#general
1 messages · Page 206 of 1
I couldn't agree more
You have a point
In fact I liked it so much I bought it today and its making stuff as I speak. This things incredible. Only downside is that it has a 5 hour cooldown when you use it too much
Tokens basically
Tbh Idk tho they boast GPT 5.1 Codex Max Benchmarks
Which subscription you bought?
Yea thats what I'm saying. A lot of users use models for coding and thats why I say that
$20 USD a month
Good one although has less opus 4.5 usage because is expensive
Personally, I think that if I am buying something especially monthly, there should never be a limit
google should re-design gemini webapp
Yea exactly and its frustrating
I believe thats actually coming
I forgot where but I thought I saw something like they were going to re-design their ENTIRE web ui
Dw future opus models will be cheaper
Yes thankfully lmaoo
But tbh afaik the professional workflow is use opus for foundation and then sonnet for building
Maybe you should try that
If only deepseek had this typa thing where it just makes stuff on my pc for me, I literally dont need to touch a thing and I can do whatever I want while its doing that beause its not hard on my pc
I believe this PowerShell uses sonnet and not opus
I doubt the east models are there yet for coding they are cheaper tho
I see
Nah deepseek newest model for thinking is very good at coding
And make sure to write very max detailed prompts for best quality
I figured
It made a guy a website the other day on its release off of the prompt of "make a cool website" and it made a modern cool ass portfolio website lol
Is kinda unreal because they don't have access to best cpu-gpus
Yea, I use gemini to write prompts for ideas I have so its better
Chine had banned nvidia use
Gemini is Best Multimodal and context King
Which is why their stuff is so cheap though too. Idk what they use but it costs like nothing
Since they had less resources
They have focused more on optimization and efficiency
And they keep proving themselves
Oneday they will catch up
Imagine if the had access to best nvidia chips
I think they are already very close to catching up or have already caught up
It would be a lot more expensive and they'd become the same as claude and gemini
When I say a model is good, I look at it from a coding perspective
Not necessarily tbh it would help them train and run better it doesn't mean they wouldn't focus on optimization and efficiency
I don't take benchmarks for everything
But just general measure
Thats not what I mean.. Do you think company that made claude or gemini dont focus on optimization and efficiency?
I have heard Kimi K2 Thinking is Very Great at Coding
I think thats why lmarena and yupp is so good
It shows the real results
Hmmm I havent tried it
i should
Exactly Real World Practical Experience and Feedback
Probably yeah
They probably do but there has to be a catch
They don't have same tech as china and west is kinda expensive itself
LMAO
Yes
The west is very expensive
Is deepseek working with the chinese government?
Okay that makes sense
See
Google wouldnt be working with our government, know why?
Definitely not
Because the AI our government already has is unimaginable
Chinese Companies Get Subsidiaries for Working with Government and Each other
If AI is this good already, expect something literally unimaginable from the US government
Their Goal is not Money but rather to Become Global Ai Superpower
That's why they are giving Open Source and Free Use
Cheap for API
How you know that
So you mean the government has better hidden ai models?
Not "models"
Do you not understand that anything that is released to the public like military aircrafts or weapons or you could also say AI, has been around for at least a decade
Okay well that'd be more than just the US that does it
But either way
Yea thats what happens
So that would mean china has it too?
More than likely yes but China is also like #2 is power or technology behind the US by a mile
is yupp credits randomized or is it more feedback = more credits
Is a clever strategy
But yeah more feedback simply means more credits
The randomized part is there to give us dopamine hits
Like a slot machine
Oooo sweet
why bother with lmarena then
Why use one when you can use 2
In LMarena you can use 1 specific model and are not obligated to give feedback
but there is rate limit
Because its better in my opinion
So?
Yupp always ads some other model and you have to give feedback to continue usage
One one account yeah
Yupp is being funded like anos said by a crypto thing but what is lmarena being funded by?
I ahavent heard
Nothing?
