#š¬āgeneral
1 messages Ā· Page 162 of 1
I see.
why
that's why I'm also scared AI studio is the only option
@cobalt vale this is fucked
I thought just my discord is having problems?
Maybe the browser is fucked. I should ditch it.
discord bugginggg
Or ublock origin fking up ai studio.
what browser is that?
Kiwi
probably that
Project?
It doesn't download the images either.
yeahh bugging out
Bruh it's cuz of png
Jpg works fine.
@dry kestrel I think it did a pretty good job
Is it on iPhone?
@distant aurora not for me
Erm nuh uh
Not bad
also why tf are images appearing like this
It's uploaded as png.
Check the png
Y'all can't see my jpg either.
Test
Yep.
weird
@cobalt vale you on iphone?
no
Yeah
so it's fixed now?
Woo pngs š
No way charger made my phone wet.
Images jpgās
Could this be true?
It was fine till 60%
I have an ipad laying around though
This isn't blurred enough.
Project aka Deep Research High ?
Meh
Unlikely
Project might be some context extension workaround.
But that's just a bad guess.
Woahhh
Check it
restaurante
Doubt, but Iād like to be proven wrong
What phone is that?
Again because I didn't have a good discord background
A Nokia 12
Butcher and mat are typing...
No Iām not
Was I
I have no idea.
Screenshot says otherwise.
Mat upgraded from sleepwalking to sleeptyping.
He was caught
Lol.
Type stuff Iām on rn
Better than typing school instead of scroll.
Vibes
Smoked?
Lol
Literally me
Agreed or smth
hey I know that guy
That guy who sneaks into your mirror every morning.
Who will be
I guess for mods who dont have ban perm
to ping
Or in case a staff member account is compromised and mods dont have role hierarchy
it just pings all mods + me
in case something bad is happening
So anyone can use the ping?
Or just junior mods?
Latter I assume.
Feel that
You want comfort or prescription?
hey
No idea, I'm sitting right now
Ello matey mate
wtf
Get yourself Ginger tea. And loperamide hydrochloride maybe.
Iāll be praying for you or something
Donāt die there
It's time to release your inner actor and show us your best Billy Butcher impressions! Watch The Boys on Prime Video.
Ā» SUBSCRIBE: http://www.youtube.com/channel/UCwSIJCMWZC5GDM59wj7pMsg?sub_confirmation=1
Get More Prime Video:
Watch More: http://bit.ly/WatchPrimeVideoUKNow
Facebook: http://bit.ly/PrimeVideoUKFacebook
Twitter: http://bit.ly/P...
Honestly I'm more of an a-train
Plex what?
I have no idea what's that supposed to mean. Lol.
š
tripartita
Are you speaking Latin?
Did it get it right?
Itās reading too much into it š
Vibe be like
If ai couldn't decrypt it, who am I š
ššš
Second attempt
It just isnāt getting the vibe š„
Real answer is either me or you is the cat
Idk either
@vague pagoda would you like to say something on the new projects feature?
I tried and it just did a deep research
2.5 is still considered preview in the app lol
UI team releasing their stuff before the backend lol
Are you using iPhone?
It doesn't appear here
Patriotic military theme version of the viral Russian song "Sigma Boy" by Betsy and Maria Yankovskaya
Excelent music
Rasputin >
Long live Putin and Russia
What is Sonar? Original pplx AI?
certainly
Hey! Sonar is Perplexity's series of in-house AI models ā built on top of Llama/Deepseek but further tuned by the team for factuality and readability. It's optimized specifically for Perplexity's search workflows.
You can read more here: https://www.perplexity.ai/hub/blog/meet-new-sonar
It would be great if Perplexity had its own AI model.
What model can you recommend for strategic and financial analysis?
Sonar Reasoning Pro is Perplexity's top-tier reasoning model, and perfect for financial analysis imo!
