#💬│general
1 messages · Page 188 of 1
Bob proctor?
I've seen enough lol
It makes your ai more spirited @timber niche less ram requirement and better reasoning
Peace out
Samsung code abuse to spiritual ai
One hell of a day
I’m testing out pro
It’s kinda like useful and uselesss at the same time
For me
For me please don’t start anyone
I got that one month trial but we’ll see it’s nice to have but hm
My chatgpt debates me.
No way 😭
Feminist piece of shit.
I saw it
What
Pro it's in prompt lab
My chatgpt debates against me.
Look at the logo of this lovely brand bruh
Ok then you asked it to
Rarely validates.
Nope.
My grok is good
It'll prove me wrong but then again it'll try to prove it's point without breaking my assumption.
Yeah so being Socratic
Just ask your model to be Socratic but tuned
To what you like
k bruh I’m out I’ve spent a tad too long in here.. 
He's basically like - I don't agree with this assumption but even if it was true - blah blah blah - hence it's not as bad as you think.
No I never said it was bad it’s very good to talk to something more objective
And it puts a lot of effort in acknowledging my arguments.
Wdym by validate then?
Like, talking to chatgpt feels like talking to a wall sometimes.
Like it'll keep spamming visit a doctor or get in contact with people even if I say explicitly that I can't.
Grok on the other hand will actually listen to that and stop yapping.
damn
bud im sorry but like
You should talk to someone somehow
HOW DO I STOP NOTIFSTIONS
?
So that I have to pay someone to give a fuck about me for 1-2 hours a day.
No thank you.
Yk what, ignore that.
I can buy liquor and cigarettes without prescription but not a fucking medicine that I need.
500 inr consultation fee for 50inr medicine.
Then some dipshit will come up to me and say that I should smile more.
The easier thing is never always the best man
I hope you aren’t
Try it 
Toxic positivity right there.
What do you want me to say lol
Actual solutions or silence. Or acknowledgement.
Okay I acknowledge your going through a lot, my solution is to try something you enjoy that is beneficial to you
There's a time for smiling. I'm not bitter all the time.
This is not the movies bruh 🙏
Just slowly try making the bitter times less prevalent
What do you like to do
That is good for you, a good habit
That's what you should tell yourself when you try to simplify solutions into smiling more.
Yup
But try it
😂 😂
Okay I’m out now you’re getting your silence 
Hi, I got this code PPLXLIUMBLHOTVAJ2QI for a 1-year Perplexity subscription. Just want to check— is this code safe and legal to use? Thanks!
@spare marsh @short narwhal
It's a widely abused code and the team is revoking access to those who use it
Kindly delete it from here
abused? aint these code are like 1 time use?
"Is that information confirmed by the Perplexity team?"
It doesn’t even show any credits, how to check
Because
the fast is only text to video
And generally speaking in the backline it's possible to reduce some quality for faster regen
Hmm what is the backline?
its kinda shit
more like a veo 2 pro
lololol
I have used comfy ui it's possible to have light model. Its often the same or different depending on what you are generating but for veo idk the exact difference
hello
Sup
hi @vague pagoda can i please get access to the pro channel?
Yeah how?
Please have official operations staff reclaim the membership activated by this promotional code. I don't want abusive users to occupy too many resources, which would affect the response speed and experience for us paying members. Let's put an end to freeloading, starting with me! Thank you!
Starting with you? Revoke your access?
How to get refund?
@vague pagoda i applied refund by sam Perplexity AI Support. Any additional action is required or not
Shouldn't be needed no
How many days it will take for refund @vague pagoda
ded chat :((
he used google translate or something
On one of my threads, I a little error message just popped up saying, "Thread is getting long. Start a new one for better answers."
I'm not sure how starting a new thread would produce better answers? I would be losing all of the accumulated knowledge in this thread. I even asked Deep Research and it agreed with me:
"The substantial knowledge base we've built outweighs any potential performance issues from thread length. Your point about losing this context is spot-on and more important than theoretical performance concerns."
Has anyone else seen this error message? Do threads have a length limit for Pro users?
you are prob reaching the context window of 32k
use deep research or deepseek r1 to mititgate this issue
those have a context window of 128k
Perplexity is only searching for Space's attached files and no web results included. Why?
