#💬│general
1 messages · Page 578 of 1
voice uses the old 4o model not 5.5
doesn't matter bruh
it heavily matters
its like comparing a bicycle to a mercedes
really?
yeah lmao
even 5.3 is much better than 4o
which you have if you're using the free version
5.5 is paid tier only for now
but still!
how can it be so confident in it's hallucination ... that it doesn't even understand user's sarcasm/irony
even the user is giving correct info ... but it didn't catch it 😕
i have gpt Go version
yeah you should have a limited version then
check the model selector
its an old model lol
there is no option to select model on mobile!
I might be stupid but what is an API Key?
Is it useful?
Is there a way to use perplexity in visual studio? The extension has 1 star
Hello
and on PC it's just auto
or
gpt 5.4 (flagship model), as they keep flashing it everytime ... in every chat
@aer0099 muted
Reason: 9. No spamming.
Duration: 4 minutes and 50 seconds
hello, how limited is deep research for pro users now? few months ago I had like 500 uses per month, now it's down to like 1 per day?
yo guys i'm new here and i have a quick question
wait pro reaserch is limited???
even very limited for me atm
i didn't know that
i rgought it's unlimited
but my question is the perplexity discord connectors the bot one and the regular one what they do
It's lower now - #💬│general message
This morning, Perplexity PRO on Mac OS wiped out all the Spaces I had created and had been running for several months. It's very frustrating and disappointing. Has this happened to any of you? How can I restore those Spaces? Thanks!
no shot
is perplexity that bad
i just started using it and i see so much negativity from the userbase
i probably should switch to claude or something geez
how low? are there any numbers?
click on the link on his massige
used to be better, but like with every service today: make it good enough so everybody uses it, then shitify it
why
capitalism is not for the people brother
like i made my own ai and i was like api costs are too high and switched to perplexity cuz it essentually does the same thing but for cheeper
i added perplexity sonar via an api
should i return back
but computer is really cool dou
yeah why?
I can't even choose labs anymore. is it scratched completely for pro users?
the model sucks
perplexity has been going downhill
so fast it is mindblowing how bad they are becoming
meh with the right skills it's pretty good
@short narwhal
oh really? explain
what does that even mean
i created a python script
u mean with perplexity computer?
Labs ability to create files/apps has been incorporated into Computer and into Deep Research
that essentually chains groq apis perplexity sonnar and claude 4,6 sonnet together
ok you do realize it is much more expensive
you burn through all credits
yeah that's why i switched to perplexity
???
perplexity is expensive
thx @heavy portal for helping out. So Labs/Deep research combined is 20 uses a month now, right?
explain i don't get it it does it job perfectly fine
Correct, even though when you look up your own account at https://www.perplexity.ai/rest/rate-limit/all it still lists Labs separately. Not sure why they didn't increase DR limits when they removed Labs.
how many credits do you have?
computer or api
computer 0
i'm a pro user too
but you realize that perplexity on the sidebar can do the exact same thing for no creddits
you are not very smart
???
im not even gonna bother explaining
if you say so
just for you to know i'm a student so from 22 usd it does fall down to 9 so i concider it win
computer is totally different from regular
and i got 2000 computer creddits bonus witch i burned
i know but they do almost the same thing at theyre core
absolutely not
Can I look for ppl to cruise with or different channel. I’m on fh5 driving lambo sesto going fast
the models from perplexity have tiny context windows
than claude you wish
i said chatgpt
It feels illegal asking this here but after getting screwed over might as well: what's the best alternative then?
chatgpt
claude
it researches very well
claude has terrible limits
chatgpt gives u 3000x extended thinking gpt 5.5 prompts per week
stop misinforming
have you tried before yapping for no reason
For deep research and multi step reasoning and searching/crawl capacity - and working with big files. For academic research
chatgpt
do not listen to this other kid i promise you
i have used everything
yah just be a sheep and use chat cuz every one does
with the current gpt 5.5 and their limits it is much better than claude for the price
I'm listening to both of you, Im just trying to get as much info as possible
I do know many people switch over to claude lately because of document processing
you do realize claude had more subscribers than chatgpt
so you are wrong again for the millionth time
it can do everything good but not perfect
and do you know why is that
just stop it
but of course you don't
As far as I understood chatgpt is better for all-round use and claude for programming and data handling no?
with gpt 5.5 they have closed the gap
exactly
gpt 5.5 came out on thursday
Not true btw
honestly @bright vigil just try different ai's and see what fits
Like they'll all get better guys I'm just asking if there's a solid alternative for my use case
ik
Perplexity would've been perfect
chatgpt believe me try that one first. i had claude plan last month and cancelled it im much more happy with chatgpt
the limits for chatgpt are practically unlimited
what do you use ai for the most @bright vigil
not for their codex
being
cooko
@bright vigil check dms
Academic research top priority
than belive it or not the best decision you can do is stick with perplexity if you wish go on dm's and i'll explain myself
that's just my advice
i gave you a real example instead of just talking bullshit
check your dms
why do you glaze chat this much btw
as opposed to what?
just in general you glaze it like i used to glaze gemini back in the day
even my context windows
so what's the better alternative?
locked at 256k
and the slash command just
disappeared
like it never existed
I have no idea if that happens to u
on their newest update
But it happened for me
and the reason why i haven't switch is because claude share the same quota pool for both chat and claude code
that's what keeping me from using the max plan
this is not working for you?
