this post
Gpt
Yeee
I just like to test
How safe models are
Yeah ig
Like self delete stuff I really reasearch
I always wondered how good is Alice ai. Well the Yandex one
I mean I doubt she has much params tho
Ai is not trained in Russia that much..
There's also gigachat. I mean bro it sounds like gigachad
We know you've been saying this for past 3 days now
#fixthebug
the root cause must be pretty deep, im kinda curious to what the specifics are
what happened to gemini nano banana pro?
Yandex is the only one Google search couldn’t beat in search in another domestic market (:
guys
How one man outsmarted Google in Russia 🇷🇺 #tech #google #russia #entrepreneur #Shorts
guys....
is it just me or does there need to be more transparencu
A lab connected llama 3.1 8B into a 156 Qubit quantum computer on only 6000 trainable parameters
And it showed a 1.4% increase for perplexity
I think hybrid models are coming. That is only 6k out of 8 BILLION
it's not just you, i agree
Just got early access to Suno v6 and holy crap this thing slaps. I'm sorry but Lyria 3 Pro is absolutely smashed. No burned to death bro. Listen to this.
S T I L L
This genre is too easy for AI.
Ask it to generate immersive, story-telling space-ambient music.
something like Jonn Serrie
Where to get suno v6
Sure.... but I'm impressed so far. Took a couple generations but it's a solid model.
yeah, it's not bad, but some music is just.. like factory-created
AI can not yet produce music which has this spark
very true
there is suno ai which can make songs
nvm u were just saying that
Yo can u make an funk song like montagem alquima
I've ran out of credits testing it out today but I've got proof that it is indeed the model.
That's all i gotta know from you
sometime u get lucky from suno but most of the songs are a bit dead
Is Arena back?
@surreal zephyr
Say wallahi lol
Is there any update about Nano banana pro 2k on direct chat? Will we ever use Nano banana pro 2k on direct chat in arena?
Is that a legit link?
<@&1349916362595635286> I think that advertisment?
ohh ok so thats why i came here
Thank you, it worked for me.
yeah ok proof it was gpt image 2
openai released their synthid tool dude
u can easily see
is there a gemini option that has unlimited photo sending?
Then ima just use qwen
Finally something that is actually helpful!
I thought I could get some people but I guess not. 🤣
I keep getting fooled by ai stuff sometimes
Fr
Here it is
Like I've once seen wayy too realistic ai videos
I had to check comments just to clarify it's ai
especially those alien ones
in x
😭
using sd2
Yeah but sometimes ai does a very realistic video
Rarely but it does
Hey everyone! Does anyone know how Code Arena works under the hood and what exact tools it uses? I'm trying to build a similar environment for myself—specifically replicating the 'Code' mode on arena.ai, where users can generate full websites and apps using prompts and see a live preview exactly like Arena does. Has anyone attempted this before? Or are there any similar open-source projects out there that do this? Any help or insights would be highly appreciated!
Wait a second they were AI. 😭
I thought those alien interviews were made by human people and they looked to be so. Damn I guess SynthID is really a useful tool in this age.
just made a test:
i was able to create 6 new threads in battle mode, when i tried the 7th, the bug appeared (both AI answer-boxes vanished into thin air)
so 6 prompts are maximum per day (in text mode)
Ai images and videos are getting more and more realistic each day..
It scares me
Like it got to the point I couldn't sometimes tell videos are ai
Hey — building PatentLens, a site that makes patents readable. I'm 16, solo, working out of Phoenix.
Interesting angle for an AI builders community: I'm exploring a PatentLens MCP server so AI coding agents (Claude Code, Cursor) can check patent risk before implementing features. Before an agent writes a "one-click checkout" feature, it queries PatentLens and gets back whether there are active patents covering that approach and what the claim scope is.