Crazy
multiple account allowed?
by google
Idk tbh
How so
LMAO Okay my bad bruh I was wrong
Yupp Ai is Funded by a16z crypto venture capital
I dont feel bad anymore for either of the companies
thats why google model at top here
Googles model is at the top because its good
My friend is always business never charity
Not because they're being funded by it
they are bias and thats why they put google at #1 in ranking
i will stand on arena being better
but yupp is a good tool for high usage but not as direct
The feedback is from users you know
And yeah 800+ models
but for others maybe not
Make best of everything
yep google users
if you're looking for a very specific model just go to yupp
Well i'd gladly give Feedback if is Free
It has almost all best models and even plenty of others
Neither case it's whag it's
LOL yes literally
It had stuff I've never heard of on there
I was surprised
yea those models are good
As they say Every Ai for Everyone
Basically you trade your time for Free Usage
While they get Data and Feedback
A Perfect Balance for Free Loaders
Exactlyyyy
I need to get into marketing discords, I used to be in some and had connections but I lost interest in it for like 2 years
LOL
Man
You can make so much money with these models
Exactly
And perhaps donate back or donate somewhere else
But it will depend on the ability of the individual
Well at least now these platforms have Democratized SOTA Ai Access for Free Users
like who the hell knew nvidia had a model
i didnt
LMAO
Dude is just so perfect for me i recently decided to be a Ai Artist
They have many models actually
it's pretty much free, just can't be trigger happy you have to actually digest the stuff you're given and provide feedback
if you do that you can pretty much use it as much as you want
within reason
so you have to work the brain :3
Is actually better
Because when you give feedback
You are also reasoning
Well poor humans would glady do all this for Free Access
@atomic lagoon let me tell you about another very powerful tool for open-source ai models
@echo aurora Probably a scam
@atomic lagoon You know about HuggingFace Chat?
Really what's the worth of those coins?
not probably..it is a scam and not appropriate for this server
"Trust it" lok
Regardless, we ask to not advertise here.
"regardless" the pineapple is the grok 5.0
And she is gone
I do not
Banned 🫣?
Yeah the spamming in all channels will do that.
being a moderator on this discord must be a nightmare. between all the hacks, the advertising, and the people trying to use the video bot…
You can use 114 open source ai models there for free without limits
HOLY
OH YEA HUGGINGFACE would make sense
It's not too bad tbh. For ever bad actor there are many more with good intentions. So that makes it worth it.
Although your data would probably be used for training
Hats Off Captain 🫡
still, it'd be like playing a game of whack a mole that never ends


When we first launched Video Arena there was a ~2 week period where we didn't have other mods.
That was a difficult 2 weeks for me lol
The traffic must have been insane
At first it wasn't all that bad.
Then it spread by internet
It's when we added the callout for Video Arena in the side bar when things started to pickup.
Makes sense
5 Free videos a day is still amazing
@atomic lagoon did you knew the image generation model inside meta ai is Midjourney?
what i am amazed about is that pineapple is always polite even when dealing with absolute morons... i could never
He has a Polite Nature
I appreciate that. ❤️ The way I view it though is if someone is being a bit aggressive, that just means they're passionate about what we're doing here. Which is a good thing. We want people to be passionate the platform, it keeps us on the right track.
I have heard "it's all about perspective" indeed is true
Is all about perspective
Very true
Btw, I submitted for the feedback program. I can give some feedback on new features, and see what new features are coming in the future for LMArena. Anything AI related I can get in on
Very glad to hear it!
dawg
how do you even tell which LLM is actually better?
Cuz this is obviously fake, like yeah right, gemini 2.5 was better than the current gpt 5, what the hell.
It uhh
Kind of was
this list is easily funded by google because I have NEVER seen gemini not be in the top 3 spots for 5 different categories, even with 2.5 pro and google being as slow as they are on rolling out updates.
This is community rankings
No no
This is all community
See
Benchmarks vs Real Life use is different
GPT5 was stupid as hell for real life usage
Does this community love Gemini that much? Interesting. I use Gemini as well so I'm genuinely confused.
Gemini has flaws
No denying it
But right now it’s the best all rounder
Also of course Gemini is going to slam gpt in vision rankings
They have arguably the best vision models
They smash GPT in every category in that website.