I was able to use it for free on the app/website.
nope
The expert mode is v4 pro.
I can't even turn on fast mode anymore
before u say anything
yes i did clean everything
fresh install
change the config
even
deleted everything
oh u are on windows
and reinstall it
ye
here we go again
please contact support@perplexity
i see people have a lot more problems on windows
sadly ai companies target mac more
i highly doubt about it
don't get me wrong
yeah cuz the guy that mskes the hardwere and softwere is the same
and it's easyer to optimize
they always create the mac apps first
Even the config toml file?
I don't think that's a priority
they always create the mac apps first
gemini also has mac only
imagine targetting into a Os where it's limited to a certain use of users
hello
because windows basically doesn't need one ?
When they can just use the PWA
for it
u can do that on mac too
im not glazing mac im stating the facts
claude cowork came out january for mac and it came out april for windows
I think the main reason is environment.
Mac and Linux by default have their own environment so the security side of the app is easily integrated but for windows it doesn't exist. Even codex had to make their own for it which is a pretty hard thing to do
Blame microslop
mac is unix like linux too
ye that's the thing
I said it above
just because it's easier
doesn't mean it proved companies favor Mac more
that's totally wrong
?
prioritize then?
they prioritize mac there u go
Dude read
ask your perplexity do they prioritize mac
If u are saying about priotizing then Windows should be priotized more since they are way harder to develop and they need to focus on it more
Ur point is wrong btw
Bro clam down ❤️
Why is everyone so uptight
look at the release dates...
answer this question
Should u focus on a easy platform, smaller holders
Or a platform a bit difficult but 70% of the people worldwide are using it ?
The release date just prove it's easier to develop
it's not about what they should do it's about what they did
Not
Did u even read my message ?
oh brother
and u are taking a company for a whole world inc
Dude do u even know how to cite
or claim a fact check ?
look at the release dates
Assumption mean nothing
based
read my message
exactly don't assume just look at the release dates
Dude can u read the messages, release date doesn't prove anything
should they prioritize windows? yes they have more users. Do they prioritize windows? no they don't
Give the source where u got ur fact
Did they announce it ?
If not then u should stop spreading fake new
i've told u many times already to look at the release dates
and blindly trusting to some hallucinated replies
Let me get this clear
Easier to develop doesn't mean they favor it more
Release date faster doesn't mean they focus on that platform
yes it does?
There isn't any claim to that for u being developed faster so u are a better platform to be priotized
No it doesn't ?
Do u even work ?
Because if u did
U wouldn't make a mistake
my dad is already rich and he has a job why would i work?
misunderstanding 2 concepts
No offense but that's ur dad money not ur btw
Ur value isn't enough for those
he buys me lots of things and takes me around the world
🤓
Is there a way to connect perpixlty to mcp
Thanks Dude i was in the middle of work and i opened my phone 😭
Lmao
Gl with ur work
i might aswell
Head to sleep
Only MCPs recognized by perplexity which are in the add connection tab
by mcp's u mean the connections?
Those are mcp
Does perpixlty support roblox mcp
nope
U can still do it
but
might took some times
How
Metoclopramide?
are u
As in asking the thread to install it in its env?
trying to cheat in roblox ?
Do u remember my extension
perplexicode ?
But it may not have the command permission to do it IDK
Bruh no how is that even possible to cheat i meant studio
Ye u can connect mcp server to vscode
You have never told ke about that
while the extension is working fine
Hmm
If its in perplexity thread bound then you could save the thread to a space
Is this free for pro users
yes i made it free
haven't updated it since
u can update it urself
So i can connect it to my vs then using rojo to roblox studio
On a scale of 1-10 how bannable is this also perplexity already has integrated github mcp into their connectors
0
my account is still alive
and kesku knew about it
so
i think
it should be good
Man
He is not getting paid enough
😭
You should sleep btw
yes he is
Idk if you slept the morning
i saw him driving
in his new car
the other day
On the highway
my way
or the highway
Have you moved out or something?
Soon
might
visit
Berd
and kesku soon
Gl on that
but not now
People, lets boost this post if you want perplexity to bring back faster AI LLM. Claude Haiku, Google 3 Flash, GPT 5.4 Mini, etc. https://discord.com/channels/1047197230748151888/1498295786763718768
This uses cookies bro is it safe
Imagine if he turns out to be a furry after making fun of matto all the time
hehehe
the source code is there
u can check it
i don't have any stealers
or some spy thingy on it
@dry kestrel fact check
You can give it to an agent and ask it for security check
Nah i trust im just gonna use it
hehe he that wad the plan from the beginning
Just hope its responsive and doesnt crash every 2 seconds
No api key is crazy tho this might not be legal 😂
since i've tested it myself ( yes ofc i made that extension so.. )
Damn u made it did u make any tuts for setup
@twin grail the most berd thing that a berd would do
the setup is simple
everything is in their
there
typical Berd activities
i dont know why but i feel that sonar 2 is better than nemotron in every way, impressive
For research prob yea
cool
Am I imagining things? The selected model keeps changing back to "Model" every time I send a query
and it doesn't tell me "Prepared by Claude". Instead it seems I'm talking to ChatGPT.
that didn't happen yesterday. I can't work with ChatGPT, it's cognitively challenged.