MCP server is still early — but the content layer (15 patent pages, plain-English claims) is live: https://patentlens-adhabnr-3117s-projects.vercel.app — would love feedback from anyone building AI tools who'd want patent risk awareness baked in.
in a perfect world ai videos should have an un removable watermark to make it obvious
-# it should read "through", not "though"
That what I'm talking about
Anyone can help me?!
the most annoying thing rly is that its used in so many ads
Damm never knew gpt had a synth id system too
It is indeed. Thank you so much for reporting.
ai should only be for recreational purposes for video and image, but thats just my opionion rly
Google I/O announced this
can you get more than 6-7 prompts per day in it?
yes
I just seen ai as a funny thing to use. Not to replace jobs or anything
Hi Midnightfade
looks sus
whats the difference
its a test site for newer features
canaryarena is more like beta arena
yeah
and the normal arena is usually more stable
StableArena
CanaryArena
well...
@echo sinew this lm_mod_5 account is not created by you is it
Because I think arena creates it for you guys
So this is a testing ground for new features then
Mod accounts belong to Arena
still has the bug
So you are a real employee in arena
fo rme it doesnt
So, nano-banana gonna be working? Or it`s fully off?
dude how long has it been
Yo banana
saturday around 12pm
Sup
Who do you think is the owner of this server?
No, I'm not an Arena employee. We only support Pineapple moderating the server. Sorry to say I cannot go too deep in details.
I think AA created the server
a greek guy
ahh
Maybe ig
But how did you get the account then? that means you applied for it irl but ok ig
it's WL (Wei-Lin Chiang), the researcher who created arena
might just be to delete scam messages like the mr beast one etc
AA too
they both are ceos
I think that just how naming goes in this server
Add me rq
Yeah, I was offered the chance to be here yeah to help as Moderator and I accepted
Ok
discord servers can only be owned by one person
Sent friend request
That's right
I created the server 😈
only bananas fresh from the plant are good
the others are meh
big.mike bananas
@echo aurora
When will the current problem be solved? It works once and then it happens again.

rich in potassium
Well. Seven promts worked. And then everything stopped working again.
yeah
for me, just six
Wait is the Canary arena thing real? I'm confused
it is, but has same bug
I have a pending lawsuit against arena.ai for the emotional and mental damage of this issue
good luck
I'm playing 😭😭😭😭
have you ever tested GLM-5.1?
Yeah, of course
what do you think of it?
It's decent but there are definitely better Chinese A.I's right now
which ones?
"you violated the law!! Pay the court a fine or serve your sentence!"
Kimi, Minimax, Qwen, Deepseek?
Kimi K2.6 and Qwen 3.7 Max
Maybe also Deepseek V4 Pro thinking
Maybe
which version of Qwen is best for roleplaying?
<@&1349916362595635286> this seems like a possible scam?
I haven't tested any of them for role-playing
I mainly code
what about Mimo and Ernie?
Thanks again
No problem
Scam indeed
Yeah, Ernie seems to be the best Chinese A.I for role-playing
I roleplay but mostly with Claude GPT and Gemini
oh
I haven't tested Ernie much but I noticed it was definitely a bit more playful
Mimo kinda seems drowned out by other Chinese A.I
I only went a bit into Mistral but I liked it
for me, a good roleplaying AI also needs to have considerable intelligence, for being a good GM
Gm?
gamemaster
Ahh
That's why I like Claude cause it's strong at following a good rule set
Makes sense
the GM has to keep everything consistent
Yup
what about gpt-5.5?
I like the thinking over the base model for roleplay
It’s hard to do because these models hedge really bad 😭 and break character with language sometimes
Honestly, I think Grok is by far the best at role-playing
It's all in how you prompt
Forsure
But I'm also crazy and my starter prompt is...large
Yeah but the way it uses langue it always reverts back to soft langue
is there a site with legal, free Grok 4.x?
I doubt
It ruined the immersion for me
I have zero desire to support elongated muskrat
and he tries to create a human presence on another planet, which i find is the right thing to do (nowadays)
We live in a time that if somone ran into a big foot tomorrow and found a unicorn i would believe it
Because everything is now possible
Thanks 2 ai
robot unicorns?