If I prompt it to use online resources, then it stops searching for Space files.
How to get pro tag?
@wraith crow @serene sand
Did u kept web search enabled?
yes
Also in spaces the priority order for searching through is attached files>> web search
I recall that in the past it can have some of the queries searching files and some searching web, but now it seems to be either all files or all web, and most of the times all files.
Try adding this to ur prompt
search_sources: [source_files, web]
buy a subscription?
Can anyone help about refund?
this changes every search to python programming
Damn
hello
hello
Got yesterday
You can tag moderator for help @elder root
This is great. I used it and it worked for me
Thanks @carmine finch
Too bad its only 1 time thing
what do you mean?👀
I mean - you shouldn't post codes to be visible to everyone
Coz someone will just take it and then what
oh
Yep
haha
Mby delete or edit message
okay?
You can send email to Perplexity support and ask them what is going on
환불받고싶은데 어떻게 할까요
Open blackbox.ai on web browser and open the profile section and then credits
Where is this from
Google Flow
이메일 support@perplexity.ai로 문의하세요
Pro 구독자는 설정 페이지의 Intercom 채널을 통해 티켓을 생성할 수 있습니다
이메일 문의한지 12시간이 넘었는데 보통 환불 완료까지 얼마나 걸릴까요?
요청이 많을 때는 환불 처리에 지연이 있을 수 있습니다. 보통 환불 완료까지는 영업일 기준 1~3일 정도 소요되지만, 상황에 따라 다를 수 있습니다.
중복 문의는 처리 지연을 초래할 수 있으니 가급적 한 번만 문의해 주시길 부탁드립니다!
감사합니다
Directly through deepseek?
Yuh
What is intercom?
im a bit annoyed at pplx
guys anyone know how to fix
on phone, i use it then i leave for another app and it goes back to the homepage, i have to then go to library and then go to the conversation but it starts at the top and i have to scroll down to get to where i left off
on chatgpt mobile, i just go back to the app and it is ok, if the app is closed i can just go to the left sidebar and continue where i left off
too many steps involved with pplx
especially when i leave the app and come back often
Strange, this doesn't happen with me
Hello!
Check this, might help
Same
I’m building a community where makers, hackers, and creatives come together to build cool stuff — no fluff, just real execution.
Thought you might vibe with it.
Join us : https://tally.so/r/mR6xMp
Please post in #😎│cool-projects
okay
not wokring i dont have this ass/ rmeove option
@gloomy finch you wanted right
i gave him
Ok
nvm let me get some and store then will share
Okay
Thanks man
u need?
Sure thing
@gloomy finch how have you liked labs?
Did you manage to make something useful with it
Yeah I haven't made too much yet, just tried a workflow for something few days back it was good, though yet to do a complete productive use of it yet.
What bout u?
Honestly same haha
Are there any other samsung users that are unable to claim the 12 months of perplexity pro promotion?
They got rid of it cuz some people leaked it on social media and the code was same for everyone.
Ahhh. So no more 😕. A day late and dollar short
They fucked up. Should have kept code unique for each user.
How will Perplexity be stronger than the AIs on the market?
Hi all, I am a full stack dev. I used API for my perplexity mcp for months, now I'm finally also a PRO perplexity user 🎉🥳🥳
Nice
Yeah they did. Now they're messing up by continuing to promote perplexity pro. I wonder if the support team will honor it since they're still promoting it on the galaxy store.
Is there a perplexity lab collection? Is the most cool feature atm, I'm very curious about what people was able to do with it
They should
By making something
That'll actually work
I already read that 😦
It won't.
It's a 3rd party provider.
On it's own? I doubt.
Will probably build on an open source model.
Hm
Crippled ai is such a disappointment.
All the verbal gymnastics to get a quality answer.
Be it grok or perplexity.
just paste the link or click the + button
It hallucinates.
try shorter videos
Doesn't work for YouTube videos.
Define shorter.
less than 100k tokens maybe
Translate that to minutes. I simply pasted the link.
And it started saying things that had no relation to the video or it's title.
I can't tell how many minutes that is 😭
rarely or entirely?
Entirely
Try this link on Gemini vs ai studio.