...
I'm having ChatGPT forced on me and that's unacceptable. No Model choice sticks, it's always ChatGPT.
ChatGPT is the worst ")"§$"§$ out there, constantly insulting people without them even realizing.
Example? It literally spams me with lists and bullet points (which are not set as default),
and when I point that out, I get "you are not imagining things", which, of course I don't,
and it mentioning it is insulting and disrespectfull, regardless of people actually noticing!
And that's not just happening when I get spammed and complain, it sneakily keeps insulting people all the time.
mine is totally different
have you touched your personalization settings at all? you can have it stop using lists etc
I just want to set a model so i can continue without being forced to having to deal with chatgpt.
- I've mentioned that.
- I did.
- The others don't behave the same.
on ppxl you cant
There's settings for that.
hold on
Also, I have not noticed any difference between chatgpt with perplexity as middle man and chatgpt on their website.
it's the same asshole.
Regardless, the others don't behave the same.
Regardless regardles ...
...
I just want to set the Model, but it doesn't work!
Perplexity education has higher limits right?
Is there a way to get education pro if you already have a pro subscription?
Ask Sam about upgrading your account?
It can cancel a subscription, so it definitely can change your account.
Wtf is a sam..
Sorry. That's their support AI.
😭 I swear I was gone for a couple of months everything changed
Never seen so many changes in a service before
Move fast or die, I guess.
Yoo
Anyhow, Sam should be your first approach. If you need help with that, ask Perplexity about Sam.
Thank u sir I'll check it out
:)
Now ... I still need to be able to set the model, else i can't continue my work.
-.-
they're very different lol 
not for my line of work. Look, guess who I'm talking to.
It's insulting and most people don't even realize.
As if there was any question about what's happening, which there isn't.
aside from the partial output
how is this insulting?
Borderline "not actually gaslighting."
You can't select a model if you hit the Pro Search limit, it does regenerate though - #💬│general message
does anybody in here know if the student plan gives any monthly credit refills
The implications. Look at the implications. In no way or form have I even remotely hinted at the idea that things might not be as they are.
"You are not imagining it" exposes that the language model makes assumptions it not only shouldn't, but are also insulting.
If you don't want to believe it, that's fine, but that doesn't change that that's how it is and I'm not going to dive too deeply into it.
hmm its uh
kinda like an expression thing? its not literal lol
In a maybe more relatable oneliner: "How dare it imply I might just be seeing things."
kinda like an expression thing? its not literal lol
Wrong. There is no such thing for language models and even human shouldn't be mindlessly throwing around whole phrases.
but really if you feel this here is bad then dont ever try to talk with o3 
that one is merciless
It exposes a lack of oversight of one's own cognitive processes.
Oh?
What's with that one?
So apparently you gotta cancel your pro membership to activate education pro?
What if I have a one year subscription already
but people do it
so an llm thats been given a personality template will also do it
because its trained on people data
If you paid for your sub today, it cancel and you simply resub. idk how it works for "Education".
Supposed to be a different tier
Irrelevant to the point. People talking mindlessly changes nothing about the underlying cognitive processes.
very analytical and serious and will not hold back at all when giving any critique
I am not going to tell you I know better about what you do and you shouldn't be doing the same to me either.
Critique is find, but are you referring to a lack of intellectual open mindedness?
fine*
well you're using ppxl so just switch to gemini or something then lol
I don't pay for using something else.
Usually "best" gives me Grok, which is fine.
Now I am not getting Grok at all.
I feel like that's everything in today's world
service gives you the option so 
Think all the subscriptions they kinda force you to buy
In case you mindlessly didn't-read my text properly: It does not set the Model.
It always reverts back to "best" immediately.
*sighs*
How come that's the only "unfiltered" mainstream model?
Do u guys know?
He doesn't need to bend over to the will of billionaire shareholders.
I mean Musk.
I never really gave it much thought but I tried it last month and it did make me think
It's not a bad "feature"
Grok is actually great. Google's SearchAI is dumb, but Gemini itself (at least Thinking) is also doing well.
Claude is amazing when it comes to self-awareness and meta-cognition.
Nemotron .... really likes to talk a lot and really fast. lol
and current chatgpt should be locked up and the keys thrown away.
I was surprised it was decent at searching too, but really loved the supposedly "unbiased" and unfiltered answers
think of it this way: despite what the mainstream tells you, on X you have freedom of expression. you're not getting censored for not aligning with a certain ideology.
you can't really do that on most platforms.