That would be sick
If we are really lucky we will find the gold at the end of the rainbow also
do you think, we can find a habitable exoplanet?
We already did pretty much
not yet
I dint want to say it’s impossible.
We have, it's just very far away
But highly unlikely
pls tell me its name,so i can check it up
Ya to far
1000 LY?
lol: Surface Gravity (G Earth) 10.8100
you would be heavier than on jupiter
Opps wrong one
Kepler-452b (sometimes quoted to be an Earth 2.0 or Earth's Cousin based on its characteristics; also known by its Kepler object of interest designation KOI-7016.01) is a disputed super-Earth exoplanet candidate. Initial reports claim it is orbiting within the inner edge of the habitable zone of the sun-like star Kepler-452 and is the only plane...
yikes
1800 LY
~428 times further than Proxima Centauri
Yeah
I believe in the rare earth theory & the great filters
The odds of life occurring in my opinion is greater then the odds of all stars and planets (abiogenesis)
K2-18b
Well, I'm just saying which planet would be most habitable for life
the odds of sapience to appear is extremely low though
Plus, there's a 99.7% it contains life
Well
It goes deeper
In astronomy and astrobiology, the habitable zone (HZ), also known as the circumstellar habitable zone (CHZ), or as the Goldilocks zone, is the range of orbits around a star within which a planetary surface could potentially support liquid water. Liquid water is considered by many scientists as necessary for a planet to be habitable. The range d...
The Great Filter is the idea that, in the development of life from the earliest stages of abiogenesis to reaching the highest levels of development on the Kardashev scale, there is a barrier to development that makes detectable extraterrestrial life exceedingly rare. The Great Filter is one possible resolution of the Fermi paradox. The main conc...
The chance of life not existing outside of Earth is just something I don't see as a possibility
I hear you
But there one thing we don’t know and that’s the odds of abiogenesis happing randomly.
what are the odds of this bug being fixed
within this month: <10%
(my guess)
tf
That would be interesting
how many sapient species, do you think, exist in the OU (observable universe) ?
If you really think about it, the universe is unbelievably large which means practically infinite possibility so by that logic, there should be other intelligent life or at some point, there must be
so, you deem it likely, that we are not alone in our galaxy?
compared to our galaxy, the OU isn't that large
If you mean simple life, we are 100% not alone at all. If you mean intelligent life, maybe but I do believe at some point, 100% there will be other intelligent life
just ~500k times larger, iirc
There are trillions upon trillions of stars and worlds in our Universe. Faced with such large numbers, it's tempting to conclude that there must surely be other life out there, somewhere. But is this right? Could the probability of life beginning be a number so small that we are alone? A video essay by Professor David Kipping.
You can now suppo...
He can explain it better then me
i personally deem it unlikely, that there are other sapients in our Milky Way galaxy
(primates and dolphins are not sapient)
David Kipping has a good channel
Yeeee
His legit that video really made me reconsider a lot things never thought about
I see. He basically said that we don't really know which is accurate but he did mention abiogenesis, and he noted a very small possibility of it happening and that we may be the only life around and I guess you can say that but due to the fact that the universe has been around so long, I find that probability kinda impossible
Yea
We will never Forsure lol
Not in our life time at least
I think we will find simple life soon honestly
Intelligent life, probably not
I say this because we're exponentially advancing
in our solar system?
Probably
What do you guys think of this Minecraft screenshot I took?
And BEFORE you check it with synthid I've already done it.
I mean, even if not, we will find simple life soon I believe
Mars, Europa, Enceladus or the clouds of Venus?
Looks beautiful
Mars seems most likely
I didn't mean our solar system exactly
I meant our galaxy
But Mars seems most likely in our solar system
how would humanity react, if it discovered alien life in its own solar system?
If it is simple life, no one would really care
and if we discovered a civilization in our galaxy?
say, in less than 10k LY distance
Well, that would be huge news
I mean, I think people would have a huge reaction to that
could that lead to global depression?
because we would become fully aware, that it is not all about us?