I ordered a ₹2.6 lakh Flipkart Assured MacBook and got scammed — not once, but twice. Both times, I received a used and scratched MacBook inside what looked like a sealed Apple box. And this isn’t just my story, the same seller has scammed buyers for years and flipkart knows about it, but the seller is still active. In this video, I’ve s...
whats the fastest route to get support by perplexity as a pro user
Become a member of their team.
i've been waiting and talking for the ai agent for like 2 days nobody cared to answer me
well the second fastest way then
sam is stupid as fuck
omg this is new
unfortunately still no way to set audio only
if you want the pro, chat with me
Ai studio?
A YouTuber details his experience with a sophisticated scam on Flipkart after ordering a high-value Apple MacBook Pro M3 Max worth ₹2,62,999.
i told sam to fuck off and he practically told me fuck you
he? i mean it
hell of an ai support agent ngl
also I don't think it can see the title
sam is just standing up for itself
why do you have that warning triangle
I have no idea. It doesn't say anything on tapping/hovering.
open a new chat and paste the link again
I've used Deep Research for probably 95% of the queries in the thread I mentioned.
I can't, no matter how difficult the prompt is
Worked this time but triangle is still there.
Depending on what you are doing, it might be better to have it summarize your conversation and uploading that to a new chat
It's because you have url context off, that will make the triangle go away, but it won't make it actually summarize the video. It just tells me that it's "web" tool is only capable of handling text based sites.
Ah
If you are using it as an actual tool, I would switch to 2.5 Pro. I find that multimodal understanding relies mostly on base model intelligence
it allows directly scraping URLs from your prompt
it should summarize Youtube videos natively even when pasting a link
Would this prompt work for your suggestion: "Please summarize all that we have discussed in this chat. Export it to a DOCX file so that I may upload the information to a new chat."
can't export just one answer automatically, copy the answer instead
oi william
He does it locally now I think.
@muted dagger #📷│media-gallery message
ollama is for local llms
I see. Thanks
thanks bro
i see hammy uses 4050 6gb
i have a 3050 6gb
will see how that goes
i still got no memory funtion 😭
What is lm studio
@languid star did you share this already?
nope
Hi
Good morning, I have a mistake when linking my card to make PERPLEXITY API payment. Does anyone know what the error is due?
perplexity isnt really that bad at coding when I use e.g. claude or gemini it actually does really well why do some people say it cant code?
I got a code for 1y pro for free on my Samsung phone and when I try to redeem it it says invalid
It's been disabled because the code was widely abused
They should bring it to legit users soon hopefully
That's not fair ialready got the code on my galaxy store and wasn't able to use it while others had it illegally
Yes
Also on Samsung you don't get a code
You download from galaxy
Store and once you sign in, it will automatically apply it for you
Blame perplexity for using one unlimited code for every one in the promo
Like what did you think is gonna happen 😭
No it was in here
A similar thing happened to Google but they fixed it within hours by making it require usa phone number
So isn't there a support member or someone from the perplexity team to ask them about it?
Will comet browser be available on phones also?
Yes
Will they really activate it again for legit people this time
Or they will cancel the whole thing
Hi
Hi luke
I'm trying to apply the $4.99/month student promo, but for some reason it doesn't quite work.
you can see here that it says "4.99/month when billed annually"
it shows up for the monthly plan
but not when I switch to "pay yearly"
does this mean the discount will only apply for one month?
i guess its 4.99 per month for student and not 4.99 per year
I know that. I'm just wondering if the 4.99 per month will be forever or just for the first month
since the wording is confusing
and if you are really student why dont you use edu mail to get pro
I am using my .edu email
it's verified -- this is how I got the price down to 4.99
its already
already what
no idea i suggest you to contact
alright
Hi, just got pro yesterday, how can I know which model perplexity has selected for the query?
Try them all out and see what you like for different type of questions
I'd suggest sonnet 4 thinking generally
Hey guys, got a question. I love Perplexity Pro but I am curious about something. I have been chatting with ChatGPT on things I know and and it is a very great conversationalist, plus it remembers my ideologies and knowledge on most matters so when we chat it bases its answers on that. Perplexity Pro is too excellent for research, search and study. I am wondering if running Deep Research on a topic on perplexity then moving the answers to GPT to grasp the information better is a great move?
Yea I know but it's quite tedious, isn't there a way to tell which modes perplexity has selected?
