#💬│general
1 messages · Page 84 of 1
im triggered haha
It's time to use our brains once again. sry for yall
Same
I don't know what's going on, but I had a 1 year pro subscription, and now it's saying I don't have it, is this normal?
Those are chatgpt users immigrating to preplexity
goes outside touches grass
i have it now as a gif
Damn bro that sucks, but Emory is top tier so you're still good
is perplexity having issues?
How this status thing work? is it based on reports?
context bro everytone says its down atm
it's kind of spasming on my login
Its down
did you guys know grass was green? shieeeeeet
it's down, ok, thanks 🙂 have a good day!
im tripping
The site is down, but the app works
i really believed i was good for it man
Happening to everyone, well almost everyone. Spotty service right now
nooooooo now the site is down to
I only have one friend who goes there and he only got in cuz his brother and mother went there
I was in the middle of a project with my team and out of nowhere I no longer had access to the project, I ended up coming here, since my subscription and everyone else's had already been paid.
awe
No biggie, you got the others
hopefully
I smell a data breach.
The website is telling me to pay again, so it scared me.
im just pissed i have to wait 4 more months for my decision
I'm wondering if people are hearing it is down, then going to check, which is just making the problem worse xD
georgia tech on top
is*
way better
booo
omfg i was so scared-- was troubleshooting the crap out of my browser cuz i was so confused lol
It's good to have friends. I'll do my best to create the assignment. Thank you bro.
is back 🙏
Same for me
Wow yeah
I am graduating from Georgia tech this Saturday💀
put that on your mother
bait?
working for me
I will, she a hoe but it still counts right?
AI has made us all lazy look at us 😨
oh thank gods lmao-- i was scared it was one of the extensions i recently downloaded-- have too many to know which one it woulda been
same its workign
oh oops.
how u gonna call ur mom a hoe 🤯
this is whats fascinating about the internet
i just met all of you and we all have crazy connections
It's back hell yeah
its back up again
QUESTIONNNN is the cybersecurity masters any good?
OK it's back
suddenlycaralho?
It’ back
congrats everyone
yo anyone know how to bypass turnitin ai detection?
bro easy work cmon nowe
Cheers
Not familiar with it but I’ve heard it’s hard and jobs are good paying
WE BACCKKKKKKK
how
I’m in ece
Really, but my "helper" is making sure the bills are paid, so I'm calmly accepting that.
@gentle gale
o server nao curte quando vc poe tudo em caps :P
todo canto tem brasileiro, pqp
pegou com a vivo também?
plz mods dont report me
me too 😭
pra caralho
promoção pica
Olha o privado ae
- Grok 2 it has the most creative imo. 2. you have to have a creative prompt that has personal details. 3. then IN YOUR OWN WORDS wrirte down each sentence of the output again in your own words and 4. run it throgh a free ai checker
nahh thats another level of lazy
and probably not goona find one for free
plus im sleepy
someone find me a bypasser tho
You can always try to throw it into google translate from native language to new language, then from the new language back into native language. Usually there is some restructuring and rewording that happens.
my rage at perplexity after it comes back online: sleep
like ill deadass start sucking ya toes if you do
i already use quill bot
but does it do a better job than that
Well... don't know what to tell you. Just gonna have to do it by hand like a peasant I guess xD

does someone know if I will get charged if I add my credit card into my (gifted) pro plan?
its time to bust out the good ol' pen and paper!
everything correct except peasant
call me young elon with all these diamonds dancing on my neck
🥱
You was literally about to suck toes, why not just pay somone then xD
cos sucking toes is more worth than money in my opininon
you woulnd't want a white toe latina to give the sloppy toppy on your toes?
What's wrong with ppl now? Is it working?
Yep it is back online
No, it's still unavailable
Which region you at? I suspect it is going down in different regions. I'm in vegas. it was down, but it is back up now.
Xfinity just gave a one year membership to its subscribers . That might be the reason .
My region is Indonesia
It was down for like 40 minutes but then came back up. Possibly the same for you
Is Chatgpt working in your region? Here it does not work.
I don't use chatgpt. But I heard it was down for hours
Just chatgpt itself, I have an API that is directly linked to OpenAI and is working normally.
OK
I wonder if PPL has a deep technological moat. Why can't big teches copy it and overpass it?
good night
Is it the best tho? They only compared it to their own
That is a great question. In my opinion, of course they can, but Perplexity already built lots of partnerships with information providers, it has lots of models available within their platform, they offer unlimited queries, and at least in my opinion the answers are way more reliable than ChatGPT and other providers. I think they have a very complex vertical Agentic approach and a good RAG implementation that makes them stand out from other players.
But of course, with the amount of money OpenAI and Google have, they can do barely anything, I am just not sure if that would be a good choice for them, because they are AI model providers which puts them on another level of competition and they have other opportunities of growth that Perplexity doesn't have.
Maybe your question is similar to asking if a textile production company can copy a t-shirt company. I mean, they can, but for competition and market reasons sometimes it is not reasonable.
that leaderboard is from google ofcourse they will gemini on top
its a marketing trap to hook you
because they know you want the best
or whats on top
aladin
Haha
Comparison with your own model is definitely better than comparing yourself to others. The best approach is to try it out because even if it achieves the best results, not everyone will like it or find it suitable for their style or the purpose they need it for
seek progress not perfection judge yourself by your own virtue not by the fleeting shadows of others
brought to you by: #1218983754165387436 message
Where do I go with auth issues? (I switched my Google workspace primary domain, which changed the associated email address). My paid pro account is under the old domain / email address. When I try to log in it uses the new one, which then creates a new account under the new domain / email.
Who should I talk to? Thx!
is there a difference in amount of pro searches or messages with gpt 4o or claude 3.5 sonnet, in enterprise pro, im basically asking if enterprise pro makes the searching experience better or customisation or is it just for security?
talk to perplexity.ai with your problems
research your solutions
Hey @silent pier!
You can find all the benefits here:
https://www.perplexity.ai/hub/faq/pro-enterprise-faq
Anybody know if/how I can delete a previous image I uploaded in a thread? The conversation has shifted somewhat but the AI is still trying to answer questions in the context of that image
Aha! Figured it out 👍
Can anyone give me an idea of how to use Plex for coding
guys check out NotebookLM its free and will help you get all the degrees in college all the way to PHD.
its a product by Google officially...
genemi is bad asf
true (||I haven't used it though, but it's probably like Bing's AI||)
Gemini 1.5 Pro Deep Search is very interesting - for intensive research, it arguably outperforms pplx
tbf, the response in the screenshot took a few mins, whereas the pplx response was prob <10 secs. but i'm happy to wait when the objective is to be as thorough as possible and accurate (e.g. when researching)
Why is my Perplexity searching in English even my Language is German
for intensive research u need premium?
Did it go down?
Does Perplexity just really stink either context?
Now the reasoning model o1-mini is missing
I have been trying to engage with Perplexity Partnerships but I have not succeeded in getting a response from my emails. Is their a contact email/number I can use ? We are in the UK any ideas
why is the perplexity macos app so bad? (below expectations)
first beta had a lot of simple bugs (like task bar icon in weird place, would consume 10% cpu on idle, ...)
even after weeks, basics things are still not working. just now task bar icon disappeared, but app was still running.
and the most basic thing: when you start the app for the first time, WHY DOESN'T IT FOCUS THE TEXT INPUT?
need to click one more time until i can enter my query. come on, this is not the user experience we expect from perplexity in 2024...
(plus: can't click the taskbar icon to quit the app...)
Heeyyyy
lol
talk to @signal hamlet
money talks
i cant seem to see message limits could you help me out?