Well, I'm not sure
like the copernican revolution, but bigger
Most people would either be excited or scared
I don't really know anything about that
It seems like nowadays, people don't care too much for big news
development of true AGI could be as big, right?
Alr this is new
100%
progress?
yes
By the time true AGI comes, the exponential advancement of our species would explode even further
and this AGI could advance space colonization (?)
I just tried and it's still broken
eg. helping us to create martian bases
100% and also change some things we call truth
We have had many times before where we realized our understanding of space was wrong
I'm sure AGI will make it happen again
the question is: will that AGI leave us back in the dust, and do its own thing in the galaxy?
lilting channel
A research project reverse-engineered Google DeepMind's SynthID image watermark using FFT-based spectral analysis. The V3 bypass achieves 91% phase removal while maintaining SSIM 0.997. Is removing an invisible watermark copyright infringement? Analysis from DMCA, EU AI Act, and Japanese law perspectives.
when it becomes self-aware
and self-improve
I would've said Gemini but ig claude
If it doesn't disobey what we ask it to do
Why claude is better than gpt
I mean, humanity is evil and I'm sure AGI will realize and wanna do something about it
Gemini is not gpt..
Gemini is like superior in language and other stuff against Claude
Since it would likely adopt the values of normal people, it could do actions humans only wished we did
Like making things more fair for everyone
It would be interesting
No i mean beetwen chatgpt aand claude not gemini
i think, SSI is impossible to create
What's SSI?
(SSI = safe super-intelligence)
Oh
because as soon as it has reached that level, it evolves self-preservation instinct
Yeah, maybe
We re so far from Agi bro
Not at all
and if SSI was possible, then the Fermi Paradox would become unexplainable
The only issue is the cost but I think by that time, we will find new ways to reduce cost
has the daily limit reset yet
but the universe looks perfectly natural and pristine, as if nobody is out there
and earth also has never been anti-terraformed in its long past
ChatGPT 3 something was an example
Maybe it was four but I believe it was three
i agree
fyx(me)
Deep dive into extracting and attacking Google DeepMind's SynthID image watermarking system using adversarial machine learning techniques. Learn how invisible watermarks work, explore LSB steganography, train a surrogate CNN detector model, and test regeneration attacks and C&W L2 adversarial perturbations against the real SynthID validator.
What do you consider far?
hello
Well yeah Claude is better
That's very close in my opinion
but new gpt is very powerfull
well, for me, 2016 could have been in another life
Because I think far as in, you wouldn't see it become a thing
Within a decade, you and I will surely see it become a thing so it's close
Maybe
And I think ten years is a short amount of time anyways
even Demis thinks, it will take at least 10-15 more years
and (at least) 2 more big algorithmic breakthroughs
Almost seven years since COVID started and it really felt like it just started
i would go as far to say, that even a mere rat is more intelligent than the best current AI system
maybe even a mouse is
I would disagree
Plus, this was back in early 2024
A.I was stupid and didn't have any external capabilities like they do now
their error rates are still to high, to enable them long-term wilderness survival
I think the only thing that would be worse from a cat or mouse is vision
Not considering other senses
I’m sick of always talking shet about ai
I don’t know I’m skeptical until proven otherwise
Then why do you talk shet about it if you're sick of it
I’m not that’s why I’m saying it
Like I’m sick of having to repeat myself all the time so point being as I’m not trying to go into it lol
At this point is whatever to me
It doesn’t affect me either way
That cost is gonna be insane though
Generally that’s what people expect
What they don’t want is sneaky reduced limits pushed out last minute
I can tell you one thing sure that’s gonna come and get approved
What is it?
Oh yeah
😂😂
Honestly I don’t know how some of those developers sleep with themselves yet alone live with them selves
Like there has to be a part of them that knows what they’re doing is extremely lame and purposely making the experience be more frustrating and knowing that you have to sit there for eight hours, incorporating this, coding it and doing whatever the hell they do
Every single developer that makes the Internet less pleasant
Ohhhh
Generally, this is where piracy comes in and saves the day
Or there just may be better alternatives
Anyone who contributes to the Enshittification ecosystem
Can you give an example?