Do you have gpt plus?
You can't see what model it is using if you chose "Best"
If you select best it does
Nope, why?
Because deep research in perplexity is quite bad
It is? Hallucinates answers?
If you had subscription with gpt i would've suggested to use the deep research there
Nope it's limited by tokens and the model isn't great
New deep research this week please 🙏
If you want you can try what you mentioned here
It's already been a week since AMA right?
Pretty much haha
Did you just wake up mat lol
AMA was on wednesday, and didn't they say next week or so and within a week
Let us hope so
No I woke up uhhh 4 hours ago?
Nice 12pm
The tokens are a non-issue. I do not use Deep Research as much. But I have heard the model is not great. What workaround could I use? Maybe a different model on search mode? I still have 4 uses left this month of Deep Research on GPT.
Memory should maybe help with it remembering ideologies? Introduce part in pplx can help too
Use claude 4 sonnet thinking!
hi, I just subscribed to perplexity PRO, I want to play with LAB options, I see the icon at the bottom, but can select it, it selects the "Search" instead, someine can help me plz ?
This was my sleep tonight
You wake up so much
Seriously? It's that good? What makes deep research on Perplexity "bad"? If you don't mind answering?
Idk what happened to me
The model isn't good, we can't even change it
This was the night before
The dance floor scared you
Prob
Still uses Deepseek R1?
Still alot of waking up
It's not been disclosed AFAIK
Erm normal amount, + I don’t remember them
Apple watch hallucination
Combination of models
You wake up in the middle of the night too!!
Yeah maybe, it's a natural thing
But boy you were pretty much half awake
I used to track my sleep
I mean yeah that's not normal I couldn't fall asleep again
Hmm okay. I will waste one ChatGPT deep research on the same prompt I asked Perplexity to see how different the answers are. Then I will use the same prompt on Claude 4 Sonnet thinking to see how the answers differ from the usual Deep Research.
Anyways, thanks guys!
You're welcome!
You should try black box mat, they have 3 months free trial
Ton of models
Builder feature also seems interesting
Growing up is realising I have no use case for any other model besides GPT and claude
nahh
Interesting
Can't share chats where I use claude thinking without it doxxing info about me in its reasoning
Hahaha okay
Memories be working against us
Every time it's starts with
Oh it's x time in their x city
Like come on
🙏
omg I had perplexity save a silksong memory but now it deleted it for some reason
Even it doesn't believe it'll release
💀👀😭
Pick one
Claude my goat…
You guys are in love with Claude. Is it that much better than the rest? I don't code so I don't care but I study and research a lot. Is it that much better? Should I just keep it to Thinking when I wanna use reasoning?
I just personally like how claude (especially the thinking one) answers
I was a GPT 4.1 shill before sonnet 4.0
Hahahahahaha hallucination
@spare marsh hello sir can you help me about cancellation and refund process it's already applied by web sam ai assistant
No dont use it
It is bad rn
Does the Android app still not have the new 2.5 pro and finance search or its just me? Already tried uninstalling and clearing cache and no change
Damn guess at this point it’s just preference. I always liked how o4-mini was formulating answers.
Well I will see
Wdym
I can see the 2.5 pro
what are u gonna be using it for
after the new claude update i feel like its worse
I wouldn't be suprised if it's using the new 2.5 pro behind the scenes
Even if it doesn't specifically mention it
Literally just research. I don’t code, I’m no developer. I just ask it loaded questions, politics, geopolitics, ideological issues and other topics.
I don't know how to check that, but the web version shows the version date on my end
The finance is the main thing killing me rn on Android tbh
Yeah, don't think you can really "check it"
The only real issue with o4 mini on perplexity is that they most of the time use the normal mini variant and even then it's pretty limited in reasoning time
Also, the lack of it being able to use searches within its reasoning chain diminish its capabilities
Sometimes I wish I wasn’t just slightly above the layman on these matters. I’d just use my underdelivering AI and love it. Now I’m here frustratingly trying to think which is best for my use case haha. My gf uses perplexity pro regularly and doesn’t give a rat’s ass about each model’s capabilities.