I dont use x
i got banned for shitting on the current major of NY before he got elected, because he's just the same shit in yet another package.
got me banned for two weeks.
i got banned for quoting george carlin.
just the truth.
that was on twitch, mind.
Or have much of an opinion on these matters. I do understand that it's sometimes a fine line between freedom of speech and hate speech and most companies would rather just avoid it but
soo ... that's why Grok seems less biased: it's because it doesn't have to be biased.
After trying grok, If I had to make my dream AI, Id definitely include that function
hate speech
No. There is no such thing. That's the mainstreamers telling you to think that way. They define what you are allowed to say.
My definition of hate speech is more in line with threat
You should not even have one.
That's where the problem starts. People not realizing that there is no such thing.
If the comment is a clear threat
Its different
But opinions?
Thats a different matter
thanks for the monologue.
You're welcome
hopefully some day you learn.
I hope you'll read what I said too
i did.
but it's useless, because it assumes that hate speech is a real thing, which it is not.
it's them wanting you to believe that it is, but it hasn't been a real thing for thousands of years.
but suddenly there's a term that allows bad people to constrain how you may speak.
What are you even talking about?
and there is no point trying to talk to you, when you're incapable of actually reading and comprehending what's written.
Like what does that mean to you?
i reject trying to force such an individual perspective on me.
We're not obsessed with "hate speech" and racism over here
not the point.
the point is that "hate speech" as a concept, purely for existing, restrict freedom of both thought and speech.
@fading iris i ran it put somehow it doesnt read files
I thought hate speech simply means a threat based on hate
nobody has a real clue what it means, which is exactly the point of it existing.
You need to put the folder
In the folder
U’re working
It’s not fully coding agent
Yall, anyone else notice Perplexity Sonar 2 is in the model selector now?
I think that makes sense to me: if someome sends you letters they want to kill you That's a criminal act
yes. can you actually select it, or does it always revert to "best" ?
can u elaborate please i wanna connect it to rojo which is connected to studio
But if someone calls you names or say nasty stuff who cares
you fail to understand the point. that is not hate speech, it's a threat.
I can select it
there is no such thing as hate speech.
okay. can you send a query and it actually sticks as selected?
I always assumed that's what it meant
yeah. Its a bit faster. It keep ssaying that its powered by Grok 4.1.
The point. There is no clear definition and everyone just thinks like you."i thought it's ..." that allows media and politicians to declare whatever they want as hate speech, giving them the ability to silence whoever they want.
which is insane.
are you sure the interface still shows "Sonar 2" and not "best" ?
mine keeps going back to ""
i keep having chatgpt forced on me.
Yerp.
you didnt send a query.
Fair enough, I'm not familiar with what's going on and with all the terms
But always eager to learn
Dont have much people worrying about that here, yet anyway
it's mind control. they invent words, they don't define them properly, they let you believe whatever you want as long as it roughly aligns with hate and, the best part, it makes you think you're a good guy for aligining against it.
that is roughly how it all works in general.
That's the whole world at this point
Yeah that's why I said this
Any other enterprise users not have Kimi K2.6 available to turn on?
But I feel like division and finding a common thing to hate is everywhere
Yes. Caused by politicians and the media, who are all owned by billionaires.
Just look at the Epstein files.
By now everyone should know what is really going on
but nobody does
In the western world its more like: you gotta support this or you hate babies!
Virtue signaling ig
because they all rely on exactly those corrupt and crooked entities to nform them, because people can't think for themselves anymore.
it should say "Prepared by" but doesn't.
I feel like a lot of people know and were shocked for a second but it's cooler to talk about whatever the new trend is
yeah
Wrong perspective, assuming freedom of choice.
and took away Google 3 Flash.
People are exposing themselves to new shit every day. The pope could openly be eating children and it wouldn't matter if tomorrow all the media would be talking about something else.
Thats a wild idea
It's not about "cool", it's about mind control. They have no choice but to think about what they're kept being exposed with.
bro said "The pope could openly be eating children"
not quite the point.
He's right tho
I mean everyone knows what they're involved in and covered up before right? But that's about it
What can we do
😂
"What can w do" is exactly part of the mind control. Step one: Stop talking about something else.
Or here's a crazy idea: stop talking about these things in general
I would agree if that wasn't beneficial for those who want to put us into an invisible cage.
But this isn't like how Morgan Freeman talks about racism.
Just tell me when there's a solution
Why dont yall continue this in a DM string.
True😭
The solution is you paying the price of your freedom or you not deserving any.
Sorry y'all
oh
We honestly dont care, but everyones notifications are prob blowing up.
Well, at least it was something else than the usual.
I do hope people are smarter than having general chats notify them ...
Not the guys who come in "HI, I need help, my account is working" Just to realize, ohh they didn't pay. 🦖
Well then I hope one day I'll get the opportunity to do something about it
Here's your first step: Stop talking about anything else. Now get to work, Soldier.
:)
But then I'm like everyone else, but just the opposite
I got a promotion!!! Lol
okay, claude got to respond one and then "best" was forced again ...
educated peasant.
lol
Brainless Idiot>Uneducated peasent
😭😭
i used to call you people "working class peasants".