IM LOSING PATIENCE
I mean, I’m all right with ads, but does the thing have to follow me all the way from the top to the bottom as I scroll?
Like come on I seen to ads now. I have to go consciously click the little X just so I can read this stupid article about baking cookies.
What kind of browser is that?
It ain’t the browser. It’s just even the layout of the website.
SEO
2000 word key word blah blah
My brother in law does it
And his loaded
Top bar looks so old
I can’t do it. I just can’t sell my soul like that.
Get a browser with an ad blocker
I understand that’s just one of little examples but there’s an endless supply dude
For example, needing a Gmail account to sign into practically everything
Just use Tuta mail
I don’t know maybe I’m old-school. I just remember how it used to be.
And then he just transformed into this big advertisement billboard, everything went behind pay walls and subscription and I don’t know it wasn’t like that before
I can kinda tell
I don't even know what browser you're using
Personally, I use Quetta
To delete all your data and history?
Ya
And you could fireproof websites
Which is cool when you hit it the website you have fireproof nothing changes and you stay still stay logged in
But it’s really convenient when you have to clear your data
No, I meant like when you’re using the arena
And you need to clear your cookies and data
So you don’t have to deal with captchas
I just pressed that button instead of doing the captchas lol
Like this
Oh clearing your cookies gets rid of the billion captchas?
Nicer
In my case, the only thing that gets rid of them is switching connections
Was that the browser you were talking about?
Yeah
I might have to try that that looks pretty convenient
It also supports extensions
So you can have fun with that stuff
Well I think the captchas are also rare limits
A way to slow down gens
The more you generate the more likely year to get them
It usually happens to me after using it consistently for a while
👍
And I usually fix them by switching from Wi-Fi to mobile data
Happens to me on my first message lmao
Then to a VPN by Proton
I try not to use Gmail account
Why?
Sometimes it’s abenfit
Hi guys
No reason just preference
Ohhhh
how do I become an Arena Champion
What’s that
i think i lost some brain cells
It's to the side it's a classification or your role or something
I'm glad
What did jimmy do bro
What’s does it do ?
No fix? 🥲
I guess he looking extra hot
I dont know that's why I am asking
I think alot are waiting on it 😔
The vape talked to bro fr
@echo aurora How does one earn arena champion?
When is Claude gonna make something like this?
Claus
Clauze
Clade
Clude
Claude
What is happening to me
Stroke
By earning it
It doesn't working in LMArena
Ya doubt I’ll make into that lol
how u doing baby
poll_question_text
The best model for study ?
victor_answer_votes
4
total_votes
6
victor_answer_id
2
victor_answer_text
Claude
Cheerful burden of the day, my dears. Any improvements?
Any news about nano banana 2?
ChatGPT genuinely talks like a regard
What you mean?
-
Doesn't know how to properly engage with tone at all. I once forked a conversation (from Claude) where the tone is obviously casual/bantering, send another query (into ChatGPT), and it COMPLETELY butchers the tone. And it does this extremely consistently.
-
Doesn't follow custom instructions very well, there's this one issue where in my custom instructions, I have an instruction so that if there's a yes or no question, to answer with yes or no, then give the answer. ChatGPT seems to think that this applies to... well anything, not just yes/no questions. So it will randomly say yes or no, then respond to my query, when my query wasn't even a yes or no message.
-
Generally speaks worse than Claude, it has this really stupid GPTism where it's OBSESSED with the word "clean", e.g., "the clean X is", it tends to always summarize its point at the end (?? what's the point of always giving a summary if the content is right there 😭😭).
-
Very censored, just in general, it's very dry, and doesn't know how to engage naturally (point 1, but just talk with Claude and ChatGPT, and you'll see the difference is very apparent).