Quick question. I'm wondering why my supply isn't being delivered.
There is no real limitation for normal use. You can send many hundreds of messages per 24 hours.
An abuse protection system checks whether perplexity is being used automatically against the terms and would only intervene in such cases.
Hey @warm mulch!
Please check https://help.perplexity.supply/en/articles/10062097-why-has-my-order-not-arrived-yet
this isn't help..
even the link to this button is broken.
Did you already check the order status? What does it say?
honest opinion perplexity or chatgpt with web??
recently i have been getting issues in perplexity.
poor image generation, poor attachment issues
on the other hand chatgpt plus offering voice + sora + models like o1
what do you suggest?
Chatgpt better for most of the things
Imho
I wouldn't say Perplexity is lacking tho, it's just some annoying design choices and their system context along with how they handle things
Hello, My opinion is this.
Chatgpt is great, but only has specific models to openai.
Perplexity you can choose anthropic, openai etc..
And perplexity knows when to use reasoning models for when the task needs it.
Its your preference on whether you decide based on what you practically use and find value.
I would spend the money on one that you find yourself actually using daily or what ever task.
I just buy both for their own use cases.
my carrier gave me free preplexity pro for a year
RIP o1-mini 🤡
@halcyon coral is there a difference in enterprise pro and pro other than security e.g. Amount of pro searches or broader customisation options, the article about entpro only talks about security
is it gone?
+1 where’s o1? i need thinking model.
@caviterginsoy o1 is unnecessary (at least for now). Reasoning in pro triggers when the queries are complex automatically.
ty 🙏🏼
Images uploaded to perplexity are public on cloudinary and remain even after being removed.
I am listing this as a bug because I hope it is. When in trying to remove attached images, I followed the link to cloudinary in a private browser. Still there. Did some testing. Attachments of images at least (I didn’t try text uploads) are public and remain even when they are deleted in the perplexity space.
Thoughts?
https://www.perplexity.ai/hub/technical-faq/what-is-the-difference-between-typing-a-question-and-uploading-a-file I would assume they are deleted after 30 days
but would probably be smart to get it setup so it occurs immediately if a user manually deletes for sure
Keeping them for 30 days for future relevance is entirely diferent from making them public instead of private
when will be fixed context problem on perplexity? i’t possible to increase context, when i ask to generate a script with modification, perplexity lose the context of previous code generated missing implementation
Keeping them for 30 days for future relevance is entirely diferent from making them public instead of private
i added some fandom wiki links to my space and yet its not able to pull from them at all
like did i put the links in wrong somehow?
Hi, does anyone have any idea on this issue?
I'm in urgent🙏🏾 🙏🏾🙏🏾🙏🏾
https://discord.com/channels/1047197230748151888/1317021623894937610
ask perplexity.ai
what are you doing
wha tdo you think perplexity.ai is for
to wipe your a**?
like im just getting this anytime i try to create a story
I do love asking perplex to fix itself
yeah tell it to f itself
it will keep training that
upload that to perplexity.ai
then perplexity.ai will upgrade to Orion
i missed the public part of it
I'm experiencing similar problem with files
Edited the post to delete the attachment and instead linked it via the.... cloudinary.com link 😄 we've come full circle
at least it appears they arent easily discoverable without a direct link
Hey, can anyone with closed beta access give me the expected format of response when the return_related_questions parameter flag is passed in a chat completions call? 
is the prompt that the perplexity team uses to create it's branding images public? i really love the style and would like to create my own images.
We just opened up applications for our 2025 Campus Strategist Program!
🌐 Now open to students at any university globally.
Have there been model or performance changes/updates in the last 4ish weeks ? Anyone else noticed perplexity default not working as well?
eg not keeping track of the conversation or making a ton of errors /hallucinations? For example , I told it that Phoenix az time currently wss 11:32pm and it responded very confidently something lile 'No youre mistaken; the current time in Phoenix, Arizona is 10:32pm as its in MST. Arizona doesn't observe daylight savings.. etc'
It had me very confused and checking if somehow all my clocks were off.
i've noticed it'll make tons of mistakes with anime series i'm making stories about
like precure teams having wrong names, member count or attacks
how does perplexity download stuff in background. Is this mimicking it or is it real? Can you receive the actual output of the code?
real. in the ||aws|| cloud instances
i think i missed any ways to get perplexity pro for free
You carrier? What's that
A --> B(🥈 ChatGPT-4o-latest)
A --> C(🥉 Gemini-Exp-1121)
B --> D(🤼 Gemini-2.0-Flash-Exp)
C --> D
D --> E(🧙♂️ Gemini-Exp-1114)
E --> F(🤖 ChatGPT-4o-latest)
F --> G(🤓 o1-preview)
G --> H(💻 ChatGPT-4o-latest)
H --> I(👾 o1-mini)
I --> J(🦾 Gemini-1.5-Pro-002)
J --> K(👨💻 Gemini-1.5-Pro-Exp-0827)```
Whay you mean?
What are you gonna do when? Gemini-exp-1206 is no longer top number one
change my name to what will become #1
Real ok 🫡
Hello guys
What's up
Very nice ai model but doesn't seem to support and create pictures
today you can just google free text to image
hey man it says "I cannot directly process or extract text from the attached PDF file, as I do not have the capability to access or analyze external files. However, if you want to extract specific content (e.g., dark black text) from a PDF document, you can use tools like Adobe Acrobat or specialized PDF text extraction software."
i thought it is capable of analyzing pdf files
it says a ''pro discord'' but this just took me to the normal discord???
let me rejoin?
i saw a offer for free month of pro for students and i signed up using my college mail but i am in free subscription, what to do?
ok
Omg same
@south narwhal you have to leave and rejoin with the ''pro discord'' link you get in your perplexity pro webpage and then you geta pro role in which you can access certain new channels
now I understand
Interesting. I'm giving that a shot!
Can someone help me with how to get an image generated? I read the #guide and I tried it on the mobile and desktop site. Although I am a pro user I never get the option to create an image.
For example I forked this prompt that someone posted in the guide, but I get no option here to create an image.
I got it. Thanks a bunch!
no problem!
Hello lemon how are you doing?
Im good, thanks
Got any plans for Christmas?
Not much
No presents ??🎁
Nahhh
How old are you
17
Damn my parents stopped giving me gifts at 15 for Christmas 🙂↕️🙂↕️
ohh
lol
Well, it was nice talking to you
I’ll see you later lemon
is there a way to turn off shopping suggestions?
no
Sonnet is making me crazy, it is so dumb I just can't take it anymore, it won't obey basic instruction about message length
It clearly understand the instruction, and start its message with "I will obey the instruction and generate an answer at least 1000 words long"
Then it proceed to generate 3 miserable line and add "I'm sorry it seem I did a mistake, would you like me to proceed ?"
I reply, "yes proceed"
It say: "I am now going to generate the answer using at least 1000 words, would you like me to proceed ?"
"YES PROCEED and do not ask for an more confirmation before proceeding"
"I understand I must generate an answer of at least 1000 words without asking for any confirmation before proceeding, would you like me to proceed ?"
"Why the f*ck are you still asking for confirmation, I already told you to proceed and generate the answer, so do it !"
"I am sorry, I understand your frustration, I will generate the answer you are asking for..... WOULD YOU LIKE ME TO PROCEED ?"