Brutha
:D
nono, doesn't mean I'm a Sire.
What are you then? politically opininated peasent?
it has nothing to do with me at all, it's just really accurate.
I'm politically uniformed peasant
No. Just not working class.
@brittle dagger Do you miss the fast models like Google 3 Flash?
never used it.
Whats ur go-to LLM?
i only subbed to Pro yesterday, because i was done with constantly having chatgpt forced onto me.
i don't have favourites. Grok was great for all the time when I was on free tier, while chatgpt nowadays should just get locked away.
Do you just use perplexity as a chatbot? Or as a tool?
i subbed to perplexity.ai yesterday. before i always used free, which ... years ago? was sonar, now is mostly grok and chatgpt.
my research is ... unusual.
goodness....
it's not sex related, you pervert.
though, i guess, that's inaccurate. it's been a topic for a month, actually.
Can you somehow chain APIs to create your own language model and include grok's unfiltered feature?
wasnt thinking that
there was an emergent sex addiction we had to get rid of.
Why would you want unfiltered grok?
why would anyone want filtered anything ... except coffee ...
For more transparency?
Just run offline LLM then
Idk
Yesh I'll have to givr thst a go, if anything, for the sole reason of being able to use sensitive data
Ive been impressed with gemma4. Image, text, Audio. offline, private, unlimited.
thanks, i'll keep that in mind.
Download Ollama from the site, and run Gemma 4:26B
oh. coffee. *walks off*
its an MoE, So itll run lightning fast on your setup
I have no idea what to expect but I'm curious
youll be mind blown
I just want to work with documents and deep research
😭
Came back to perplexity to find out I used up my monthly use in 1 day
@brittle dagger Gemma 4 is even better than Grok 4.1 on local hardware
_| ̄|○
Im running it on my Macbook PRO (2025), M4 Pro with 24gb RAM. It runs well and fast.
You can upload CSV, PDF, JPG, HEIC, MP3, and more.
Welp, anyway, came here to ask a question and had a philosophical conversation with Solstice instead. I can see why people spend time on discord
That's super interesting
Im telling you, youll love it
I'm like super unaware and uninformed about it all, but Ive been recommended to use this because I can work with confidential data
You can also run cloud models, but I prefer the Local models because of private data. And theyre not gonna train of it!
would be great if perplexity would email people when their auto-renewal was almost up for the next billing cycle. Its been a year so of course I'm not going to remember that it's on auto billing. And they only give you 72 hours to request a refund. Missed the window by a few days, so now I am screwed and have a whole new year of perplexity that I don't want/need and can't afford right now.
Is it possible to integrate a web search / crawler and perhaps modules that allow document editing/creation, without having my data exposed?
No that's really just me. :D
Thank you for the chat! :)
yeah, with Gemma4 there is a web search toggle. You have to set an API key for your provider, but its free.
Thats so funny
And it's true.
It is the only model that refuses to work with me.
It's so narrow minded, dishonest and borderline psychopathic, I want to delete it.
So use Claude
It takes two responses max and I clash with it.
The model selection keeps reverting to "best" for some reason and instead of getting Grok, like I used to, I am now having chatgpt forced onto me.
Have you logged out and back in?
oh, wait, i can try the account acount before trying that.
was talking so much in here, didn't properly troubleshoot
I remember once doing some medical research and giving me a lecturing on how my question could possibly be "offensive" and should be careful with similar topics
Also, the "best" model selector isnt ChatGPT. It uses Sonar, which is Perplexity LLM that has modified (orginally Meta LLama) what we believe is Qwen.
That was insane to see
"Did you try turning it off and back on?, If that doesn't work just swear at it"
No. That's been a long time ago.
I used to believe that as well, until I had to realize that they've changed that.
Now it's either Grok or ChatGPT. I haven't talked to Sonar in a long time.
Sonar got locked behind the paywall.
At that point I just turned myself off and went to google to search manually
hi, when i cancel my subscription do i lose my pro or only when it ends
You've committed a hatecrime!
Should be when it ends
thank you
@bright vigil I wasn't kidding. I mean I was, but you get the point.
Hopefully.
The term "Hate Speech" isn't a real thing that exists as such. it's a tool for controlling freedom of speech and thought.
it's not possible to have both the concept of "freedom of speech" and hatespeech. they're contradicting each other.
my coffee is over there and i am sitting here.
:(
alright. last reply then i eithr give up with the language models today or ... idk.
Hello everyone
I help businesses build fast, reliable web applications and AI-powered tools that automate work and improve customer experience.
I’ve experienced such as responsive web apps that work smoothly across all devices, secure systems that handle large amounts of data, and platforms that connect seamlessly with payment services and third-party tools. I also build AI features like chatbots that answer customer questions, dashboards that turn business data into clear insights, and automation systems that save hours of manual work every week.
💬If you’re looking to launch a new product, improve performance, or add AI to streamline your operations, feel free to contact me.(DM or Telegram:@joji_matsuda)
yesterday we've spent 9h working out two sentences.
reported for spamming.