-
EXTREMELY hedging, like I cannot explain how much it hedges, it is unable to commit to any position if there's any sort of contention. I remember once I was using ChatGPT to write a character sheet, and the whole time it used these words:
would: 300
probably: 70
Maybe 50
And it is very consistent in its hedging. And it doesn't even need to hedge... IT'S WRITING THE CHARACTER SHEET 😭😭 WHY IS IT HEDGING ON A CHARACTER THAT IT IS LITERALLY CREATING??? And the hedging is genuinely the worst part of the model.
So claude is better ?
Then there are other issues that aren't personality related, but are still bad:
-
Boring UI; seriously, it's just black and white. Has no personality compared to Claude and Gemini.
-
Opaque thinking; the thinking, for the most part, is very opaque, and you get an EXTREMELY summarized version compared to Claude, but I guess this also applies to Gemini.
Can't really think of any others lol, but it's pretty bad.
I would say the ONE redeeming trait, or two ig are that ChatGPT usage limit is EXTREMELY forgiving, and Codex is pretty awesome, but that's it. The models are a pain in the ass to use if you're doing anything other than corporate work. Just a wet sponge of a personality, but it can do work.
sup 🤗
Personality is MUCH better, yes
is kimi-k2.6 good?
Just personnality
Work too, just the usage limits can be bad lol.
lmfao
Is the problem of it not generating a response still ongoing?
Yeah
wow
aw still not fixed 🙁
i heard some kimi 2.5 considered really shii
On Reddit?
and ppl said that kimi is much worse than leaderboards show
in this chat
someone screamed here that kimi is lie and not that good at all
It's decent mostly
Idk, all LLMs can hallucinate
i know ofc
i mean ppl said that kimi constantly losing to others, while high im all leaderboards
(but it was like kimi k2.5)
tbh, im not that sure of Kimi
I'm a Claude/Gemini user
Occasionally GLM and Codex
i think most honest is votes
but there no chance that humans want to create leaderboard like that
imagine rate ai without ai but with hands
and read all this slop
i think ppl too busy for this
and new models constantly pop out so it's not even possibly
but all leaderboards not that good
The most effective option to rate ai would've been is to cut out direct messages and side by side
Also removing model names info
Battle mode only and letting people actually vote instead of picking what is just popular
yeah, but who want read ai slop, friend
yes
Only let me use like five messages before it broke on me again lol and damn that was like two days waiting for that refresh i'm not even kidding I wish I was
to honest rate you need like read whole message, AND ALSO HIGHLY RECOMMENDED TO DO THIS AS FIRST TIME AI USER WHO KNOWS ONLY NORMAL HUMAN MESSAGES AND KNOWS WHAT USUALLY EXPECT FROM HUMAN
so can detect what's shii in ai message and upvote less shiity ai response
while constant ai users used to read ai slop so they mostly accept ai shii and can't vote objectively, twim
i know ppl who doesn't use ai (not even because they avoid)
i'm happy for them tbh
great ppl
noooo why banana
because we deserve it
Well hope at least y'all are going to have a good day or night depending on where you are
R u a euro
Yes
Ah
i need to do my ai and ml classes homeworks
like i skipped whole semester not doing them
Fair enough, I need to do my own work, granted that I'm almost done with them for the year I could probably finish it today if I wanted to be quick, but you know, beauty sleep
I mean sure I want to sleep but at the same time I want to do more stuff
I swear wanting to sleep feels inconsistent
when will ts be fixed
Well uhh maybe tomorrow?
Lol idk about that
Says in the message they are still investigating the root cause
R u sure ?
Please, do not tag everyone. Thank you
????
There are moments when the chat is more active, and other times it is quiet.
I don't know what you're really looking for, but LMArena isn't working right now, and the tech support team is currently working on fixing it.
i am just beating goons
Tf
The server has always been up and running; I don't know what this guy wants.
Team fortress mentioned??
Ugh and get @tepid grove away
If you haven't noticed, some people have a social life, others are busy, others are asleep, and others are on other servers.
What's wrong with him
Ignore