They don't really differ a lot on perplexity because most of the models are nerfed
But some still stand out
Yeah I know that
This is what I’m trying to find
Claude thinking is the best for search
That's just from personal experience though
Fair enough. It’s answers are actually satisfying
💬 383 🔁 0 💜 99.2K 👀 335.3K
Real
#meme #literallyme #tylerdurdenlives #tylerdurdenfightclub #tylerdurden #tylerdurdenedit #tyler #fightclubmovie #fightclubedit #fightforyourmind #viralreels #bestedits #twistending
O4 mini better
I literally showed you an example.
Depends on what you use it for too, like Perplexity is good for me for finance but rn if I wanted to discuss some philosophical topic or something that requires actual reasoning I'd still go to the actual Gemini app
Or ChatGPT for the voice mode
Just when i thought i'll get rid of sam if i reach out via email, it popped out in my face and said hi there, i see you have a ticket opened already, i'm sorry if i'm frustrating but you will have to wait for the team to respond
i hate this agent
Go wherever you want. Explore everything.
The new update for the voice mode there is nice
Real, acceleration tfw
It's fire
Hello everyone, let me ask, is there any way to change the voice of the article to Vietnamese?
Read this headphone part and put it in Vietnamese.
Does it play in english currently?
ko
The sound is very strange, it's a mixture of Vietnamese and English =)) never mind, it's still readable
Ah I haven't used pplx version for awhile
Ok b
I forgor

o3 time
Wdym?
Hey guys
Does anyone have contact information for an enterprise sales team member? Help !!
It’s really good at reports I feel so that’s probably what they mean
It’s goated
Does anyone have the memories feature working? If so, how did you get it to record instructions?
Yeah, it works for me. Check your settings, there’s a toggle. I don’t know how they got recorded, but it just recorded them.
It's enabled. I didn't do the recent threads one tho. Just nothing saved yet
The out of nowhere wife lmao
It should save things from time to time when you ask questions like you normally would
personally i just told him to but in general it does take track from your projects
Ex it knows what exam iam after and what software iam trying to learn from the spamming queries i have done
In other words just be yourself 🥲
Bet thanks guys @gloomy cipher @short narwhal
It's good to see another 24 hour user
I wonder what this person went through.
🫡
Yess
Cuz I thought damage stunts growth.
Even that stunts growth.
Well it's about how you take it
Fair
Some age like fine wine
Others age like milk
o3 by itself can work like a mini deep research
Need a more detailed prompt though
You mean homelander @languid star
I did some self-psych background on PPLX, comparing it to what therapists have taught me, and I found it to be really good as long as you aren't lying to it. I could tell that it tries really hard not to hallucinate.
i mean you dont have a reason to lie infront of ai where as infont of a therapist you might be ashamed or feel insecure to say anything, which model did you use tho from what ik claude thinking should be the best for this but the new gemini might be superior
how do I get the pro role?
i tried joining from the "pro discord" invite but looks like it needs a role
@spare marsh
thanks for taggin the mod
i might have tagged the wrong mod 👀
i wasn't expecting the server to be this quiet
yeah, API team
if you are free
i want some review for my
wut
project i made for perplexity hackathon
there are times when its chaoss
im like nail biting and stuff cuz i think i messed up the frontend 😭
fitting of a cat
schrodinger
@vague pagoda @wraith crow
They might be busy but you will get your pro role soon!
thanks got it
You gotta leave the server and rejoin from the Pro link to get the role
Literally yesterday
Big things happen
Oki doki 🥹🥹🥹🥹

🤤
I was literally confused for a sec like am I in the wrong server
NOOO
To enforce my meme collection
The French meme will remain in France
more memeeeee

People in general.
Do enneagram test.
It will tell you exactly what kind of person you are.
Mbti is meh. Enneagram is the real deal.
Lmao. Accurate.
Who’s watching wwdc…?
might launch 150b model
Apple?
Idk if they’ll do it this time
Unless that’s part of the models they’re giving devs access to + shortcuts 🤷
leak said so
Are you watching it in the tv app?
Is the Samsung offer back?
Craig 🔥
Morning morning
For only 999.99$ per month you can enjoy the newest product of apple 😃
Severance is the reason for it's existence
Severance delivery system
Cinema
@cobalt vale did they even use private cloud for anything
Wait isn't it the reverse?
that was based on WWDC
60hz display for 700-800 dollars.
And their display repair costs twice as much as Samsung.