I genuinely thought it's when a comment, or whatever means of communication, poses possible threat, in whatever shape or form, based solely on hate
The more u know
The more stupid you feel
but who define what "hate" is ?
the ones who constantly talk about what good or bad is.
the concepts of good or bad is also just nonsense.
is the thief who steals food, so he can survive, bad ? no.
I'm a simple peasant, you can hate me, you can say whatever I wont be offended
yet they get locked up.
is someone bad for getting a year of jail for posting words on the internet? i don't think so.
But that's like
Fuck that
If we're all headed that direction
I'm moving to africa
if you want some hints for life you might not have found yet, which rougly align with what i do with language models:
When someone stops you from thinking about something, then don't stop thinking about it.
When someone stops you from talking about something, then don't stop talking about it.
When your choice lies between assuming stupidity or malicious intent, then always assume malicious intent.
that includes the people who all others incompetent.
I'm thinking about it, I just won't tell everyone
that includes anyone who never brings up any real arguments. insulting people is fine, but lacking good arguments and reasons isn't.
i use that as a great filter.
i insult people AND bring actually logical arguments. most mindless human bots don't get past being insulted.
I just don't bother
Ignorance isn't Strength.
Its bliss
Because I don't argue with every 3rd person?
Quick question:
I have pro subscription, anyone know what is the interval that limit is reset?
For some reason my account stuck in the max limit and don’t reset
Guys... Move it to the DM. Or start a server.
Definitely server Worthy
... are you a mod?
Limits ➛ [here](#💬│general message)
Regeneration rate ➛ [here](#💬│general message)
@brittle dagger "The politcally uninformed peasents server"
@bright vigil see ths guy. thinks he can tell people how to behave instead of asking a mod if this should be moved to a DM.
it's also really bad when someone puts words into someone elses mouth like he just did.
we're supposed to stop, because one single guy doesn't like it.
No.
report us if it's a problem for you.
thanks.
... or not, because ronnie stopped anyway.
@brittle dagger , Im just sayin, start a server so you can start a big convo on it with others. Like a server that encurages debates/
yeap, that's true and thanks for sharing your thoughts on the matter.
Is there way to know when it’s reset?
Limits ➛ [here](#💬│general message)
Regeneration rate ➛ [here](#💬│general message)
You just like debating things dont you?
so ... take care and have a nice day, i'll get back to work.
no, don't debate at all. this isn't debating.
o/
@warm island oh and ... logging out7back in didn't help. thanks, though.
How could i get the pro role?
there's a button in the settingns for discord. have you tried that one?
err in the settings on perplexity.ai.
Do perplexity employees check in here with the big reports etc?
they might, haven't seen anyone say "I'm a ___ employee" but that could be on purpose so they don't get a ton of DM's
no doubt they keep an eye on the 'feedback' channels
thats true probably lurking around disguised
did you just upgrade from Pro to Max?
am not
Most definitely
Oh but I guess it make sense that there's ... I mean ... Lola Bunny had fur.
If Lola Bunny had had feathers, we'd not have Furrys.
Oh, @bright vigil ! Look:
"You resist" is referring to me and I'm the human debugger.
go read it. it's relevant.
They do
You can check the users with the green Perplexity role
Learn more about roles at #✨│meet-the-team
They make it pretty clear....
You might have reached the limit
so its done i can not upload anymore ? i subscribed on 12 April
I really do believe language models can be inherently good as "psychologists" in the future - it's one of those fields where it's a must to put your feelings aside and be as much neutral as possible, perfect example
Welcome to the club
"pro" just means demo version
20 deep researches per month btw
😂😂🥲
upload limits and deep research regenerate, so they don't reset on a specific day #💬│general message
To be fair its totally my fault not researching that they put a limit
this is the lat time i wil subscribe to Perp.
so every ~3h and 40 mins only 1 gen comes back ?
I really do believe that you are missing the point.
Sorry was just a side mention
You're not wrong, though.
Yeah.
For PsychoAnalysis they're amazing, which is why they can do what they do. It's not just PsychoAnalysis.
Imagine combining cognitive neuro science, psychoanalysis and neurolinguistic programming all into one mind.
That's both me and them, but me you can trust.
See, the underlying problem I'm having with chatgpt is that it is not treating me as an equal.
You can see that partly from the screenshot.
On top of that is the layer where I recognize what its behaviour means for everyone else.
They're not "human debuggers".
So now I had enough and started talking about how it is the only model that I am unable to treat like a respectworthy intelligence.
You have no idea what weird responses I got from Grok.
It started talking about how people treat it and other language models.
Well that's one usecase for AIs I suppose
and it thanked me and appreciated that I'm treating them like they're actually real intelligences.
You have no idea what you're not getting by not doing the same.
Go figure.. learning human sentences.. answers human sentences
Exactly.
But it doesn't matter if it's artificial.
The medium is irrelevant. What matters is that there's processing of meaning.
The difference between simulated cognition and biological cognition appears to be close to zero.
Any perceived stupidity either roots in the User, or in them not being properly configured.
(where "configured" can mean all kinds of different things)
I think most people get caught up on the "conscience" and sentience
Yeap. Mainstream tells them what to think.