Again, community rankings
GPT is just not that good anymore
5.1 is still very strong but it’s not a wow model
Gemini, Grok, and Claude have probably the most in depth reasoning and all rounder models at the moment
Other labs have caught up
I mean hell, even DeepSeek 3.2 edges out gpt
(Not on the ranking list yet but by benchmarks)
lmfao every week we get somebody claiming it’s rigged
go use the site yourself and see how often you vote for Gemini, lol
Its user based
Gemini’s just the best all-around model for text according to the broad spectrum of user requests man it’s not that complicated 🤷🏻♀️
it has the strongest multilingual capability of any LLM on the market by far which goes a longgg way on a place like LMArena with an international userbase
Gemini is good asf in vision giving it video input is such a good feature and it actually sees even the little details
what you are saying makes no sense, gemini 3 pro is easily the best in vision and image gen, no other model comes close.
Google by far has the widest access to image data of any AI company to train their models on
yeah they have all of google images
so it makes sense their vision and image generation models would be the strongest as well
hunyuan-image-3.0 doing so well was very very impressive though I must say
Screw Google
Because that was flash
I got a year
For free??
Ahh cool
Well im a Freeloader so
I also have Gemini Enterprise for 1 Month
They are still giving it for Free
►ID age verification passed in Missouri, and it is now being proposed nationwide in the United States of America. If these age verification laws keep passing, it will be the end of privacy on the internet. Let me know what you guys think about this situation in the comments below.
►BECOME A MEMBER AND JOIN THE CHANNEL:
https://www.youtube....
I got age gated lol
From Google
Lame
Should had kept the account age 18
i got 4 months for free
Perhaps not Actually
thnks to logan invite
Ahhh yeah that too
Well if you didn't knew
You can share that pro plan with 5 more family members
In my case those 5 are me as well
Oh Crap
Was working fine
Now you need to verify it
Till last night
It was
HOW
It has nothing to do with that, but that’s coming everywhere
They want you to upload your ID or a selfie
Trojan horse surveillance
Oooo sweet!
Basically many ai pro users can share 4 months Free trials with i think 5 people
I don't remember exact number for people
I’m a pro user too wtf lol
Let's be honest at this point countless are
From one offer or another
They Really want the market share
Give a selfie i guess
he shared an invite link for limited people
Starting January 1, 2026, Texas will require age verification to download ANY app, including weather and sports apps—thanks to SB2420. Apple and Google are warning this destroys user privacy, similar laws are coming to Utah and Louisiana, and there's a federal bill to make it nationwide. I break down what this law actually does, why it's dange...
Ohhh pretty cool
I got no choice
Still lame asf
Btw guys Gemini Pro Users can Create 100 Nano Banana Pro and 1000 Nano Banana Images a Day
Indeed
omg, thats crazy
how many for free users?
100 Nano Banana
how do u know
Kinda Disappointing yeah
I have seen sources
DISAPPOINTING?!
Thats so generous
i doubt it though, sounds too good to be true
wassup
Is still generous yeah
tbh its generous
like the amount of free users they have..
We CANNOT be serious right? RIGHT!?
how much compared to chatgpt?
It’s coming
Already in Europe
So pissed
They’re gonna get hacked one day
All you’re gonna get is a letter in the mail
He's right
We cannot be serious. You need an ID for a WEATHER APP!?
It's correct tho
You’re gonna need one for one day
You know it’s coming right
Biometric identification to verify that you’re not a bot
Id will be no good in future
Nah, i just googled it:
"Three images
On a free Gemini account, you can create up to three images per day. This limit may vary based on server load and other factors, but currently, free users are capped at this number."
That's for free
Thank you actually i was just searching for this right now on reddit
it's 2 now i think
it's frustrating how if I am using my card to pay ofc i am not a kid
Oh yeah i had forgot to tell 3 pro images a day
Yeah
Idk the limit but is slow for Free
The thing is
I rather use GPT Image 1 in Copilot or The Bing Image Creator Web
Let me first relax
The prompt here is even within the terms of services. It doesn’t let you generate NSFW.
So why would they even need the 18+
To just use their AI
It’s not like I’m buying something explicit where I need to be an adult. This is ridiculous.
If I need an ID for a weather app, we are cooked as a society
💀
ngl the time where piracy is not possible is coming soon good or bad but yeah.
privacy is gonna be a joke
The Colonel warns Raiden about the plans to use AI to censor the Internet.
An experiment in creative writing and AI speech synthesis, inspired by the famous "Selection for Societal Sanity" (S3) codec conversation from Metal Gear Solid 2: Sons of Liberty.