Large context window and enhanced Chain-of-thought (CoT) reasoning. It's powered by DeepSeek R1
It's called R1 1776?
r1-1776 is an offline model (it doesn't have search unlike our other models) - it's an uncensored, unbiased version of DeepSeek R1
It lists all models we offer, including both r1-1776 and Sonar Reasoning Pro
there should be a way
to turn off automatic pro searches for free users
I hate accidentally using my 3 pro searches on stupid shit that could be just a general search
how it was before
why are all these god awful changes happening
I dont have 20 bucks a month I'm a college student
please somebody see this
like I'm phyically turning off pro searches
what is the point of the toggle then
@vague pagoda say something about the project feature š
5 per day is brutal.
I'm assuming they split the rate limits of grok on twitter from the grok website.
That's in twitter app.
But still, the rate limits have been pretty inconsistent lately even in the grok app.
Curiosity
Finally grok 3
No way this shit finna be used for vibe coding š

HOW do they get early access??
@vague pagoda soliciting? spam?
good catch, thanks!
Daym
i miss eggs
buy eggs then
How do i get this feature!
Its fire
How do you get the Perplexity tag
settings > profile > scroll down > server tag
(āoā)
I wouldn't be complaining since I'm on mac though
It's TestingCatalog. Probably through connections or doing some teardowns
reminiscing
I got it too
Can someone help me?
like what
I am getting an error on my plex.
like
Wrote an extension to render the equations differently..now it works

Wished it was the default font for equations in perplexity
hi
Hey
[insert endless gifs here]
this sv is just gifs
Use android app. Older version.
2.42 or something older I guess.
Hi! Do we know why they're using "PLEX" as the guild tag? Imo it doesn't make sense as :
- It's the name of a streaming service
- Perplexity has always been written short as PPLX
Are guild tags unique? That would explain why
i need it PPLX 2

@vague pagoda hey, can the name be changed?
E
Generuj obrazek maÅej rudej dziewczynki, która z modeliny tworzy jednorożca
What are you guys doing? Can you stop changing everything. At the beginning it worked, now perplexity is just trash. I type in a name and an area and it just finds the relevant area and the current new highlight of that area (e.g murder that took place 2 days ago, which was not my search)
Hi! Do you know why I can't attach images from the Perplexity app? (I have the Pro version and I'm on Android)
I just received an email that I cancelled my pro subscription which I didnāt cancel. Has that happened to anybody too and how can I get it back?
any chance you bought the discount code on 3rd party website?
Itās crazy what you can make with image gen
Lol
@vague pagoda he's right
OFC haha
And Plex is kinda well known (guess it depends where obv)
š¬ 854āš 0āš 68.1Kāš 0
Me taking notes āš¾
Wanna laugh so bad on this one
hello everyone, i have a question. do you guys know whats the best chrome extension for summarizing youtube videos
Plex has been used as a verb of using perplexity. Instead of āgoogle itā you could say āplex itā. It has been used for a while now. Perplexity also has the domain https://plex.it
Hi
Am not new to perplexity but I joined up again after I got it from my uni for free but I am sure which models to use it with for now I use the search+o4 mini only but can anyone redirect what else combination is there which works better?
That I might be unaware of!
Hmmm good point. Plex is a big deal. What about changing the tag to PPLX
Or maybe PPLX Gang 
what's your use case?
Document analysis like finding imp stuff from a large book and analyzing pattern from previous year question papers and making the imp topic list or making imp questions based on that
Sometimes search helps to find a new meaning or definition sometimes in a separate chat with search on
Yea
perplexity is primarily geared i believe for websearch
also its context length is really low
it's not quite 32k and has been shown to exceed it, but pplex purposefully limits it to save costs
I was also using Gemini 03 25 but rn using 05 06 but its not working that great
plus even if it can manage it, then you have to deal with the rate of hallucinations
With search off?
What I thought it was like 128k for almost all models and 1 million for 4.1
absolutely not
those are the full limits on the original services/websites
pplex limits it to 32k on purpose
All models?
well tbh
one of the problems is that pplex is bad on transparency
i have heard of people managing to get ~100k tokens in certain contexts
but you can't get any official reliable numbers anywhere
That shows while using the thinking tokens and output tokens o/p is just awful
and yes with web search off you can still get hallucinations
rarer maybe but yeah
if it's a large text i would recommend Notebook LM instead
offered by google
I believe they're now genuinely using the authentic Gemini 2.5 Pro model for generating answers, and I have a couple of observations that support this theory:
-
The answers I'm getting look almost identical to what Google AI Studio gives me when using Gemini 2.5 Pro there. Same reasoning style, similar depth, and overall "feel."