So I made the mistake of subscribing and it works, but now I'm wondering how the guy could sell me this...?
so you.com is overall better than perplexity?? Plus unlimited use of gpt 4o is not possible for 25 a month...
you.com sounds to good to be true...
the risk of pro subscription being terminated
Yes, I would be disappointed to have paid for nothing. But it is rather at the level of the users' accounts of perplexity. How could the guy activate the pro mode on my account? Either it is a security breach, or it is a shared system...
most likely tos violation by abusing the term applied discount coupon
||or credit card fraud||
Ok, merci!
So I can understand getting scammed by paying someone who has nothing to do with Perplexity, but I don't understand why the coupon was activated on my account... The coupon seems valid in the eyes of Perplexity...
Thanks, we'll see what happens...
@silent pier I haven't used pro searches much yet. And I don't see where the counter of pro searches is.
Is it possible to disable this "shopping experience"? A primary use-case for me is doing product research and the current ux with these "related products" is completely distracting and ruins the content focused usage of this tool for me.
why does it say default model ''claude 2??'

Anyone knows how to get in touch with MODS? I have something to ask
that is just an old backend string. it will use claude 3.5
most the dm's are open you can just start a chat
Oh cool! Thanks a lot :))
hey
just a quick question: is there way to get a perplexity bot into another Discord server? Has anyone thought about making a Discord app/activity for it?
The attached text file won't load
?
it started as one
they got rid of it
Hey, it's not allowing me to upload text files at the moment. It uploads but it doesn't register that anything is in the text file
"I apologize, but I'm still unable to access the content you're trying to share. The search results indicate that the attached file named "paste.txt" failed to load, and no content was found"
how big is the content ?
Have tried it with several different files, ranging from 500 words to 4000
or however much the minimum is roughly
I got the same issue
This needs to be fixed! https://discord.com/channels/1047197230748151888/1307056160515424339
same issue. its been happening for about 30 minutes
is it working for anybody
hi
does o1 already being added to ppx? it's has been a long time since the last time i use ppx.
We don’t have any of the o1 models, we use to have o1 mini, but they removed it from the front end, I think you still get 10 uses per day via complexity. I don’t have much faith we will get full o1
In my experience perplexity is pretty bad at remembering chat context, so i think it's better you split ur prompt by numbering them instead of spreading them to different messages
never
whatsoup
ask perplexity.ai
View From The Top with Aravind Srinivas, Cofounder and CEO of Perplexity
Aravind Srinivas discusses the vision behind Perplexity AI, the challenges of building cutting-edge technology, and his approach to navigating complexity in today’s tech landscape in Stanford's View From The Top series. ☕
nice
Since nothing is getting answered in discord, how do I even get support as a pro user
I can really sense the potential of gemini one of these days sir
yes sir good to see you sir. whats ur favorite LLM today sir
Gemini Sir. I see their stream live good sir and context sir. Chatgpt o1 also sir. They are trying to innovate sir to improvement sir. Unlike here sir. 😂😂😂
I can say intelligence improvement sir Gemini and chatgpt are leading this AI competition sir. I appreciate that from google sir.
u can go to twitter and use grok 2 for free sir
is Grok 2 good sir? never tried elon musk sir.
i didn't realise that
it's not a great model imo... but yeahdidn't realise it was freely availble now
cot is not solution i think so
according to some chart I think it was on xAi site, it outperforms Opus sir.
illya said that
@cursive jacinth
Hi
Hello
Can I transfer my subscription to another account?
I can't access my Google account right now
ask support email / Pro support
I'm sorry for the question but can you direct me on how to reach support?
Thank you guys
#💬│general I accidentally deleted all my threads instead of the last thread. How to recover it? Anyone please.
I'm also facing the same issue.. is this only with sonnet? Or same with other models?
Only sonnet, and only sonnet on perplexity, works perfectly fine with the official version
:(
how did you get the free year ?
Goood night
-# why is my life not good damn
I definitely like the unlimited opus ngl
I somehow have a premium claude subscription that’s free for me, but the limit on opus is annoying ngl
You.com student discount was useful for unlimited opus tbh
Hi, Is Chat GPT o1 available through Perplexity (pro subscription) ? Because I can not see in my available list of models in Perplexity options.
Hey @exotic shard!
Only the models that are displayed in the model list are available.
Is Chat GPT o1 going to be added to the list of available models ?
i guess back to ChatGPT plus then...
Weird bug using firefox, if someone can replicate it and tell me it they have the same
I was role-playing and sent this prompt :
"I walk in the cave, disappearing inthe dark"
My keyboard is a bit broken and often when I press the space bar, it either don't register anything, or register space 2 or 3 times
So since it did not register the space between "in" and "the dark", I edited the prompt to add the space
But when adding the space, I felt some resistance in my space bar, and this is generally the moment it register multiple space, but yet it didn't, I only got 1 space, and I found this a bit suspicious, so I pressed space again and this time it didn't register, tried again and it didn't, and again, and again, but no matter how many time I was pressing space it did not add a single one, so I went to the search bar and tried to press space there and it worked fine, adding a space when I pressed space
So now I had this weird bug on perplexity where it wouldn't register more than 1 space between words when EDITING a prompt
So I sent the prompt to see if the multiple space I added would show up, but they didn't
It show "I walk in the cave, disappearing in the dark" when in reality it should show "I walk in the cave, disappearing in the dark"
So to be sure I click on edit, copy the line from perplexity and paste it in the search bar, and then it properly show the multiple spaces "in the dark" ...weird
Even weirder, I next ask perplexity to repeat my last prompt, and it repeat "I walk in the cave, disappearing in the dark" without the spaces
BUT when I ask it to repeat it taking into account all the spaces, it repeat "I walk in the cave, disappearing in the dark" BUT it say :
"...with a notable space between "in" and "the dark" which may emphasize the gradual transition into complete darkness."
So it doesn't show the spaces when edit the prompt or ask it to repeat it, but it knows they are here
Oh and if I write something in the text box at the bottom, it shows all the space I add but when I send the message it turn all the space into a single one, but again if I click on edit and copy paste the text elsewhere, the spaces appears
if cost is the issue why doesnt perplexity just allow us to pay per use sort of like api specifcally for the o1 models? perplexitys searching algorithm is top tier so it would pair well with o1 if one needs sources or write essays
Why is the voice recognition feature on Preplexity so bad? Are there any plans to improve it?
its not sir
Will Pro users be given the option to use any of the newly released OpenAI models, or Gemini 2.0 that came out over the last week?
If not, what iteration is the current GPT-4o, available under settings?
Typically, Perplexity is very quick to adopt new models as they are released, but for the first time since I became a Pro user (early adopter) I' feel like I'm uses Gen-1 AI models right now, with the same old rote replies and generic overall feel with hallucinations, etc.
Having a reasoning model, that checks its own "facts" would not only be hugely beneficial, but it seems like the literal one thing that a platform like Perplexity has been waiting for since Day-1 (a reasoning model on a reasoning platform that aggregates data to provide accurate answers).
Please let me know
https://bunkerwars-meta.k8s.bunkerwars.game/f/ZFlSEtse
i found this using perplexity
and very fun
Anybody know why Perplexity Pro keeps giving me the wrong answers when I ask for current NFL game Scores, but with regular Perplexity without Pro turned on it gives the correct game scores?
Aye mates, anyone know if the windows app you can find over GitHub and download over three corners via sourceforge is legit?
It likely isn’t
Just use the website to stay safe
thx
In Perplexity, I’d just like proper image generation and reliable file handling – that would be enough for me. 😉 Perplexity has a big advantage because of Sonnet. If GPT-4o worked as well as Sonnet, I definitely wouldn’t be using Perplexity. Search is also much faster in ChatGPT and perfectly meets my needs.