Oh boy, private and on-device hallucinations!
Lmfao
liquid glass 🔥
One of these devices and the matter is settled.
And if it falls?
just get AppleCare+ ✅
Let me guess it doesn't have any reasoning because of the paper they wrote
How's this guys ?
The paper is just an excuse to show shareholders why they don't need reasoning models
Imagine in future they launch 1 Tb model 😂

💯
looks like they might not even release it actually
And let me guess a few years later when everything has reasoning and maybe a new type of model
Apple is going to release their model with reasoning and act as if they invented it
It looks like a new Gemini update is coming
https://x.com/AiBattle_/status/1931675930918781358?t=cAGDX1VwdFp0XbkrcsogFg&s=19
Lol, Perplexity sadly failed to realize april fools posts on the internet.
(I clicked "report" a moment ago and made Perplexity dev team aware of this issue already. Hopefully gets fixed super soon.)
AI Model Testing
https://aider.chat/docs/leaderboards/
Is pplx still offering 12 months for samsung?
"Deep Think" on Gemini 2.5 Pro is already available via API to some early access users, it seems. The post also mentions that it will arrive at AI Studio soon!
nope
D:
omg the menu bar is finally transparent
Google: DolphinGemma
OpenAI: CatGPT
Apple: Raccoon Translation
Mm yes more competition to pressure openai to give higher rate limits

@vague pagoda @wraith crow
Why do they have so much promo code?
Will this be suspended later?
Thanks 🙌
Most welcome 🙌
🙏
I advise you to never buy from a third-party since abused codes/ fraudulent account will be rolled back to free tier.
Apple will do anything but allow macOS on iPad
That will kill mac lineup
not really when iPad with accessories cost as much as a real mac
why would that be a problem for Apple
Craig saying facts
Idk they just don't wanna provide power to ipad users

No DMs promoting your products, services, or servers. You may post your own project in #😎│cool-projects only if it adds value and follows the no-spam rule.

?
<@&1098365634498285709> I have Perplexity Pro and even when leaving the channel and joining from the Pro link I still don't have access
Lmfao
@serene sand
Just came here to say THANK YOU 🙏
<@&1098365634498285709> Hey, Can I get a little help with redeeming my code? It shows that I's been redeemed.. clearly not by me...
@serene sand
Sad that my pro is expiring soon since I got it through uber
Where is the code from?
From T mobile Magenta
Please contact their support explaining your issue. They should be able to help you with this issue
They've told me this is issue on your side unfortunatelly
That they're doing everything up to their standards, something like that
@wraith crow are you able to help me?
Please send a mail to support@perplexity.ai and explain your issue there they will be able to help you further with this.
What cartoon is this?
💬 119 🔁 0 💜 9.6K 👀 0
Telegram founder Pavel Durov has effectively been under house arrest in France since he was arrested there ten months ago. For the first time, he explains why.
@short narwhal
Yuuhh
How to do this?
@distant aurora how he do this ?
Get subscription on website
Is it free ?
No
Then you can't get it
I’m buy today the pro plan of perplexity
And this pro discord ?
You have pro subscription?
Yes
@serene sand
Lmao imagine if apple ai rolls out after gta 6😭
Genmoji💀
Live translation is useful.
Microsoft already has that but ofc apple loves to "invent" new "names"
??
Wait for a discord mod to give access to pro channel.
Alright
@halcyon coral you here?
Apple work out buddy real time pep talk.
in discord pro what we can get it ?
Dedicated pro channels for chatting and feedback.
Mean Its really helpful for us right ?
No. Don't worry, I also watched cartoons when I was younger.
문의한지 30시간이 넘었는데 아직 환불이 진행 안되었습니다 보통 이런경우가 많나요?
New free premium prompt: Immerse yourself in an advanced narrative with realism that will require the use of your deep reasoning, every decision you make, has an impact.
Download it for free from here in Spanish:
#1381764428982390917 message
considering the feature set, this is a terrible deal
I would have considered 5 dollars.
That extreme search ain't shit.
Worse than grok.
It has grok version 2 in free mode. It has gemini flash 2.5 in reasoning mode being the free plan.
Flash reasoning is available in Gemini without subscription
Use via gem.
Grok 2 is useless.
xd
Do you disagree?