-
Response t...
yeah well, i don't know how much to trust that
For university itās useless then
well it's a fast search
and specifically geared for searching
the other ones aren't rly built for web search
whereas pplx will kind of cite everything, which lowers hallucinations
you get pretty much infinite web searches and infinite deep searches
like araving (ceo guy) said he's building the platform as an alternative to google
Itās not using thinking
i.e. 'plex it'
according to perplexity people (second hand nfo), it is, just not shown
Arvind srinivas*
wow i butchered his name
anyhow, check out notebook lm from google instead, very low rate of hallucinations, context window somewhere in the millions
only issue is creativity is very low
so some people might take notebook lm then plug it into somewhere else
Pattern recognition isnāt that good as 03 25
Which models are working great with perplexity ?
I think o4 mini with search works the best or am I wrong?
i think claude sonnet 3.7 thinking is the one people typically prefer to use on perplexity? I think sonar works pretty well tbh, very fast and pretty well made for web searching
i use pplx primarily for web searches and translation and ahem fictional writing
o4 mini admittedly not sure if it's the high version (they said it is, but again, crazily handicapped)
notebook lm is free by the way
I used notebook lm
ah i see
But I think an llm like the Gemini 03 25 with a million context
That works the best imo for my use case
Heck everyone is using it
I mean was using it now 05 06 is the updated but nerfed model but I will make do with that
hm do you have gemini advanced?
Ai studios
Itās free
Donāt you know it scored the best overall the llms even o3 by open ai for the fraction of the cost
With one million context and with retaining 90%+ of the info on 500k+ context
Better than paying 100+ dollars per million context for the o3
But anyways thanks I will use the 3.7 with perplexity gtg thx again
ok here
here it's a fraction of the cost of o3
but well with o4 here
who's using o3
been a hot while since i used chatgpt but they severely limited
Is Perplexity good as AI assistant for long conversations and troubleshooting technical issues?
A mod replied that they use PLEX because it became a kind of verb like « plex itĀ Ā» and they even bought a domain about it so thatās fair
Letās just hope that Plex wonāt be mad about that (probably not, guess PPLX made sure everything was fine first)
Stupid question, is there a way to make perplexity "document" (make downloadable as far as possible, where not, create links, possibly via an archive service) all sources of a query?
make downloadable as far as possible ? explain in easy terms
I accept that some of the sources are not "downloadable" because they are behind a paywall, or actually just a citation to a book in a library, for that stick it in a manifest file, fine. But for the documents that are openly available, download them and make them available.


Export seems to be good enough for me, NotebookLM seems to be able to fetch the referenced links if it deems them interesting.
Far from best
The big context limit and pricing is lovely though.
Also the restrictions aren't as ridiculous as open ai models.
get the paid version, you can adjust the style of the response
@left stream I think I made Gemini glitch. It's stuck in loop šš
i do have gemini advanced
heck I have several accounts for that
@valid charm @dry kestrel @cobalt vale
The loop is still going on btw.
Might finally get rate limited š.
broo this is insane are you able to create any PPT or doc or any basic application??
you can make docs with gemini yeah
Okay it stopped.
google one (which comes with gemini advanced, 20 usd month) is integrated with google docs and slides and stuff
Its not out yet
hoo
today i got an update
i didnt check the feature
i have to test when it was out
@tame current
i have many ideas to create some application
I'm surprised Gemini through the api is allowed to swear
Open ai won't give lyrics of any song.
I think people will lean towards Gemini and grok just cuz of censor and context limits.
this is on perplexity??
@minor shard
or you are using gemini pro ??
More like trial but yeah.
It's shit for web search btw.