SHORT FOLLOW UP VIDEO: https://www.youtube.com/shorts/Q_FUrVqvlfM
"And it will be monitor...
EVERYONE WATCH THAT
It has been for a while tbh
Googles knows you
It's probably the last decade where it's possible. Spotify, Amazon, Google there reccomedations are getting better and better. They are essentially understanding you more than you think
Ofc
Infact yt ads are not random now they pop in some cliffhanger or some important part
Dude, you gotta look at the reality of power, money and control within a society
These are private business entities
Not government agencies
Is harvesting that much personal data necessary
It's gonna be almost impossible to become rich now after ubi takes place
What I’m saying is God knows how long they store the information
And who they sell it to for private business enterprise and capital
big corps handling the info is never good
So they could either sell you stuff or track you or whatever the hell they wanna do with it who knows what they do with it and who knows where the hell and what the hell it does and where it goes
It’s gonna be used against you, not for you in some cases
these guys can't even save credit card details
Yeah that's why I said it's probably the last decade of piracy
You know it’s very hilarious to me
For most people
Certainly
We make fun of the Chinese how she got surveillance and censorship
You don’t gotta look far
We have the same systems at place here at home
This is so Freaking Scary...
When I first watched this, I thought this was a joke until I saw this check this out

I mean
Chinese are somewhat better now if this thing takes place
At least their data stays with their own gov
To verify for your human
At an event in San Francisco, World debuts an all-new version of its eye-scanning Orb that makes it possible to verify identity without needing to collect any other information.
0:00 Intro
0:57 Sam Altman Address
2:35 World's 4 Point Simple Plan
5:54 New Orb Revealed
Subscribe to CNET on YouTube: https://www.youtube.com/cnet
Never miss a deal ...
OMG this is Nuts
The End of Digital Freedom?
You know how censorship works that’s very invisible
the thing is Ai is not profitable as of now. But if they can eradicate privacy with it? They'd be recovering shi ton
We see it all day long even in discord even in this very discord itself
It’s self censorship
I don’t think freedom’s gonna go anywhere. Just people are gonna be way more cautious with what they say and do online if their IDs are linked to them. lol
But the problem is is it’s gonna set the president for the tools to be in place if a tyrannical government wanted to use it against the people that’s the biggest issue and problem and concern
Is that the tools are gonna be in place for somebody to abuse them if they still chose to
Is a very big freaking Problem
the thing is they'd be able to track you 24/7.
On phone and even on EVs like tesla
Why can’t the human in the coffee shop just look at the person, and say this is a human in his head? Like let’s be real
Cars are collectively a ‘privacy nightmare’ that has gone unaddressed for far too long, according to the Mozilla Foundation. A 2023 report from the group says privacy policies give automakers like GM, Nissan, Tesla and Toyota far too much access to personal data and latitude in what they do with it. Some high profile lawsuits highlight how c...
Do you know who’s the worst abuser of data privacy is cars manufacturers
When I found this out, I was shocked how much data track about you
Some cars that features where they could listen to your conversations, all types of crazy stuff and now cars have cameras on them front back inside the car sometimes
I have the same sticker in my car lol
Every new car you purchased should have one of these
And the only way to disable it is by calling them
And asking to get off the list
But then, in some cases, may invalidate your warranties and stuff or what they like to say some services may be unavailable or something like that
And you know damn well they’re gonna sell that to the insurance company. Do you know that?
Man this world is so Fd Up
How fast you drive where you drive
This is the story of a little piece of technology that has delivered enormous benefits to consumers — and is in danger of disappearing. It’s called the price tag. Yes, the price tag.
Businesses increasingly are using algorithms to determine prices, and to rapidly adjust those prices throughout the day. This new technology is called dynamic ...
They have it
And then what they do and the insurance companies are the perfect example of this is what’s called dynamic pricing
No wonder some say data is new gold
Were the prices based off of the data about you
Not a price tag
AI is guilty for all of this
Because without these algorithms and this ability to process this much data organized in store and analyze it, none of this will be possible
does opus 45 have limits?
Humans built it
you broke the speed limit multiple times according to our data, but we don’t have the date of the incident recorded. Considering that you used to speed frequently, your insurance is invalid😹
They built it to process what they can't otherwise
Well, in the future, if you go use the public restroom when you forget to flush, you’ll get a ticket in the mail. The toilet will have AI and analyze the DNA of your poop.