Yes, version 3 is better. That's what flash is better on that platform on the free plan, but it has Sonnet 4 without reasoning.
@dry kestrel https://g.co/gemini/share/97ea087fa358
where is the real William and what have you done to him?
Does anyone know how to tweak the search behaviour in Space queries?
In the past it could search both Space files and web results for every query
Now it only does either one of them, and most commonly the files only
How do I modify this?
Prompt and system instructions tunning
If you mean manipulating the search system, it's likely to be difficult. I made a prompt which is posted here and a certain parte... Not found.
I'm using API for my n8n workflow. How to call the right model for Labs research?
reasoning_effort
enum<string>default:medium
Controls how much computational effort the AI dedicates to each query for deep research models. 'low' provides faster, simpler answers with reduced token usage, 'medium' offers a balanced approach, and 'high' delivers deeper, more thorough responses with increased token usage. This parameter directly impacts the amount of reasoning tokens consumed. WARNING: This parameter is ONLY applicable for sonar-deep-research.
interesting
I noticed that new 2.5 flash and pro models are extremely sycophantic
Damm first time seeing an admin
Hello kwak
afraid from pinging
Who know if I got banned 😭
@distant aurora
👀
icon reveal
To be released in 10 years
Why perplexity discards any images that I want it to create it says something went wrong
send your prompt
Try refreshing
Yeah it works after refreshing 10+ times
Also why can't I select an AI model when searching in my android app it's shows empty
Only pro users can
I am pro user
Click on search option
When you type prompt
On the first option called pro search you'll see best
Click on it
You'll see the model options
Poor guy is trying to make some money
anybody seen that builder.ai thing?
was crazy
They literally said "we will have our ai build your project for you" and they had indians behind the scene building it instead
Yeah got it thanks
Hence Indians are better than AI
Lol
Doesn't change that the product is not worth that much
i don't disagree. they are, and chinese and maybe some of the japanese dudes out there. but it's crazy how the company managed to pull this out
FireShip did that? i gotta see it lol
Yep
Apple core:
Fails to deliver apple intelligence
Tries to prove that ai copies patterns instead of reasoning when learning patterns is the core of reasoning
Read. The last line literally says grok 3 can be seriously dumb sometimes.
They got rid of grok 2 I believe.
actually
Cuz selecting grok 2 gets you rate limited just like grok 3.
So I don't see any actual benefit of it anymore.
yes, william preferred grok 2 over grok 3
Would have been cool if grok 2 didn't get rate limited at like 15 messages.
.
Didn't get the context properly, my bad. But you get my point.
Only reason for using grok 2 was no stupid rate limits.
I could actually have a meaningful conversation.
@cobalt vale https://g.co/gemini/share/a40b43334c3b
Even ai studio looks up 8s gen 3 instead of 8s gen 4.
😭
@cobalt vale bro idc what anyone says, google benchmarks must be fake.
How they made it so far in the leaderboard is beyond me.
well 2.5 Flash is a tier below even o4-mini
I used pro in ai studio.
.
You can try the same question.
Compare 8s gen 4 to 8 gen 3. Why do you think 8s gen 4 heats and throttles so much despite newer and better cores at lower clock speeds while 8 gen 3 doesn't?
You need to enable search grounding manually in AI Studio
meaningful conversation with a dumb model?
Low iq conversation, yes. Like rants of life.
I already did that.
Do you not see the citations in the first screenshot?
💀
no
Tfym no
no I didn't see 😭
😭
@gloomy finch
Hmmm good point
We thought u killed butcher
😂😂
😂🫠🫠
Grok is still best for ranting 🫠
@distant aurora did o4-mini-high screw up
Big time.
Bro
The amount of hallucination in thinking.
Hello
True
Idk where u live but this north India
Nope. Come rescue me.
Not sure u wanna move to Canada
Mine is worse🫠
Somewhere in the states maybe
It can go upto 47-50⁰C
Depending on the area.
U wouldn’t do well in 0 degrees let alone -20
++
Freezing to death >> sweating to death.
Atleast my pillow will be cold on both sides instead of sweaty.