Search in Perplexity is slow and I don’t like it. UI / UX is also terrible 😅 Claude 3.5 Sonnet is great 🥰
proper image generation and reliable file handling
very true
image gen is so eh on pplx
almost just a gimmick
i don't know anymore man, ppx has been feels like downgrade from last few month.
I hasn't use it anymore in these past few month and went for deepseek or gpt for anything even when i already sub to ppx for 1 year haha.
also it's puzzle me why don't they try to work with Qwen or Deepseek if the cost is the problem with reasoning model, maybe because they base on US so they can't do deals with them? i am not sure.
I see your point. I use Perplexity only when I need Sonnet. For everything else I use other services 😅
QwQ is reasoning model where for math it's 9/10 of o1 but for other thing it's mid than o1 but it's litteraly small fraction of the cost compare with o1, if it about cost i don't think there will be a problem.
$200
perplexity is the goated Ai
i need support about perplexity supply
Is this Sam actually a chatbot? It just repeats the same thing and doesn't help at all.
Why sell something if you're not going to ship it?
its a bot
message alex directly
Can we use elevenlabs for custom our voice reading ?
ask perplexity.ai
read it here this is youtube video to text with source video https://www.copilotforyoutube.com/search/build-anything-with-perplexity-heres-how-Jz-PnGoASvLhrH-frWFgMO
this. a legit tool checkout the features: https://www.perplexity.ai/page/ai-copilot-for-youtube-k3w9dMgYTqOOH9d.52mNvQ
Do you want to join my team? Apply here: https://forms.gle/2iz4xmFvDCGnj2iZA
If you're serious about AI, and want access to my code, click here: https://www.skool.com/new-society
Get 50% Off Perplexity: https://perplexity.ai/pro?referral_code=141ZIRU5
Follow me on Twitter - https://x.com/DavidOndrej1
Please Subscribe.
With Perplexity and VS...
This video explains how to create and use the Perplexity AI chatbot’s collections as personal agents. You’ll learn how to use and reuse these agents to help you save time and improve your prompts’ effectiveness over time!
Of course, it also comes with a loving spoonful of the “so what” baked in 🍞
Leave questions in the comments below!
📺 Links...
Appreciate that sir!
hello
Anyone tried chatllm or Felo and perplexity. Which are you preferring?
Hiii,
Can I change email in perplexity ??
Well yall ensured I won't renew my pro. Should have stayed out of politics and not donated a million to a felon.
Hey @solid pumice! Please take a look: https://discord.com/channels/1047197230748151888/1225557012532695060
Begins at 10am PT
Join Kevin Weil, Adam Fry and Cristina Scheau to hear about new releases and watch live demos.
shots fired
hey folks - i am looking to have a list of daily updates on certain keywords with a brief summary from the news. how do i best set it up
any pointers would be awesomely help, thanks a ton
You would have to create your own code for this and use api
Or you can just make a prompt, and ask perplexity daily about this key words all in one.
Jm
Jm what do you say?
Hola
Hey everyone, does anyone know why my API call isn't using search? How do I force search within the API call?
"You are a professional SEO expert. Your task is to create a product description for the following item. "
"Avoid using terms such as 'mitigate', 'treat', 'cure', 'prevent', or any health-related claims.\n\n"
f"Product Title: {product_title}\n"
f"Store Type: {store_type}\n"
"Instructions:\n"
"1) Using your search powers, search for this product on the internet.\n"
"2) Analyze the highest-ranking website descriptions.\n"
"3) Create a product description following this structure:\n"
"<h2>Product Name</h2>\n"
"<p>A few paragraph summary highlighting the product and its primary features.</p>\n"
"<h2>Key Features</h2>\n"
"<ul>\n"
" <li>Feature 1: Description</li>\n"
" <li>Feature 2: Description</li>\n"
" <li>Feature 3: Description</li>\n"
"</ul>\n"
"<h2>Usage Instructions</h2>\n"
"<p>Provide clear instructions on how to use the product</p>\n"
"4) Ensure the description is fully optimized for SEO.\n"
"5) Include metadata such as a page title and meta description.\n"
"6) Output only the HTML content.\n"
"7) Do not use markdown formatting, code fences, or backticks.\n"
" For example, do not include anything like:\n"
" ``html\n"
" or\n"
" ``\n"
" These must not appear in the final output.\n"
"8) Provide only the final raw HTML. No code blocks, no markdown."```
This should trigger it, but my API DASHBOARD is showing 0 search queries
headers = {
'accept': 'application/json',
'content-type': 'application/json',
'Authorization': f'Bearer {perplexity_api_key}'
}
data = {
"model": "llama-3.1-sonar-large-128k-online",
"messages": [
{
"role": "system",
"content": "Be precise and concise."
},
{
"role": "user",
"content": prompt
}
]
}```
Interesting how OpenAI seems to adopt every great feature Perplexity introduces. Is it innovation or just playing catch-up?
I'm referring to Projects and GPT Search, where many features or options seem to have been copied from Perplexity.
Not really
Projects was first done by claude (but on the team plan only??)
At this point everyone is copying each other lol
And just renaming it
Anyone using perplexity for coding ? What type of "active element" should I use ?
Am I able to actually provide links or did that feature get taken out?
Hey
Just a question, what make Perplexity more unique than other services? I just want decide whether I should to Perplexity or not...
what is with the model silently being switched to a dumber model? The quality seems to drop massively and is clear that they are using a dumber faster model but dont tell the user.
money
Nice, now we can give subscriptions. Who want to test that and give me 1 month? 
I wouldnt mind if they chose to just throttle me or tell me to use a cheaper model instead but doing a switch like this is pretty irritating, especially since Im paying for it
you're paying?
which model are you talking about then
opus? sonnet versions? gpt4 vs 4o?
sonnet. It is clear when it switches
On perplexity?
yep
oh damn there's poop sticker in the announcement. oh well
When does it happen and how do you notice it?
ill try that
idk seems random, I can only tell by the output but it is clear because the responses become lightning fast and pretty generic
😭
it seems to work with complexity.
I think the perplexity UI is silently switching to haiku cause screw me I guess. Technically its still 3.5
nvm its still terrible
Join my AI Academy - https://www.skool.com/postagiprepardness
🐤 Follow Me on Twitter https://twitter.com/TheAiGrid
🌐 Checkout My website - https://theaigrid.com/
Links From Todays Video:
https://deepmind.google/technologies/veo/veo-2/
Welcome to my channel where i bring you the latest breakthroughs in AI. From deep learning to robotics, i co...
i noticed as well
it says im using claude 3.5, but when i use free claude from anthropic's website the responses are astronomically better -- seems like the changes happened when they silently removed the option to use no sources in spaces
correction: i think i was just being stupid. i probably hit my 300 pro uses without knowing or something. it's working fine now.
@signal hamlet hi can you read my dm?
hii
It’s perplexity pro good?
I got a free 1 year plan.
what are ur expectations that equvalents to good
Hello @everyone , I have pro subscription , Can you any one guide me where i can get Perplexity API token ?
is there any mod can help me?
it is illegal for him to ignore your dm
just be patient you will see his face in your dm
morning
what is api
please have a look here https://docs.perplexity.ai/guides/getting-started
Hi, what are the differences between the models? Do you get different answers when using gpt versus claude?
Hi
The way i see openai innovates and now their websearch sir and to the current state of how perplexity is stagnant to improvements really bothers me sir and this is a threat sir. I dont know sir. Just an opinion sir.
Hello gang
yes sir
Thank you so much for the response . Let me explaore the link 🙂
Even though we have pro user account , For API again we need to register is it ?
The perplexity website and the api are two different products. So you'll have to add a payment method there as well before you can use it
God calls you to repentance. We sinners deserve hell for breaking God’s Law; our own conscience condemns us.