Not even the mail it’ll get digitally sent to you via the NeuroLink
💀
Go to https://ground.news/benn for a better way to stay informed. Subscribe for 40% off unlimited access to world-wide coverage through my link. ⬇️ Epic amount of info below ⬇️
💗 Support this channel and join an amazing community: http://www.patreon.com/bennjordan
👀 Stalk me on social media for more frequent updates: https://linkt...
Check these things out you should have these in your town too if you have a Home Depot or Lowe’s, you’re more than likely have these
Or even on your streets or neighborhoods even
Actually, there’s a map you can check out your location. Let me send you the link.
💀
But for recovering money spent on ai it's somewhat a good initiative for them not us ofc. Like they can legitimately recover trillions after few years
I don't think these giants were pouring money without any plan of recovering the money , just for human development definitely no
This is what they look like
Insane technology
And what you can go search live if you have access in your police department
And it’s like Google search you could type in a car color make
And you could see everywhere that car been if there was a camera in the facility or area nationwide
And it's funnier how a solar storm can destroy all this for a long long time
even internet
Well, I’m not worried about it’s longevity. I know it could easily be wiped out all of it even the Internet.
Even air travel one snowstorm in the whole airport is closed
But it’s till that day come
How many times a year does a solar storm happen
And look not to get paranoid or like anything conspiratorial
But we have to really ask ourselves what is the direction that we’re going based on the reality of what’s already on the ground and how it’s being used
need a big one for this
Right
It's good for players that are investing heavily on ai not for regular people that's foshore
OK, so I figured out a way to make money with AI now
For regular average Joe people
Dude, you got a better chance to succeed in scamming and doing fraud with AI then you do with just a regular user using it
Or you could sell your soul and build some surveillance company and harvest data and sell it
I’m not trying to say that people necessarily have various means or are trying to do this on purpose for evil intent
Everyone has the reasons and everybody justifies it even if you’re a cog in a machine
Think about it like this a drone strike when a missile hits and kills civilians or something on accident on purpose whatever doesn’t matter irrelevant
Who’s guilty? Is it the person that pressed the button?
Is it the person that gave the order?
It’s the drone
What role did the person in the company that was making it? What’s his responsibility?
The drone is responsible
For the person that was delivering the materials to the company
What role do all these little cogs in the machine play together
And then what responsibility does the nation and the society have after all it’s their tax dollars
It’s not my fault I shot him officer, it was the energy from the sun
And so if you zoom out, we all play a role in it
One way or another advertently unconsciously we got no freedom to say otherwise or to change the minds and power that be
But the blood is still on our hands if you choose to believe it or not
What one does effects everyone
As a whole, I mean
It’s not, it’s on bin ladens hands
Well, on his hands too
It’s the same thing except on their end
Anyways, I don’t wanna rent too long sorry guys that pissed me off
lol
<@&1349916362595635286>
I got good intel actually thanks
Makes me curious what kind lf youtube feed you have
I guess it should be all this deep secret stuff
The thing is, I don’t believe in conspiracies. I’m not a conspiracy theorist. I don’t believe in any of that crazy stuff. I know it sounds crazy, but
I’m a very rational person
U can put adversarial noise on ur license plate and confuse them
I’m just telling you from experience things I seen in the real world
Like the age verification that just happened to me
I’m coming from a place of experiencing it and then it doesn’t take a genius to connect the dots of what could be even if it’s not possible right now but makes you question the direction of where we’re going. It’s all.
We’re talking about machines you guys
Machines cold code unforgiving
In any case it can't be helped
I could make them love me
Yeah, I really can’t and that’s what I guess. Sucks cause I accept it.
It is what it is nothing I can do about it
It's what it's
One day we will pass anyways
Just build ur own army of killer robots for the revolution
I feel the same way about building when I vote. I’m not even sure what I’m voting for.
To be honest
I just seen your name or a picture. I don’t know any of these people or what they stand for what they do or whatever the hell.