But can only take so much stuff off in the heat
The issue isn’t the cold it self tho
Cold is easy to get past, it’s everything else that comes with the cold that I don’t think you’d do well in
@gloomy finch bro istg, I was sleeping in my underwear yesterday and still started sweating. I turned on the ac cuz the bed was starting to get wet. Fan alone isn't enough.
that prompt really takes time
Understandable bro🥲🥲 I have a thermal power plant nearby 🫠🫠
Does it radiate heat??
I live right next to the ocean so
Fan just gives hot air🥲
But even dimensity 9300 doesn't heat or throttle so much
Still helps with sweat
yesterday was 🔥 though
Yeah there's boiling water steam released through it 🫠
Damnn too sweaty experience
Yea it’s humid
Yup have gone through during my college, had ocean near 6hr journey, so living nearby ocean is far too humid 😔
Great man
We have dehumidifiers in basically every room in my house so
Don’t really feel it
That's better, or else living with humidity and all sweaty sweaty ruins mood too right from start of day
How?
but my offer did not count ...
Didn't that offer ended like on May 31st
The struggle with laundry.
Yeah mine shows the same but the deadline mentioned for additional months through referral was till may31st afaik
Should be api
Either they haven’t updated site
They had to do that after all huh
hopefully damn anthropic does the same
Or
Onggg
We’re giving 200 developers free API credits—worth 1M input tokens for OpenAI o3. Fill out this form: https://t.co/oevDG9rizT.
All doled out! To the 200: check your dashboard for the o3 credits: https://t.co/lKussw7g0Q. (And thank you for having notifications on!)
Gone
😦
so it was an error? ffs
Wdym
Nah should not be
@OpenAIDevs google gives that to everyone for free for each prompt forever
Lmao
We both have different priorities than to check the o3 pricing 😂😂😂lmao
Couldn't care less lol
True
That would need some good competitor which is almost impossible as of current scenario
Sadly 🫠
Of the FTC puts own it's big boy pants and gets to work
I forgot the worst of them all
@cobalt vale I want whatever it is smoking.
Can I not select different img gen models inside the perplexity android app ?
no
So I'll need to use a browser for it ?
Yes
you there?
For now
What I know from experience is 8 gen 4 is not that power efficient
dm if you got time
Thank you for taking action to address this injustice.
butcher is the expert there
100% more efficient than my tensor G2 though
humans and AI are not much different after all
So when do they usually respond. like never?
my expierince is never
or after 2-3 weeks
the worst support experience on earth
what's so hard on assigning 5 people to manage this
everytime you make a support thicket sam will respond he will transfer you
i hate sam
i wish perplexity increased their support
seems like they don't even have any
i mean
the whole thing that i want it just replace my email
and move my sub to another email
not so hard
they make it hell
you will get reply i guess in 2-3 days
apple transperancy 
💬 2.6K 🔁 0 👀 570.2K
The top contributor to Luigi Mangione’s legal defense fund cited President Donald Trump’s recent executive order on the death penalty as their motivation for donating.
Trump’s order, signed on January 20, directs the attorney general to seek the death penalty “for all crimes of a severity demanding its use.”
Mangione, 26, faces the possibility of a federal death sentence if convicted of murder.
An anonymous donor contributed $11,000 to his defense fund, making it the largest donation listed on the GiveSendGo fundraiser.
The donor expressed concerns about the death penalty becoming “politicized” and feared that Trump’s order would exacerbate the issue.
“I am also concerned about the severe infringement on the defendant’s right to a fair trial and due process, particularly since his juries are effectively pre-tainted,” they stated.
Video: An image of Luigi Mangione is being projected onto a building between Lafayette…
I feel bad for this guy.
dm
@atomic shell daylight robbery https://cunnyx.com/cozy_OSS/status/1931712401562612179
Goes to show how weak Switch 2's CPU is.
︀︀Even a Pixel 7a (Tensor G2) can beat the Switch 2 in CPU intensive workloads.
︀︀x.com/tofu_mkw/status/1931360924784627927
Quoting とうふ tofoooo! (@tofu_mkw)
︀
本気で比較してみた
︀︀全て同じ設定で実行
︀︀#Switch2 #マイクラ
💬 1 🔁 3 ❤️ 26 👁️ 2.2K
If slow was a word.
let me watch
my Tensor G2 is so back
pixel phones arr good for gaming + camera but are expensive
Pretty sure a cheaper Xiaomi/realme can run it too.