That is why Jesus Christ (God in the flesh) died on the cross. The Law was broken, but Christ paid the penalty. He suffered in our place, as a Substitute, and rose again after three days, conquering death. Now He is seated at the right hand of the Father.
Repent and place your faith in the Person and work of Christ, and you will be saved from the wrath of God.
“For by grace you have been saved through faith, and that not of yourselves; it is the gift of God, not of works, lest anyone should boast.” (Ephesians 2:8-9)
evangelism spam?
never seen that before - kinda refreshing ig (about as credible as most crypto currencies)
for those who want to use referral code, you can use mine
https://perplexity.ai/pro?referral_code=P089QUFO
it still doesn't have video-to-text and does not allow you to chat with video. Read it here https://www.copilotforyoutube.com/search/search12-days-of-openai-day-8-rQQJ3bPMn1WyaUvBqMUw9F app explainer https://www.perplexity.ai/page/ai-copilot-for-youtube-k3w9dMgYTqOOH9d.52mNvQ
Transforming Contact Centers with GPT-4o Multi-Agent Crews and Human-in-the-Loop: Building agents with OpenAI o1 and GPT-4o for automation, quality assurance, and human-in-the-loop solutions.
Presenter: Maik Hummel, Principal AI Evangelist, Parloa
How to find support?
Hello, I’ve just started using perplexity AI yesterday as a friend of mine gave me his code for 1 year of pro for free and it is amazing. Way better for research than I expected and way better at getting the responses I want from an AI than I have been getting with chat GPT.
Does anyone know of a place I can check to know the difference of all AI models I can choose from and which model is better for what or is it just preference?
hello
It’s pretty much between Sonnet 3.5 and gpt-4o. I’d say sonnet is better generally, but it’s variable enough to be preference at this point.
Is there any plan to bring the complexity style features to the mac app?
So is there any word of Perplexity brining o1 as a core model that we can leverage for search?
I hope that Perplexity really steps up with the new models coming out. So much has been released by Google and OpenAI, and I love the pro search with Perplexity, but I want to see them add new features. It’s been quiet for a while with new features here.
perplexity has been useless for a fat minute
claude and gpt's own clients perform better in every way for me
incl search where perp takes 5 years to give me jack
while gpt now almost instantly does its search
Hi sir, Its ok sir, Perplexity was my first love sir. Actually I thought they are really promising at first sir, but i think they stop their vision nowadays sir cause of friday pizza sir.
Seems to me they are heading to a new direction sir. Not competing anymore sir. Let them do it sir. Probably we dont need their platform in a few months sir.
No offense sir. Dont mute me for this sir. Just saying sir and i am not even mad about this sir.
I was muted last month for saying so much facts sir.
If i am gone you know why sir.
I don’t won’t anything like this 😎 I use it for work, not to watch a snowfall effect
Are the Claude sonnet 3.5 temperatures all over the place the past week for anyone else? Keeps randomly going way too low.
Oh thank god it’s back to normal again.
hey, isn't o1 api has been release? will it be on perplexity?.
It is out, yep
dunno when perplexity will implement it or if they will
tried it out a ton on nano-gpt back when it first released. not very impressed.
slow, expensive, still dim-witted
Can perplexity team allow adding sources behind authentication.
Like if I pass cookies so that it can use the website content as source or any other better ways
put your idea into ideas channel
thanks
they never will, they need to get jeff bezos attention so jeff can invest in perplexity again
how to install perplexity on 3ds
So this new custom web source is this just us telling the search to only trust sources we tell it to trust basically
does it work now?
I would like to search the way the video searched using spaces, but unable to do so..
@signal hamlet
Are we able to gift a trial version to friends?
that’s not possible
you can see the options available here https://perplexity.supply/shop/perplexity-subscription
guys is the FLUX.1 in perplexity the FLUX.1 [pro] version ?
ask perplexity.ai
pay food for your family not for dissapointment
stop this
when is the windows app coming for peplexity?
[dev]
("Le chat" by Mistral uses the [pro] version for free if you want.)
Cannot show the video to perplexity, thus your suggestion is invalid
yes you can
allhumdulilah
believe you can
i have android and i can download perplexity ai in my android with no problem
I have it in my android already,However the issue is i cannot use the feature that was posted in announcement recently.
sorry
💅
🫂
fr
If it seems too expensive and working with Qwen/Deepseek isn’t possible, then maybe making a deal with Meta, which is still developing its reasoning model, could be a good option.
you.com is the biggest rip off site
stop supporting cheap 01's they its just the string name being changed from gpt-3
Guys I have a question, if I choose the ai model to be GPT 4o and then I use normal quick searches, will it still use GPT 4o and will there be any limit?
It gonna use different model to do the search then the result gonna be given to GPT4o where it being use as additional data to predict the next token.
But in short, if you have pro account it still gonna have the same limit that is 300x if you use pro search if you don't use pro search then there is no limit as i has try but the quality not as good as pro search.
Good!
So the search itself would use the default model whereas the final answer will be given to gpt4o and what I'll see is the combination of the two, something like that?
More like one model scower the internet then extract the data, summirize it then give it to the Gpt4o as prompt with your prompt
Oh yeah I mean it's still very good. I tried asking the model about itself it said its GPT 4 developed by OpenAI and with some enhancements by Perplexity, is it GPT 4o for sure?
And thank you for answering my question!
I guess the model doesn't fully know itself so I won't count on that
But I've heard in the past about people complaining that they felt the pro search with GPT 4o selected was a bit downgraded
Mostly it will be Gpt4o, i mean ppx could get in trouble if they lie.
Also if you ask model about it self and they answare right mostly it has been prompt in the background with that fact.
Because if you use model from their API without system prompt they could identified themself as other model haha.
In your testing, what would be the best model for doing researches with pro search?
Oh lol 😆
For me, i use claude 3.5 sonnet.
Also that downgraded feels mostly are because either the system prompt or even the search result messup the next token prediction of the model.
Note: it also possible that its because many people use the same API at the same time where if it using something like batching/caching it could make it worse.
I suggest you using space to improve the quality of the model with your own additional system prompt.
Oh could be that...
You know what made me pay for perplexity, I was just thinking to myself, maybe I should invest that in Perplexity since it has multiple model inside and it'd practically unlimited like you said if I don't use quick search, oh and the fact that it's the best tool for doing academic researches and stuff other than Gemini 1.5 Pro with Deep Research (which I've heard, can hallucinate a lot and cite wrong information)
I just hope perplexity would be able to keep up to google did not a long time ago with deep research... but I guess sometimes searching from too many sites can be a bad idea because some of them would have false information and so on...
From my experience when i don't use pro search and with my pro account i not yet hit limit doing the quick search, but you can test it out yourself to make sure of it.
Also almost every LLM will hallucinate even with additional factual infromation, it just gonna be less thanks to it.
Yeah but I guess Perplexity doesn't do that that often unless the topic you are asking about is very complex with a lot of details...
And the website it searched for had false information that perplexity thought were true
But it should be able to compare and see if it's just one or two website that were false (because all of the other websites says the opposite)
?
You know, sometime i feels hallucination is a feature rather than bug. I mean if we look at human... how do we come up with new thing? we imagine thing that doesn't exist yet where it could have been branch of some base fundamental or even no base at all. Where we then do 'TRIAL AND ERROR' to ensure if it true or not, where if it true then we success at discovering something new if not then we gonna need to imagine it differently next time.
But i agree for production or something that already exist it could be bad as we want to learn it not modified and try something new with it.