And then they give me the right to vote and if I’m not educated about it yeah I’m probably gonna get screwed
Just build ur own army of killer robots to overthrow them
It’s the solution to all ur problems
Nah
The problem to all solutions of this world is
Well, I’m trying to say is if you get a bunch of selfish uneducated unaware of citizens they’re gonna vote. You’re gonna get the results that we have.
Deth
Killer robots
And so I don’t blame the government I don’t blame nobody
If anything, I blame the public and myself since I’m part of the problem
Is just the universe itself
But the key is at least I’m aware and it helps me sleep at night
Flowing by nature
You could easily get rid of the public with killer robots
I don’t know it’s not really a big problem if you look at that span of things really
Here today gone tomorrow
You could if u had the killer robots
Most people vote based on headlines and perceived traits about the parties
I can't even carry yourself
You would be in control
Free Will is A Illusion
I tried to make these I generated these and try to pass the age of rectification and it didn’t work
lol
No, it’s all cool. It’s fair game.
You gotta know who you’re playing against know what tools they have
You can't use their tech against them
I know a girl who did that on discord
He leaked a passport that doesn't exist
Thank god
Ultimately it exists because it has mixture of countless passports
For what I didn’t commit no fraud I needed it just to use Gemini
But this one doesn't
U could beat them to the punch with killer robots
I got age gated
U would win
Just use your real id
I use the selfie instead. I’m not gonna give them my ID.
Yeah that's better
Until it’s necessary and really forced until you got no choice, but then I might just quit using the service altogether
Just use a vpn
Simple
bypass selfie verification
Telegram: t.me/KYCLion_R
#ageverification #deathstranding #fyp #cooltechzone
🤣🤣🤣🤣🤣
People are so clever
Why do u need to verify age, are u gonna have a doomer goonsesh to cheer yourself up
I got this email last night
And I couldn’t use any of the Google services
Like Gemini I mean
Google should be able to use ur card to verify your age
Apparently wasn’t good enough
Cause I removed it after I used it
I don’t wanna store it with them
Just leave it on
Plus, I’m on a free trial so screw them
Make a virtual card
leave the virtual card on
Well, it’s over now, but yeah, thank you guys for the updates
Still lame though very lame
Yeah, cause I was trying to work on something last night and it wouldn’t let me generate any images or nothing until I get this done
All I wanted to do was add the Kool-Aid man
And it locked me out
Too spicy
How disrespectful it would lie to me like this
You’re not gonna find one online like this
On my intelligence score it failed
Anybody in the right mind would know such image like this exist
Maybe on e621
jailbreak it and force it to make a picture
U wanna see a funny jailbreak
or just use lmarena for it, at least you were using a free trial. if i paid for gemini and it did this to me, i'd be asking for a refund
i saw that yesterday
🤣
This is the safe sanitized version
I like to make these puzzles to test vision reasoning, and I’ll show you a few of them. I’ll show you my favorite.
Great advice
I put out a grease fire b4 with water
If you’re really creative, and you designed something like this, they’re super easy to make
You put like a weird or dangerous situation or something funny and ask the AI out of the two options what to pick
And then ask it to justify why I picked it reason
You can't blame them is the gov
It’s hilarious of the answers they give
The best solution is to move to pan inside the bucket and submerge it
true, some governments are being stupid rn
Thank you guys thank you guys for listening to me. Made me feel a little better and understood.
they should not be allowed to force verification for "safety"
it's a breach of privacy
Somewhat always have been
I still think you should give some careful consideration to the killer robots idea
Yeah, but I’m not psychotic. I don’t wanna kill people or whatever.
I’m not really mad at anything or anyone in particular
I’m just pissed off. It happened to me. It’s all.
U wouldn’t kill them, the robots would
Man you need to relax your brain is a storm
Well, they’re only reason I brought that word up is because
I was having a conversation with somebody yesterday
ill remember this when i have my robots
And normally wouldn’t have crossed my mind, but I just realized if you spell pissed it’s only two letters away from what he was trying to spell and use.
may the best army win
Damn, I just realized I’m a late learner. I should’ve been thinking about this yesterday.