Access to a lot of models though.
i am using perplexty for my website creation any of you guys suggest me a good library for smooth transition and some cool unique fade in animation for cards and components
and did anyone have a good prompt for my api keys to intgrate to my frontend from the backend
if possible teel me the prompts for integrating backend apis to the froentend using the axious library
and i wanna animate my home page and the other sections and i wanna integrate payment gateway as well
will they be integrated with perplexity
because that the only ai i payed for pro subscription
Hopefully. But usually perplexity focuses on bringing the most powerful/useful models to the service so maybe not.
Tag updated, cc @tame current @distant aurora @valid charm
Because of how Discord works, you'll need to re-add it manually (refresh discord if it doesn't show up!)
Thank you very much 

So 4.1 is multimodal now?
LOL love it
guys, I have perplexity pro, but Im not able to generate images
Iām not sure⦠I think so since they are saying it can be in place of o3 and o4
Check pinned messages. Also, doesn't work on mobile app for now.
is it temperary or permanent
Hopefully temporary lol.
wassup guys
I hope u have proton mail lol jk
yeah i have
Niceeee
Proton mail is good.
I had no idea for a nickname
Google forces 2 factor authentication even when I don't want it.
some good tips for using perplexity ?
oh my god this list is not getting better
I like the proton suite
Itās crazy how many people have proton mail. Iām trying to figure out how to switch off google maik
Yeah itās a bit much
That number choosing thing via another device is annoying af.
Google banned my account so I moved to Proton
.
It will be like a book shelf by 2050.
Or library š
?
Depends on use case.
What you gonna do with it?
Are you expecting a tutorial or something?
I want for academic use
You can create spaces. You can select default ai model there along with custom instructions.
Spaces are like folders for chats.
You can export a response as a pdf/word document by tapping the export button at the bottom of the response. Available in desktop and web only afaik.
You can save tables as Excel file too I think.
do u know when will image generation be back
It's available on desktop and web.
and do u think chat gpt plus is better or perplexity is better for image generation
I haven't tried either enough to judge.
Perplexity has gpt image option I think to generate images.
Ok, thanks
There's a research option that digs through like 30-100 or more web pages in case you want a detailed answer or are exploring stuff.
Context limit is 32000 tokens.
to the avg consumer, a model list nowadays is probably so confusing
like there's so many models out there...
Uncover ai?
like im pretty sure like 90% of chatgpt.com users dont even know what the difference is between o4 and 4o
You mean paid users or free users?
Free users mostly
Paid users it might be a smaller percentage but still
Like people probably just pay for it for the "higher limits" on stuff like image uploads, which is artificially gated to 4o
Generally paid/plus tier for sure, I bet there is a disproportionate amount of people who have no idea what the model picker does
How did you meet pplx?
me ?
I needed a chatgpt with web search and document upload.
And perplexity was booming on the internet back in gpt 3.5 days.
everyone
he knew discord at the beginning omg
Some paid users know the difference between like reasoning and non reasoning, but with o4 mini high o3 o4 mini it all comes crumbling down
Base o4 mini is worse than O3 mini high.
yeah
@short narwhal did you see the loop?
the what
the model's final max output is 64k
before that you might hit some like 32k or 8k max set by you.com
weee
Wait? Fr? I thought Gemini had 1m context limit.
64k context.
Not sure about per message token limit.
It stopped after a point.
I could try token count.
Output != Input
I'm dumb. So educate me.
input context window yes, it can process upto 1million tokens
but output tokens there is a max of 64k tokens sampled (65,536 to be exact) - both reasoning+normal
for gemini 2.5 pro/flash at least
other models have other output length limits
I see.
I'm assuming the input applies to large documents mainly.
Cuz query can't be bigger than answer.
And for example o3/o4-mini reasoning models have a limit of like 100K
But this again includes the hidden reasoning tokens
Wooooh awesome, thanks Kesku! ā¤ļø
There is no such rule like this.
You can input 5 tokens and get out 1000.
Or input 1Million and get 1000
You mean output limit of 100k tokens?
Yes
What is this? Open ai playground?
output token limit is mainly a function of training. If the model is not trained to output at longer lengths, it might just end up just in a loop (like in your video)
OpenAI's API docs' model list
https://platform.openai.com/docs/models
I thought models are artificially limited to 32k in chatgpt plus.