Well that depend on the model that do the jobs of searching, if it being train as you said and being prompt to pick the one that have highest probability of right base on how many time it come up then i guess it will do that.
That's an interesting perspective to look st it! And you know, it makes sense cause for example I know that what gpt and claude do is to predict the next token but the fact that it gives so many details about a very simple questions can seem like hallucinations so yeah, I know exactly what you mean...
But I've never thought of it like that before I just took it for granted. Seemed normal to me that it would give me such a long answer for such a simple and small question
Yeah... people have probably already built spaces that do these things
I am curios about combining those text model with some simulation generation model where they could simulate environment to do the 'TRIAL AND ERROR' everytime it being ask question about something from the real world.
I mean for now they limited with text base data where it could be hard to apply true world physc or rule to it, its like "you know it but you don't really know it".
Sounds like fun to test!
Yeah, that wouldn’t be great
is that true?
Okej ich habe es gefunden:)
Check this perplexity answer for AI copilot for Youtube - https://www.perplexity.ai/page/ai-copilot-for-youtube-k3w9dMgYTqOOH9d.52mNvQ
Begins at 10am PT
Join Kevin Weil, Antonia Woodford, and Amadou Crookes to hear about new releases and watch live demos.
Hi there, I have a question regarding the new update for Spaces (adding links as sources). Does the tool update its information regularly or it downloads it only once.
I'm asking to be sure if the information will be updated or not.
Congrats on the Carbon acquisition
Nothing interesting today... still cool to be able to dial to ChatGPT
I'm currently facing an issue with Perplexity where I keep getting the message: "Please try Perplexity again later. Please sign up for higher personalized rate limits. See Perplexity Terms of Service for details."
It seems like I might have hit the rate limit for requests. According to their documentation, there are specific limits on how many requests can be made per minute depending on the model used. If anyone has encountered this before, could you share how you resolved it? Also, does anyone know if signing up for a higher tier actually helps with these limits?
Thanks in advance for your help!
OpenAI just released o1 on their API so that's why...
Yesterday
Ye, surprised me tbh took forever to release it to the api, o1 with search feature is kinda interesting
By the way, are the ai models in you.com actually the ones shown in the site? or is it just system instructed in a way that mimics these models...
They’re actually the ones. All the models I’ve seen can’t answer a riddle I have that o1 and o1 preview can only solve. Gemini 2.0 sucks and can’t even solve it
Github o1, you.com o1, and chatgpt o1 solved it correctly
Don't forget it's the flash model...
Not that one
1206 was confirmed to be gemini 2.0
Adobe?
Lol
Ngl I would say yes
I only used it because it had unlimited claude opus, but when they released o1 and it finally became available to api users, that became a bonus for me
How many queries can you do with pro?
I guess Perplexity would do the same
Ikr. I was expecting more from 1206 but it didn’t solve 2 riddles that o1 solved… though there was a different ai that did solve questions way better than 1206
Forgot the name, it wasn’t o1 and it almost solved all of them unlike gemini 2.0
But I've heard the models in perplexity aren't the same ones? or maybe I misunderstood something
Yeah... the only one cool thing you got is the Deep Research
Qwen max with a specific system prompt begins to solve them at o1 preview level but without it, its like nerfed…
But I guess even that hallucinates a lot
Nah, they are the same ones
They just hallucinate sometimes if you ask it to confirm to identify themselves
Thats normal
But the model who's looking for the searches is the default one right?
That's what I've heard
I’ll try digging more info about this, can’t remember from the top of my head
Thanks!
And have you tested Gemini 1.5 Pro with Deep Research?
And you know what 30$ a month for unlimited queries and 200k context window ain't that bad
But that's for teams I guess
Yes, ngl it surprised me yesterday
In a good way or a bad way?
I guess in a good way
OK I really want to get pro now because of you 😅
I DMed you
Good way, the deep search was impressive in a way
Nice! I will probably buy it once it gets Gemini 2.0 Pro support (if there will even be a Pro version). actually, the value you get even now is quite good because you still get 2TB of storage!
And forgive me if my English sounds weird at times
No worries all good
Yes, I would definitely wait, gemini 2.0 is definitely good for long context memory, but for reasoning I would easily go with o1 as the best. I hope it beats o1 one day
Gemini 2.0 pro will be godly
I think there might eventually be a 2.0 pro
I know it sounds stupid but I was impressed by the fact that it was able to count how many ("r")'s are in the word strawberry lol
Because unlike o1, it doesn't have that same thinking process as far as I know
The you.com usage of o1 is about 20 - 30 a day. It fluctuates though amd it has 64k context
I have a question, for academic use, shall I use Claude 3.5 sonnet or sonar huge 405B/ Grok-2?
I am not really sure as perplexity states that their Claude 3.5 Sonnet model is up to date as of October 22 but there are new models after that
For academic, I recommend claude 3.5 sonnet or claude 3 opus. They’re good at explaining and writing
alright thanks
you might like sonar or claude better . i didnt really like grok but could be good
also in my opinion haiku is the newest updated one on there so its actually very good atleast has been for me just as good or better in some ways than sonnet
you like Orion?
100%
me too
you can use orion in you.com for $5 lifetime and unlimited
How to get long answers from perplexity ai ?
Dms
Ong, best thing ever. It has 1 million context memory too
Those R problem mostly depend on tokenizing, LLM don't use word but token where if it actually gave the right answer without doing breakdown of it, mostly its because it has been overfit with it.
But i understand why there some people still using that as a test, if the model make mistake on the first try then doing the second try by breaking down the word into separate token then we can say they can reason through problem.
Space => system prompt => "Insert your prompt that prompting it to do details breakdown of every step in the output if not it gonna get shutdown."
when is the discount next time?
deep dive
Wow 😯 Perplexity mentioned in official Github mail
𝕄𝕒𝕘𝕚𝕔 𝕀𝕤 𝕀𝕟 𝕋𝕙𝕖 𝔻𝕖𝕥𝕒𝕚𝕝𝕤 ✨
This is exactly why I was impressed. I was aware of that and knowing that makes it even more impressive because it shouldn't even do that.
yeah basically llms are kinda dumb they just memorize stuff and sometimes they get lucky but if you trick them by changing the words a little they get confused like a toddler trying to solve a puzzle theyve only seen once also who cares about these tests anyway lets talk about something fun like kittens or explosions or maybe kitten explosions
What about the teams plan? That is unlimited or does it have the same limit? Was thinking about buying a teams plan on you.com
200k context window but I guess everything is unlimited... I just know that almost all the models are unlimited with the pro plan anyway
Hmm okok. I'll have to buy a plan soon 😅
That's what I said yesterday but unfortunately I couldn't get the student discount 😔
Could have saved me half the price
They offer student discount huh
Yeah if you have an .edu you get that automatically
Uhh 🥲 what if your mail doesn't got the .edu but is a student mail?
Try
And if it doesn't work
You can submit a form or something
But the reason why I couldn't do it is because I'm not a student 😂
Hmm alright, well, I want o1 so gonna go for teams anyways, don't care much about the discount at this point 🙂
But I have a friend who's in college
I'll be honest I would say the same since my pension already covers almost everything but huh sounds very expensive
Hahaha I'll ask him in person about his email
Waitt
I am a student though, I don't need this I assume 😅
Won't it just get deleted
oh okay
Would very much want a private mail
Sheshh I need it more than you I'm not even a student 😔
Are you sure that's safe
I mean it's a temp mail gen, anyone can see the mail if they lucky 😂
And reset your password
I've got nothing to lose
Well I say you create your own temp mail 😅
Nothing hard 🙂
Wait it is temporarily
I thought it only gives you access to change the email temporarily or something like that
Seems strange to have an education temp mail
Ahh doubt it 🙂 or idk
I know what temp mails are
Just mail with subdomain edu
Anyone can create
They probably cross check with listed unis or something but idk
Don't you need a website with that domain to even create a Gmail like this one?