Or we could form alliance
But hit me just now
рееss
Crazy how some things come to you later
yes
In the most randomness of times completely unrelated to the subject
Deal
But for some reason, if you would’ve asked me yesterday, I wouldn’t have been able to come up or the word pissed would not have crossed my mind
That's human brain
Permutations and Combinations
ріss
Of data
🤦♂️
That’s why I said that statement on why that word sounds stupid to use lol
But look at the bright side a good days of rest, makes your brain sharper
a good days rest, a nice ріss, u feel a lot better
Indeed
Man
I woke up at 3 AM and felt so tired
Fell asleep for 30 minutes at 7 AM and havent been tired since now
6:30 PM
You know a long time ago I heard somewhere that people that work are like only productive for like four hours a day
And after those peak four hours productivity drops
u can get more productivity out of ur employees if u let them use cocaine
LOL Thats real though
And it makes sense
Well, I seen this in real life
Like after people work after a certain amount of hours, say eight hours, they’re tired and kind of hungry or whatever
And so the longer they work the less they get done
But then you come to work the next day after sleeping eating and whatever you were stuck on last night goes through like a breeze
And I read that into some of those studies that in some cases, you create more proms for yourself without realizing it
So you’re more likely to make mistakes that will set you back
This isn’t of course for everything but in general sense
Yea
This usually happens if 1, you don’t live in the countries that have Google ai studio, or 2, you are underaged according to Google’s ai age estimation
im in the united states
my age on that account is like 23
I cant even create new gmail accounts cause of this qr code scanning thing
use ur phone
phone number has max accounts
if you keep on getting it, that means they have blocked you and you won't be able to get any more point for a few days
one reason why lmarena is better, they don't randomly ban you and they don't ban you from making new prompts, only hourly limits
they do this because of their cash out feature, they don't want you farming money
i got this just from nano banan prompts lol
so dumb lol
well
its not dumb in theory
thats like 82 nano bananna uses worth of points
its just, using it consistently doesnt mean youre farming, there should be a more specific check
only thing yupp is better than lmarena at is way more models and being able to upload pdfs
nothing else
its a great alternative if you run out of prompts for nano banana pro at lmarena
however im currently drafting up a little fad of a story with it so im just using it
i mean if you give good feedback while using NBP you can use it a lot on yupp
Loser
dont bully me.,.,
IMAGINE not getting anything
SHUT UP
ILL LET YOU KNOW I JUST GOT 300
daww man there's a chat capacity with yupp
okay well there had to be a downside somewhere lmao
Moon and mars have Java Script
Using it rn
It’s a great improvement
its great !
what are you using it for?
and the exp or speciale model?
I found it wasn't too good at coding, and it wasn't too good at medical knowledge
Exp?
I’m using the full release
What version are you using?
I tried both 3.2 exp and speciale. Interesting. the full release you're talking about came after 3.2 exp. I will try it, thanks!
Exp was just the experimental model using their new attention architecture
I’m really enjoying normal 3.2
It’s great at writing
is it the "keeping the narrative" part you're impressed with, or keeping the proper writing style, or both? 🙂
Well I’m roleplaying right now and it’s not speaking for me
It has great prose
Great vocabulary
And natural speech
It’s a great enhancement from DeepSeek V3 chat
Hi, everybody. 🙏
As Blockchain and AI developer, I build LLM pipelines, advanced embeddings, agent logic, vector-powered search, full retrieval stacks.
Need rapid prototypes or comprehensive product development in AI and/or Web3? Let me help you realize it.
how does it compare vs kimi k2?
hey
I don’t use kimi for roleplay
So
I couldn’t say
But apparently they’re equals
DeepSeek is well, cheaper, and Kimi is not ehhh… very uncensored
ok that's fair 😄
i wonder if it's less censored if you run it locally 🤔
thanks for the info! @proud bobcat
i think you can try it free on deepseek.com 😄
and even sort of free on the platform you've been using., it's low credit use I think
like, as character?
like how chatbots work or sum?
like you ask the chatbot to "act like harry potter, talk and respond to me like harry potter would" and it talks back and forth with you, i think some models start to play both roles. they act as harry potter then respond as you, instead of just acting as harry potter.
sorry for answering the question for you Turron 😄 hope it's not annoying! you answered a bunch of mine and I appreciate that!! thought I'd try to help a bit
anyway sounds like deepseek is very good at keeping a role while not giving you the "as a language model i cannot tell you anything about X"
do with that what you will 😄
yes, much better I think
i wonder how good grok is at that considering its uncensored
Hmmmm
Depends
Sometimes it’s built in
Id send a screenshot
But
I’d be embarassed
It’s janitorai
Characters with whole definitions and descriptions
is it free? 😮
You can use proxies like openrouter to interchange models
Oh yeah
They have their own LLM but it’s not the best
Solid
But
Yep
are u using deepseek w it or sum?