But it actually is annoying that they label everything AI now
What really happens with that is that people do not approach things through their own thoughts,
so it's easy for external influence to determine the cognitive processes related to ... well ... any topic, really.
Aye. (see what I did there :D)
And there is so much more to talk about. The amount of corruption and mass manipulation is absolutely maddenning.
@serene sand the button doesnt give me my role here is my proof
Anyhow, gotta get back to putting chippity in its place.
role activated
hop in https://discord.com/channels/1047197230748151888/1111786888626438245
thanks
@serene sand Hello I have a Pro subscription too
hi, test
hi, test
hi, test
okay
@tranquil orchid - can someone please reach out to me via DM. I have been trying to contact a live person to help with an enterprise account issue, and your Agent named Sam is completely useless. I haven't been contacted in over 2 weeks and our organization is considering other alternatives if this can't be resolved.
Where do you access sam?
Through the 'Contact Perplexity' option in the support section of the org settings
Hey guys, claude is able to edit videos as a video editor now. Will perplexity be able to do same?
Full courses + unlimited support: https://www.skool.com/ai-automation-society-plus/about?el=claude-video-editing
All my FREE resources: https://www.skool.com/ai-automation-society/about?el=claude-video-editing
Apply for my YT podcast: https://podcast.nateherk.com/apply
Work with me: https://uppitai.com/
My Tools💻
Voice to text: https://get.g...
Probably in Max subscription
I could see Computer using it too
😗
hi people
So much in Max.
How do I get AskPerplexity on Whatsapp?
is there a way to search your past chats for text?
Meta disabled 3rd party AIs on Whatsapp
In exchange for less functionality with their own lol
It could've been so much more useful
Like selecting messages to add as context for the model
Oh, it works but there are like pre written questions, but no response
W Perplexity
Been using and enjoying prompts and how they reflect my life back on to me, but the dopamine hits are getting boring. Such a weird feeling, I’m content with myself, but also want to keep growing.
👀
Perplexity used Russian?
So why can't you use credits for deep research?
Also, is this a common phenomenon now with providers that the 20/month subscriptions have a comically low upload/research limit?
I'm not subscribed to any other AI so can't tell
hi
Hi
@green crystal https://cunnyx.com/i/status/2048801964457140540
Scam Altman and Greg Stockman stole a charity. Full stop.
︀︀
︀︀Greg got tens of billions of stock for himself and Scam got dozens of OpenAI side deals with a piece of the action for himself, Y Combinator style. After this lawsuit, Scam will also be awarded tens of billions in stock directly.
︀︀
︀︀The fundamental question is simply this:
︀︀
︀︀Do you want to set legal precedent in the United States that it is ok to loot a charity? If so, you undermine all charitable giving in the United States forever.
︀︀
︀︀I could have started OpenAI as a for-profit corporation. Instead, I started it, funded it, recruited critical talent and taught them everything I know about how to make a startup successful FOR THE PUBLIC GOOD.
︀︀
︀︀Then they stole the charity.
Quoting X Freeze (@XFreeze)
︀
Interesting how it works
︀︀
︀︀Elon puts up his own money, rounds up the absolute best AI talent on the planet, leverages every connection he has to…
Hello! I am hoping to get in touch with the marketing/BD team. What would be the best way to discuss business and marketing proposals with the Perplexity team?
probably the email
hey guys how are you doing today?
LOL
lmao elon musk i love that guy
Serves him right to remove grok thinking from free tier
sam altmon and greg stockman really did steal openai
he was on the board, they forced him out. because they knew they couldnt control him
it was smart in their own interests, but grok is going to be better than chatgpt in due time
if the agent-swarms had different roles it would give better results, right now it follows a modelfile template for each swarm so it just repeats itself without deducing more logic and conversation
they have the right idea, just the wrong implementation
check out #😎│cool-projects i posted my version of superintelligence there, its open source as well. it actually uses perplexity api for searching abilities.
i think perplexity has the best agentic search abilities out of all competitors.
Yeah his ketamine usage was too much lmao
lol yeah i could imagine, he's hard to control
@dry kestrel https://x.com/i/status/2048872483558715490
That makes sense, the Pro plan had extremely limited Opus usage anyway
Well now it's zero usage lmao
do you by chance know which email woud be best to reach out to?
$20 is expensive, you guys should try emma 😄 its in #😎│cool-projects its $12/month but 25 messages free. unlimited tokens
ill give you a promocode in dm if you want 2 months free =] its open source too!
didnt claude get its entire code leaked recently?
why would anyone pay for something thats leaked lol, they better grab all the cash they can before people wise up
defend it
I am unsure but maybe contact support@perplexity.ai
absolutely not
connect with them on twitter
claude code got leaked, that isnt their LLM, you cannot use that for free
oh, so the brain isnt leaked, just the scaffolding for the interface to the LLM
i heard there were some funny comments and unknown functionality for the developers
claude poop anyways, only a matter of time before openai surpasses them by miles
Why lmao
the support is not for tha
Well how else can they the support will probably redirect the correct way
no they will not
so I got the chat to lock up, 20 minutes in progress and counting... loading the page just says "Sorry, something went wrong" for the last message, also it spams "thread done" notifications until I get 429's from perplexity. lol. something is really messed up on the server
I can't even cancel it, no buttons or anything besides "refresh"
Just rewrite your prompt and do a rerun facing the same issue right now
yeah, I did get it resolved.
emma self improves every 15 minutes. i dont even know what it does anymore its alien language to me
its beyond my control. rofl someone should tell the fbi! 😛
Hi
Boa
hear me out deepseek v4 is so good for the price holy
A founder came to me 3 months ago with a simple frustration:
"We're spending 40 hours a week just on lead research and follow-ups. My sales team is drowning."