Uh, website, no. Just a domain
And a hosting provider (can be free)
Many cheap domains ranging from $1-$10/y
Not bad but there is no real way to know if they'll work for sure
Yeah🙂
It's always fun to try
Curiosity is what led humans to invent stuff after all
Anajrim, you said I can have you, what if I can have your friend's student mail 😏
But how would that even help me if it's just temporarily
That's what I don't understand...
What if my friend decides to take your account 😎
I'll ask them to change my email once I buy the subscription with it 😎
There must be a way to do this
Ah, I mean. I don't think it'd be a problem if a friend wanna give up their mail 😌
Actually, the temp mail would work just fine for your case
Don't think they check
But again like
What do I do if want to keep my account
Do what I said
To ask them to change the mail or something
You know it will probably ask for a verification code at some point no
?
Otherwise you can take my mail though you'll need to fill out the form, I don't mind
Probably will, yes
Will you really agree to that
I want teams plan, pro is useless to me
Discount only on pro
😍❤️💋
I love you 😍 💋
Give me a kiss
🤣
I added u, I gotta go
I text you my mail and such later
Don't trust a stranger tho 😅, just a warning for future and change the mail asap after u done 😂😂
Of course that's what I wanted to do after you gave it to me
Nice nice gtg now
Bye for now!
Cya
is that just to get perplexxity school discount thing
Not for perplexity no
For perplexity, I will just pay the full price...
But I want to test out you.com and my friend is in college already but I think his email isn't .edu
#💬│general I have an query :
- In perplexcity api is there is any toggle or api parameter to switch between web search or in the knowledge base whatever we have provided?
Please give me a hand!!!
NotebookLM (The video must be public and have been online for at least 24 hours.)
does perplexity updates the gpt4-o model to the latest api updates?
ask perplexity.ai
it is the weapon for research
search for your answer
Gemini 2.0 Thinking experimental is on fire sir
You are the person that's always saying "sir"
Right
Begins at 10am PT
Join Kevin Weil, Justin Rushing, and John Nastos to hear about new releases and watch live demos.
i don't think they disclose that, didn't find an answer
Am I crazy or was there a christmas gift guide tool published on the Instagram stories of perplexity? I'm trying to find it.
Have you guys tried this new thinking model? It is pretty cool and remindes me of the scratchpad plugin for complexity
Going to test it right now
I haven't delved into it too much but it seems to be onto something. It might even be better than the advanced just because of the CoT
There is one small thing that I wonder about, can you get similar results to what Gemini 1.5 Pro with Deep Research can do or is it more like o1 with the chain of thoughts and all
It might just be CoT
Deep Research is nuts tbh. Imagine with Gemini 2.0
Lol wanna hear something about this exact question, got 3 "r"s with the normal Gemini 2.0 Flash Experimental first try lol but that could have been luck
I was getting "1" or "2"
Weird
Imagine with Gemini 2.0 Pro which hasn't been released yet
Advanced is available
Like I said, could be just that I was lucky
Only if you pay for it no?
Like when you do the Google One thingy
go to aistudio.google
Not through the API
It is there in the API studio
Only the flash models and Gemini 1.5 Pro are available to me
And some Gemma ones
Isn't advanced this one?
I meant Gemini 2.0 Pro even though it doesn't even exist
I've heard of it
But I'm really not sure
I heard it's better than GPT 4o in some cases
But until they call it Gemini 2.0 Pro, I can't tell you for sure...
Fair
I hope some Aggregators will add Gemini 2.0 Flash to their service
I wish I was well versed in using APIs because I would like to use my $5 in credit to just mess around with discord bots that pull from the PPLX API
I use you.com and it only has Gemini 1.5 Pro and Flash (it has many others but not google's)
Sounds cool!
I do feel as if the answers I get from you.com are "worse" than the actual models but that's only because of the way they search on the web... I tried mathmatical questions and it gives the same results
This is a good sign they use the actual models and don't just system instructs it in a way that would mimic those models
I mean, they did add custom system instructions but they shouldn't effect that as much
From my testing at least
Does perplexity have a referral system? I'm getting my entire frat on it
It does.
By the way, I know it has nothing to do with AI but DC just released a teaser trailer for a new Superman movie
It was so random for me at least
😂
Yeah it's great!
The other thing is, I heard from my friend at AZ state that you get free pro if you sign in with a .edu emaill but that hasnt worked for my rommate
using this link
That I did not know
Nice to know
Watch me singlehandely get the 500 sign ups
LOL
Hello
i try to ask perplexity to complete and translate description of a list of game in french
he translate the first of 800 game and ask if i want to continue
i try everithing to proceed to the entire liste but he continue to do only the first 2 and say would you continue for all the file ? (in french)
new developpement, now he say, i can't
i want to proceed all the file because i want trad and shorten desc
Did you ban that guy?
Is it possible that you f*cking STOP updating the website while we are writing something in a thread ???
I was writing a supper long prompt, got a doubt on the spelling of a word so I go search for that work in a new tab, and when I come back the page refresh with a message telling me it have been updated and everything I wrote is lost !
Understand your frustration but this is exactly why you should always Ctrl+ A + Ctrl + C before you hit Enter...
This was not before hitting enter to send the prompt, this was just me switching tabs for literally 3 seconds to check the spelling of a word
The website shouldn't update while I'm on it ! so how about they fix it instead of requiring their users to take this kind of precaution ?
Fair enough
unless its a small prompt or quick question I pretty much never write my actual question/prompt within any AI boxes I put it all in Obsidian where I can sort them
and then just copy paste
Rip
I doubt it, They have some timeout far more than '3' seconds if the perplexity tab is indeed inactive.
Dont start typing a fat prompt and then go afk.
I've already subscribed via Google Play on Android but it seem to has no such payment method for setting up API, please help
Subscribing and API is completely separate.
But it should give $5 free for having pro
Im using only Google Play for payment, there is no setup for that.
Guys this channel where people will answer you super fast, use that channel wisely.
👉 #⚡│ask-community
This could mean a few things, since the clue is cryptic:
• Focus on the letter “O”: Switching from “H” to “O” might suggest something related to the letter O or the shape of O’s, as if pointing toward something circular, something zero-based, or the initial of a product name.
• A play on sound or spelling: “Oh oh oh” may hint at surprise or a revelation—perhaps a reference that whatever announcement he’s teasing isn’t just a holiday treat but something else entirely.
• From Santa to something else: “Ho ho ho” is instantly recognizable as Santa’s laugh, but “oh oh oh” is more neutral and might mean we should think beyond a Christmas theme. It could be directing attention away from a holiday surprise toward something more open-ended or unexpected.