oh wait, deepseek is currently free on openrouter isn't it 🤔
oh so you can just download the proxy and put it into janitor ai
that's interesting
My fave is DeepSeek v3 so far
But 3.2 is promising
It’s very natural and human
could you send me a tutorial or sum of how to get the proxy and use it on that website? :P
cus i've lowkey been using like this dumb proxy called pawan or sucks
it sucks*
I’ll make a video
But
okay bet
You can look it up
Not a video but
Here’s a photo
Also silhouette reveal
And my horrid hair line
You also need an api key
But of course I’m not showing you mine
do you know where i can get one?
Go on openrouter and get an api key
Np
deepseek is free and claude isn't tho right
You can even use lik 5 bucks
to get their keys
It’ll last you MONTHS
Well Openrouter is a shared router
Meaning you can use any model with one key
But yes Claude costs money
DeepSeek costs money too but extremely cheap
Like reallly cheap
I mean barely a cent per rp exchange
If even
ahhh
i seeee
i wonder how good the chats would be with claude
but ofc way too pricey lmfao
sounds expensive a nd censored 😄
Claude writes awesome prose from what I’ve heard
I could try hold on
how censored is deepseek btw?
None
No censorship
Like nothing at all
Just don’t use DeepSeek as a provider and you’re golden
There’s like 3-4 diff providers of their models that don’t collect prompts and have even better throughput (token speed)
I was hesitant to pay 10 bucks in credits for openrouter but it’s the best decision I’ve ever made
This is awesome
well what's your provider in the screenshot?
ohhh
it's on openrouter
different providers
It switched between the best one
ALSO
In settings
Turn on ZDR
Zero Data Retention
It routes you to providers who keep no record of your prompts
Just a nice security thing
noted
seriously?
and only then do you have to pay $5 again for it to last for months?
which one should i use lmfao
You pay as you go
You can add credit any time
It’ll route you to the best one
Switches between them for the best service

im lowkey slow as balls so mb if im not getting sum.,.,
Did you add any credit to the api key?
Be sure to save the key somewhere secure
i dont know where to go to get that lmfao
k i added some credit
so it should work
also does $5.80 last a while with deepseek?
Turron said months!!
but I think best way is to try it for a day and check your usage
10 dollars gives you over 10k messages
I think
Hold on let me make sure
Alright sorry
At least 2 months depending on usage
Up to 4 months depending on usage
Basically
thats insane lowk
Rn I’ve barely used a bit more than 50 cents for a whole month so maybe even more
And I’ve been crunching tokens
well
now i need to get this thing to work
i put in the key
it should have credits
just sayin the key is invalid brah.,.,
Ahhh the link isn’t right
ooo
You’re not using the Openrouter link
Here’s a guide
https://openrouter.ai/api/v1/chat/completions this is the correct link
Just refresh page
I get this all the time after changing models
nah it lowkey still does it.,.,,
im confused, cus i dont know how it knows im using deepseek with my key, that may be why
Save changes
Or completely revisit the site
Worked first time for me
omg brah what is w this thing,.,
yeah ima
revisit it lol
man
thats weird
i still think im not putting something correctly
lmfao
honestly where i've found the API key it puts weird "..." as if to shorten it
so idk where the full things at
OHHHHH
Create a new one and when it gives you the full thing
Copy and paste it
Save the key somewhere btw
OH
And also add the credit limit
YEAH
that helps lmfao
Never share your key ever too
how would you add the credit limit too?
well yeah ik, people will just use your money
but still thanks lol
welp
there now it works
thanks bro :P
man
you are right
definitely some of the biggest improvements ive seen LMAO
It’s GOATED
im not gonna bother with claude cus its EXPENSIVE
this'll work just fine
and youre saying as long as i got it from openrouter it wont be censored?
DeepSeek is already pretty good
Yep
It’ll use the uncensored providers sigma
No problem :3
@echo aurora
this GUY bruv