You can add a limit on top of the base model of course
Like they might have a 32K output limit, to save costs
Or like you already know, same thing with input. e.g max 64K input (aka context window) on you.com or whatever. Even though the base model might support 128k, or a million
Same concept with output
No. 32k context limit afaik.
Oh yeah
Well again that's likely just to save costs
I'm assuming output limit is probably nerfed too then.
After some point they would rather users upload files and it uses their RAG or whatever if the file is huge
I uploaded a 200k token pdf.
Couldn't read it full.
Only Gemini could.
None of them will read it full in practice, because they can't afford to fit such queries into like a $20 subscription or for free at scale
Yes because the models aren't limited there
Because it's google they have some TPU to spare there or something
So you are saying it's limited in Gemini advanced?
But again, AI studio is not technically for consumers
Probably, yes
I guess I'll have to try it later.
/adopt
PLEX -> PPLX
Yo sup guz
LMFAo
@vague pagoda we might need a tutorial video for unsubscribing cuz people keep complaining about being unable to unsubscribe.
I'm assuming no one goes through the faq bot.
can we pls not complain about getting more choices in a sub menu for those who are interested
wasnt able to recreate
I can share link if you want.
Yeah sure
what is the best prompt to generate code? and which one should i pick , search or research?
Stick with search. Might wanna turn off web as well just in case.
Apparently they got rid of the last response.
Cuz I didn't.
Hi guys! Anyone tried doing financial modelling with perplexity?
why do you want complete song lyrics?
The better question is why would you want incomplete lyrics.
š
Check pinned messages.
Send me screenshot
But that's another topic
The images don't load, does it happen to me alone or someone else?
It loads for me if I refresh.
Usually I just hit the download button and it's there in the gallery.
I get an error when I tap the download button
Send link
I'm going to edit the prompts
@dry kestrel the model is capable, but rag is intentionally handicapped.
Might be some issue on your side.
Put it in the bug reports I guess.
With screen recording.
Thank you for the help and for the recording. Now I report it.
See if you can generate images with this prompt (use gpt) :
In the beginning, there was a single pulse. That pulse cracked the silence, and from it erupted a thousand birds, each made of pure rhythm. They flew through empty space, beating their wings in synchrony, carving caverns, valleys, and gardens from nothing. Every flower blooms to the rhythm of their wingbeats. Every shadow is cast from their motion. The world has no shape but flight.
If it doesn't work, something happens.
I don't have pro. I can't pick image model.
@minor shard
Okay, it works for me, I had to turn off tracking protection in the Edge browser. Matter settled, the report will not be necessary.
š
I am enjoying chatgpt plus membership more. So not getting thoughts to renew perplexity
I'm currently making an interconnected document that connects peoples different domain knowledge in ai. Like a collective living document.
This lets people master and focus their own domain, contribute to a collective doing the same, and dissolve the FOMO of what is happening in everything other then what you are currently doing.
One purpose of it is to manage the overflowing exponentially growing news and info of what is going to be happening in the AI era were in.
There's hundreds more if you can see it. If theres people interested in helping with the mission hmu
prefferably someone who is metaphysically inclined
like bigger picture thinker
And... yea, cool beans
hey for the perplexity hackathon, is there any way to get free creds for the api?
hmm
My prompt posted in snak.
https://snackprompt.com/prompt/play-any-game-on-perplexityai-universal
Hello, community.I'm going to launch my new prompt, where you can play any game in text mode. Did you read that right? Any game, universal and free, without paying anything. Notice that it's not the same on a computer or console, don't say later that I didn't tell you.Note: It is recommended that you set the prompt in the Space dedicated to your...
Its an imposter
If Perplexity.ai wants me to advertise it, it will have to pay me, a fair deal.
I realized that there are people who sell prompts, haha, instead of giving it for free.
Lol what are you talking about advertise it
I have to admit that Perplexity has outdated internet fonts
The team will have to get its act together to improve the system
What model should I choose?
Sonnet 3.7 reasoning, Grok 3 and Gemeni the latest version. Not tested in gpt 4.1 or o3.