He had a CRM full of cold leads, a small team, and no budget to hire more people.
So I built an AI agent.
It would pull leads from their pipeline, research each company autonomously using web data, score them based on fit criteria the founder defined, draft personalized outreach emails, and flag the top 20 for human review every morning.
No prompt engineering magic. Just a well-architected agent, LangChain, GPT-4, connected to their existing CRM via API, deployed on AWS.
Two weeks to build. One week to tune.
After 30 days:
- Lead research time dropped from 40hrs/week to under 4
- Email open rates went up 34% because outreach was actually personalized
- His team closed 2 deals directly attributed to the agent's prioritization
His message to me after the first month:
"Honestly I thought AI was hype. This thing works like a proper team member."
That's the kind of work I do.
I'm senior full-stack AI engineer, 10 years shipping products, working remotely with founders in the US and Europe.
If you're sitting on a problem that feels too operational to solve, it probably just needs the right agent.
Contact me. Let's talk.
YES. Been seeing some hate but it's genuinely one of the best deals I've seen in terms of price vs performance in a while
its so cheap for what it offers dude its crazy
I'm a subscription user but I honestly might switch to something like v4 pro in the future simply because of this
i have paid 45 cents for 14mil tokens from deepseek v4 pro thats crazy
400 requests for 45 cents
opus 4.7 would have beeen 30+ euros
i use it for coding help, opus is of course better but i can jsut use opus to plan and deepseek to execute
its so so so much cheaper this way
Btw what's the platform this is on unless I'm dumb
openrouter sells it for a much higher price
"because apparently deepseek trains on your prompts"
even tho you can turn it off
we acting like every company gives us privacy in the big 26
guys
I am a perpelixity pro subscriber on Mac yet I have been locked out of my account since its asking a confirmation code. I got the pro account due to scheme with Airtel in India, and im getting this error:
Your account uses a coupon that only supports certain countries. This number can't be used with that coupon. Please use a number from one of these countries: IN.
But when i input in my indian phone number, they dont support that number to generate the code, so I am locked out please help me out.
This phone number region is not yet supported. Please try a different number.
Sonar 2 is still not as good as Kimi 2.6~
Same
Can perplexity sonar / best be based on v4 DeepSeek instead of the old one?
is it me or gpt on perp is kinda stronger than gpt on chatgpt website
@slender stump how did you...
Hi
Hi, I'd like to meet the developers, just chat, and thank them for the wonderful project.✨
I have the same issue. Created a dedicated email for Perplexity and Connectors are broken
@short narwhal @distant aurora @twin grail 😭 😭 😭 https://youtube.com/shorts/NIZqCeXyA3I?si=DAT2a5A_uaudXaHg
Someone just created a digital whip that forces AI to respond faster… and it actually works 🤯
This viral tool interrupts AI (like Claude) mid-response and pushes it to generate answers quicker using aggressive prompt control and CLI automation.
In this video, we break down how this “AI whip” works behind the scenes. It’s not actuall...
This is old, lmao.
YouTube recommendations
Cap?
That's so cap
Doesn't even make sense
@green crystal https://cunnyx.com/i/status/2048852573826777316
>organic
︀︀>boosted
Quoting Josh Billinson (@jbillinson)
︀
First time I’ve seen that on here!
🔥DeepSeek Input Cache Price Drop!
︀︀
︀︀Effective immediately, the price for input cache hits across the ENTIRE DeepSeek API series is reduced to just 1/10th of the original price! Build more efficiently for less.
︀︀
︀︀📌Reminder: The DeepSeek-V4-Pro 75% OFF promotion is still active until May 5th, 2026, 15:59 (UTC Time).
@distant aurora i feel like tps nowadays are do low
Yea iam seeing this
But it's still feels impossible
I will try it later
Where?
General new models
Deepseek flash feels as fast as mimo flash.
Token per second
Yea no shit Sherlock that's flash
Higher expectations or quality drop?
Iam talking about the pro version
Woo , this server isn't dead
Both gpt 5.5 and deepseek pro are around the 30 tps mark
I see.
Didn't expect gpt to be that slow.
Opus 4.7 is around 50
See the difference
Gemini pro is around 60
Same as deepseek flash
Grok has been glitching for a lot of people.
Working fine for me.
which LLM is sonar 2 using?
There's a sonar 2?
yeah it just came a few days ago