Without additional context, the clue remains open to interpretation. The key idea is that the intended hint involves the sound “oh” rather than “ho,” suggesting that the nature of the surprise or announcement is tied to that change—something about the letter O, a zero, or simply indicating that the initial assumption (a holiday gift) may not be the right track.```
Hopefully it’s o2, most likely 🔥
you are right it has something to do with taht
I was thinking Orion
I'm thinking its orion (aka GPT-5) since o1 is short hand for the full name GPT-4o1
its either o2 or Orion
Yes, either will definitely be exciting. I already had fun with full model o1, but o2 or orion definitely will be interesting
If we try to interpret the clue symbolically, “Orion” could be considered because the constellation’s three-star belt might be loosely connected to the repeated “oh oh oh” (three O’s). However, this is purely speculative. There’s no direct indication from the original context that Orion is meant to be part of the puzzle. Without further context from Sam Altman or a known link, it’s likely just a coincidence if “Orion” comes to mind.```
In the context of AI releases and model upgrades, it could be hinting at a new version or a next-level model that emphasizes “O,” possibly signifying an O1 successor or something that’s even more advanced. It’s reminiscent of how each new model might be given a code name or a letter-number combination (like GPT-4, O1, etc.). By drawing attention to “oh oh oh,” maybe he’s nudging us toward anticipating something that pivots around the “O” naming scheme—like a direct hint that a new model version or a special announcement is coming soon.
Ultimately, without an official follow-up, it’s all speculative. But given his past patterns of cryptic tweets before big announcements, this seems like a subtle way of saying, “Watch closely—something new is on its way.” ```
That’s what it told me
Oh shoot, you got a point about orion
hi
guys dont forget the moment of the openai whistleblower who got bounty hunted
Chill with the tin foil
i declined ur friend rwquest sorry @wary umbra
mmm
lol they just released o1 full, why on earth would you think they would already have o2. I need to know what your putting in your coffee.
They would do it since Google has pretty much upstaged them this whole week and they just launched a new reasoning model today for free, and its pretty good considering its FREE through their AI studio.
Hello, I'm a AI Backend Engineer with over 6 years of expertise in designing and deploying scalable web applications and AI-driven solutions. My skill set spans both front-end and back-end technologies, as well as specialized knowledge in AI, knowledge graphs, and vector databases.
What is this
Nice
#💬│general Anyone? 👀
Hello there,
Now I am developing one project with Perplexity API. But I think system prompt is not working via API.
Are there anyone who faced same problem or who knows how to make it working?
Now I am working these models.
https://docs.perplexity.ai/guides/model-cards
Hey @fading barn!
The sonar online models will always use the web to get the latest information. There is no additional switch
Hey @willow umbra!
Please note the following:
The search component of our models do not attend to the system prompt. You can use the system prompt to provide instructions related to style, tone, and language of the response.
I found Perplexity
I seriously thought this was true for a second 😂
Check out this page maybe the product there helps you: https://www.perplexity.ai/page/ai-copilot-for-youtube-k3w9dMgYTqOOH9d.52mNvQ
Here's a concise summary of the key updates for Day 11
**Native Desktop Apps: **ChatGPT now has fully native desktop apps for Mac (and soon Windows). These are described as more lightweight and resource-efficient than the browser version, enhancing user experience.
"Work with Apps" Feature: This allows ChatGPT to directly interact with various applications on your computer. It automatically pulls context from the selected app to understand your requests, then executes tasks within the app. This eliminates much of the copy-pasting currently needed.
Supported Applications: The video specifically demonstrates this feature working with Xcode (an IDE), Warp (a terminal), Apple Notes, Notion, and Quip. More support is promised.
Model Selection & Capabilities: The presenters switch between different OpenAI models to highlight their varying strengths. Model 001 is shown as effective for complex coding tasks, and the Chain of Thought model reveals its reasoning process step-by-step. Advanced Data Analysis is also used to create a graph from data extracted from an app.
**Advanced Voice Mode: **A new voice mode lets you interact with ChatGPT conversationally, providing spoken prompts and receiving spoken responses. This mode is shown being used to edit a set list for a holiday party.
**Security & Privacy: **OpenAI emphasizes that the "Work with Apps" feature only accesses the contents of an application after the user explicitly selects it. This ensures user control over data shared with ChatGPT.
**In short: **The video showcases substantial enhancements to the ChatGPT desktop experience, primarily through the "Work with Apps" feature, which increases automation and streamlines interaction with various productivity tools. The addition of voice input and various model options adds further versatility.
lmao funny it looks like perplexity
Does Perplexity now have voice support?
ask perplexity.ai
GPT 5 sir but there is a catch with the message cap sir. 50 uses per month sir. 😂 on the unlimited side sir, 500 dollars per month sir. 😂😂😂
Node pay with millions of users launched this lol
Is this not just that one open source project wrapped with a sidebar
Like that exact UI I have seen somewhere
this
this is literally that
Yes this
how original 🔥
What's the best way to get support? I sent an email on Monday regarding my promotion code that shows as invalid for Pro. The auto-response mentioned 1-2 working days for a response, but I haven’t heard back yet.
Hey @half pike! Please DM me your account details. I'll check.
Sent you a friend request so I can message you. Thanks
definetly inspired by the awesome product that is perplexity
OpenAI o3 and o3-mini—12 Days of OpenAI: Day 12 - read it here and follow this link for follow up questions and chat with video with AI Copilot for Youtube
https://www.copilotforyoutube.com/search/openai-o3-and-o3-mini12-days-of-openai-day-12-T7sbiQRKxbMdlrWTddGC9L
Does Perplexity intend on supporting o1 anytime soon? The o1 on other platforms is pretty amazing when hooked up to search and I would love to see a real implementation by Perplexity.
Perplexity cant afford o1 sir, our subs are just for friday pizzas sir. FYI sir.
Yes sir
aw man
go try
thank you @oak mango for this video
Well I guess Perplexity is cooked then since naturally OpenAI will hookup o1 to SearchGPT sooner rather than later and if they don't then Gemini 2.0 Deep thinking mode will connect with Deep research mode and take over search.
Seems like they dont care sir, they seems to be not interested anymore actually with the competition sir. Probably lot of people here already noticed that sir.
look around you sir, try checking their announcement section sir. what do you see sir. Where are they heading sir.
though perplexity was really promising sir. I just dont know what happened after this friday pizza thing sir. They just suddenly sir.
let me tell you why, cause all llm wrappers / agents will get cooked eventually because they're able to get replaced by the creator of the APIs they use:)
There is no incentive for API creators to destroy their API clients. This means they will compete agressively in some areas yes but your statement of "all llm wrappers / agents will get cooked eventually..." is incorrect.
Maybe not all, but if people think Google, OpenAI, etc. are going to leave a market like search (which appears to be Perplexity’s target) for a third party to take… then I have a bridge to sell you
I don't think so, they won't compete aggressively. In fact, the "Big Four" won't compete at all.
They'll eventually get replaced and forgetten but until then there will be unicorns using their services' wrappers intensively cause they have to get backed.
Monopolies don't work like that:) How fortune or unfortune. You're right on maybe not all though, the future will show. Or we'll show the future to shape.
Will ppx have their own reasoning model? i has try o1 and then back to ppx, the different is actually quite clear for complex problem.
they dont
Agreed - but the players with their own models have a significant advantage.
isn't that a bad thing? the competition already have that even in the opensource community they already have that kind of model.
a lot of things are going on behind the scene from canadain studen discounts to all other things, they are hanging on a thread while jeff bezos is bobbing his head at anthrophic ceo. brb im gonna have dinner.
Maybe they should stop giving student discounts then 😅
If the institution wants, they could be pay for perplexity for all of their students...
i dont know how that campaign is going
im not sure if perplexit.yai is popular in canada
i could be wrong
because they have access to the statistics of from where do people use perplexit.yai most
And where is it?
and the case usage are probably mostly from canada hence why they target students from canadian schools
Which place/country
🇨🇦
that is just my theory, dont take my word
yes if you are student in canada
I'm not...
I remember you sent a link to temp mail generator. Can anyone just abuse this?
Not good for perplexity
I was able to get perplexity for 10$ for one month because of an issue I had...
I think you need to have student email to actually get that discount, i could be wrong tho.