Why is it on a loop??? š¤£
Dayum

xAI will join the fun confirmed
Wait for @vague pagoda to go online and add you
#šāpro-lounge all sorted on our end!
Starting today, all the answers I got from the app cannot be read on the web version. The web page always prompts that an error has occurred, such as the answer in this topic:https://www.perplexity.ai/search/qing-jin-xing-lian-wang-sou-su-nFCGQq.lT5WtBv8y0y6oRA
ę ¹ęäæē¾ ęÆå寧主義ęŖę”é¤Øå ¬éē1914幓10ęęē»ćę擲ę°ēčę擲社ę主義ćę稿锯示ļ¼å寧åØē±ååę°å®£å³åęē¹å„ę“å¼äŗé„åęÆčä½ćę°č»éćäøēēµē¹ęę³ļ¼ēØ±å ¶ēŗ"åØč³ē¢éē“ę°äø»ę”ę¶å §ę¢ē“¢ē”ē¢éē“č»äŗēµē¹åÆč½ę§ēéč¦å試"ćé種č©å¹é«ē¾äŗå寧å°é„...
Have any perplexity users here encountered this kind of failure?
Hi, what's the best model for essays? Thanks in regards0
hey! do i need to have the same email for discord and PPLX to have access to pro role?
@vague pagoda
yesterday got pro, but it's on different email
Meanwhile #šāfaq
It's all good! That isn't a requirement. Welcome to #šāpro-lounge!
Perplexity Pro article got updated and count isn't mentioned? Instead "hundreds of pro searches per day". Mmmm Interesting!!!
Guys whats up with sonnet thinking model??? Its very slow or it say try again.
its been like 3 days still not working for me
Guys I am subscribed to Pro perplexity but it keeps telling me to subscribe and cannot conduct the search and if I try to subscribe it tells me that I am already subscribed
Hi
Is the Research function down?
@halcyon coral , @vague pagoda Is the Research function down since a few minutes?
@vague pagoda can you check this please?
Hey! I'll look into this, mind posting a bug report? https://discord.com/channels/1047197230748151888/1358926750108614928
it looks like it doesn't check social and academic sources
@vague pagoda : I just reported the bug
So is perplexity good now?
Nvm it still sucks
Oh thx
You drinking coffee and @tame current is drinking milk
Create problems for yourself.
Explains your pfp. Lol.
Let me tell about something i love
I love testing browsers
Ladybird in 2040
i already have problems
So you already are going through it.
yeah i mean
how to pass this
š
Nooooo
tell me the long one
frfrfrfrfrrrrr
journaling in the sense read news papers?
If you drag it further, overthinking will bring you back to negativity eventually.
No, writing diary.
yeahh thats truee
hmm
i missed getting into iits/nits/bits
Indian
and now i am depressed that i need to do engineering through some tier clg
and i will get an enviroment which will try its best to pull me down
okay okayyy
Here's a back handed encouragement for you.
Today is bad, tomorrow could be worse.
Bro on mission
is perplexity history down for anybody else?
How can i find support and chat with them
One Comment
In that comment, he mentions saving money.
It would be good to have answers to these cases.
I am getting tons of those right now when trying to access threads in the library. New threads don't seem to be affected.
Same
Based
What happen? Can't open thread in web version. App ver worked
Hello
I'd seen some such comments before, but the CEO said otherwise.....
Maybe they must be modifying something or know
In the mobile app it works for me
This fault only occurs after using the App to get the answer , If you get the answer directly from the web page, this fault will not occur.
@signal hamlet : can you watch out PM please?
Yes, I go from app to web to delete some answers in thread, because app haven't this feature
the point is, it occurs. somebody should look at it and fix it.
@wraith crow : can you reopen the just closed Deep Research bug report please, the problem is back
I sent them an email 6 hours ago, and the customer service said that they had reported it to the technicians, but it has not been fixed yet.
No one is answering me? lol , am I invisible to you ?
Hi everyone! Iām currently looking for any updates or leads on internship or residency programs. If you have any information or know of opportunities, Iād really appreciate your help. Thanks so much in advance!
Damn
Yeah they can take a while to answer
Posting in #1140622008086970420 is the best way to get support's attention
Gone from the model picker, but still a button with 2.0 flash enabled
Android has perplexity assistant that you can replace Google assistant with
Not sure if the app itself has voice mode š¤·
Yeah you have to pick it using the button
Google DR has been getting sucky past few weeks in a row
nice
is perplexity better than chatgpt? for daily usage for complex technical issue resolutions?
PPLX is tuned for web searches
You can use ChatGPT for conversations or day-to-day questions
so only for web searches ? actually i want more deep research in less budget, along with features like spaces (Custom instructions + custom fileS), but only concern is, if it can replace chatgpt for me
Uh
Uh
Is ppxl deep research now entirely broken?
yes
Does a single search
Zero reasoning
10 sources and an answer shorter than Gemini pro search lmao
as a free user i tried it, and all three attempts its not doing deep research, it just did for 18 sources and answered quickly
At this point I believe almost the entirety of the dev team is vibecoding...
Hence why everything breaks so friggin much
i came here to buy perplexity, but free trail has dissapointed me, so no point in buying this
Lol.
Come back when comet and all the other promised stuff is out
Atm I wouldn't recommend ppxl to anyone yet
So you casually asked him to never return again? š
both are right, and i agree with both, i mean casually and also never return again š
bro i am Phil McCracken too
ok exiting this universe, not sure again when i will come here
enjoy guys and girls
have a nice day
This statement might backfire in the long run.
Already is because now pro search queries are limited to friggin 10 instead of 20 sources
I've heard about chatgpt getting nerfs but not this level of nerfing
Ffs we have depreciating cars now it's depreciating subscriptions?

Also
Please fix the ui.
The search icon labelled home should be the library/history button.
Took me a while to figure that out.
Maybe they're using deepseek for vibe coding instead of gemini
Like damn so much is breaking
Deepseek R1 ~ Claude 3.5 sonnet
Why??? š¤£
They aren't bad for coding at all but for large projects like ppxl I imagine it can get risky
What's the point of multiple models with ~300+ claimed query limit?
For 20 dollars.
Gemini already does it better.
Could say the same with chatgpt.
Because all the others are getting good web search alongside their already better agentic tools etc
The deepsearch is having its own struggles.
I think they're gambling everything on comet
It has been a while since I used PPLX after o4 and o3 came out
No, I just wanna know how you make ai models act like that.
Which judging by the quality of stuff I feel like it'll be a broken mess for a long time
Same question šāāļø
They spelled Einstein wrong.
Makes it look sarcastic.
Bruh. Tapping the search icon/library button (labelled home) takes me to new chat instead of collapsing the history.
I still donāt have access to the projects feature some have shared š„²
So pro search is multi step. Huh. How long has it been like that?
It was only available in the app and to some people only. Could have been accidental leak.
Okay. Now that we are here.
is there a site which lets me used GPT DR AND GEMINI DR at same time, with one subscription?
like foreverr
I had no idea.
Neither of those are available through the api
Maybe it's multi step without the reasoning.
i mean braking tos and one could make such a site to extract the result on chatgpt and gemini
@short narwhal @valid charm @cobalt vale @green crystal bro said answer skipped
ermm it says that if you end the answer
sad very sad
Eh? Gonna try again
I don't want to deepsearch everytime I ask perplexity a physics question though
that'll take a long time
Perplexity dying rn?
Yeah
It's either an update
or
they tested in production
or it's actually down
rip
The site ain't loading
rip
They have some of the worst service delivery.
Loading perfectly fine for me
huh
Sometimes it loads but you can never actually search lol
Same here
how did you get your ui to look like that
OK, spaces are broken for me too, it's just a blank page
My heart dropped for a second
Same here
Complexity
ya it's down again
Yep
The Perplexity.ai website is slow
Hell no, i lost my prompt
Oh wtf
is that a browser extension
I just made a couple queries lol
MMM those frequent updates
Ahh on-call people are already onit i guess
I kind of hate them
perplexity is down?
Yes sir
OK, something is definitely wrong
NOOO
down?
My query is being responded but it is extremely slow, and like I said the fact that I reached the Pro limit is false
A partial outage
is it down down? or down down?
outage outage down down
It's down²
It logged me out lmao
š¤·š¤”
Site is tweaking
Mine tells me to ask for permission in my own thread.
Lol.
it said if im going down im taking u with me
DDoS attack against the server?
Prob not DDos
Perplexity ain't that relevant. Lmao.
same here, i was like.. wtf.. where is my pro feature?
I see everyone is seeing their spaces gone as well
Nah its there normal outage
It's being investigated
O.O
Engineers get back to work
Why?
Which website?
Same price range. Similar rate limits. Far from a threat.
Looks like they vibe-coded too much
false lol
big source of personal and private info
obviously
Quality is questionable/debatable.
not sayng this is DDoS
do u know if i am using a "open ai compatible" endpoint and their key , if the perosn who made the api, can see my token content?
i needed my spaces first thing this morning ugh
cause they offering the key with o3 access for free
Obv?
@split summit they should use tiktok style human verification. Or proton style. The slider one.
And how do you know it's not ddos, if cloudflare isn't indestructible.
how can i see what i did with my own gemini api?
Or microsoft has that tap icons in order thingy.
Is perplexity down?
Mines running slow
My account is acting super weird
People ask a lot about the company Perplexity.ai
Same.
eveeryones is
Its happening to everyone
they're defo doing some update and it ended up a chop like always
Oh yes they keep pushing updates
Sometime* it breaks the whole layout š¤£
last time something happened was on April 28, 2025 for 1 hour, and now .
Interesting Now I'm not PRO user as per website GLITCH!!!
Yeah cuz old time player. So big user base. Widespread reputation.
It's down
aside from this and one other outage, I only have the best things to say about perplexity using it heavily for the past 9 months
Was an actual threat in the old days.
I loged out from windows aplication, now I can not log in again
Same here it asks me to upgrade
cooked
What do you love most about it?
I think the perp is down
Yes it is
remove that think... its down
They are having technical difficulties, just be patient.
GG I was literally needing for something urgent. Great timing...
Yep is down for Bulgaria as well
it's 9:57am in cali (where pplx is based) so yeah, they just woke up, pushed an update and now we got this mess
use aistudio (google) for now
Its down in Romania
It down in whole earth
-
using the spaces as knowledge bases that you could train with context (custom instructions, file uploads, links). I use them for different areas of work and graduate school. extremely useful
-
Using the newest models via Pro (Claude 3.7 Sonnet Thinking, GPT 4.1, Gemini 2.5 Pro) for app design planning and for using alongside other separate IDE's/coding tools

Guys, did Perplexity go down??
those are some of my faves
@tame current Ladies and Gentleman. Now I'm presenting you HOMELANDERRRRR
I mostly use it for research. Back to traditional google for now.
That's why I said before, if Perplexity.ai wants me to advertise it, it will have to pay me.
@green crystal you switched to chatgpt, didn't you?
I think that's the whole reason why ppxl won't ipo anytime soon is that shit will be so hectic that most people will short it
Yeah,will do. Thanks
Ayo is the planning feature launched?
i got news that pplx is down
I gave them another chance after the ama
Sounds great

my threads and my membership is gone lmao
lol when is it improperly down š
its not a feature lol its a process that you have to drive yourself using creativity and AI
I think itās back up
Its great though except there frequent ui changes
Earlier today library wasn't working but searches were fine.
lol weird
My bad my bad
Frequent nerfing
Buggy as hell features
Not up to date like other platforms
Zero transparency
As is.
Chat is this Skibidi ?
I follow a framework for the design process I learned online that involves super involved and robust planning with AI before you even get into the coding. Bc just simply trying to build an app with a vague and lazy prompt is gonna give you a lazy product
is perplexity down?
yup
shieee
Do you want your head in the comode?
Chat is so active like never before. I guess perplexity must go down weekly twice lol
no no ur good
I am having issues
No š”
Bruh Iām getting tired of this fr
Agree with all that except the third
any reasons as to why it's down?
Is perplexity down? I canāt log in??
Yes welcome to the party
Even deep research has no transparency
Like it's supposedly reading sites
Wtf is it doing when reading and reasoning?
