#š¬āgeneral
1 messages Ā· Page 86 of 1
Oh iām sorry
Hello guy's
Guys I need help! Iāve connected my LinkedIn with the perplexity how to get one year premium now
can someone help? ;-;
Didn't find a solution? Is there any way to bypass this so that PP obeys the instructions?
Is it pay plan or free plan?
No indication as of yet
Nope, not me
Hey, I'm new to this server. Where would be an appropriate place to put my findings on having Perplexity do numerological talismanic magic for me and outsourcing it's finding? Where would people be interested in hearing using Ai to further occult and spiritual studies?
Why did I choose the sonnet model, but the response clearly isn't using it? The result is fast but clumsy.
I'm definitely not renewing. First, you removed the Opus model, and now you've secretly downgraded the models. It's terrible.
Hey @boreal badger!
If you find the original message helpful, please consider reacting to it with the :star: emoji. If the post is appreciated by the community and receives 5 stars, it will go to the https://discord.com/channels/1047197230748151888/1082806833938436228 channel and the post author will get the <@&1082034222778302614> role on Perplexity.
real XD
how long does it take O1 model to generate?
Hi guys, need help here. I subs perplexity on my Apple account, but i can't login same account on android apps. Someone can guide me?
Please go to https://www.perplexity.ai/settings/account and copy the email there. Then use login via email on android and paste that apple relay email address there.
How about private apple id? is it will work the same? I can't login with the option you give to me
Should work the same. Just copy the email that you see on the settings page and use that one to login via email on android. If that still doesnāt work then please send an email to support@perplexity.ai
that is plastic grass* for visuals
Alright dude, thanks for the insights
hello, compiled few pplx numbers from my non-scientific tests. some may be outdated (havent re-tested everything), but I think it is a nice overview, so users can see why pplx behaves sometimes like it does, or for comparing products https://monnef.gitlab.io/by-ai/2025/pplx-tech-props
Interesting, and useful š
how does this work really? With pro you get access to all other ais premium feature or?
Hey @hard torrent!
With a Pro subscription you get access to premium AI models such as GPT-4o, Claude 3.5 and Grok 2.
Further information: https://www.perplexity.ai/hub/faq/what-is-perplexity-pro
amazing, thanks
Is there anyway to customize the temperature levels on the LLM's to make the response more chat like ?
No
Hello š
is anyone else having slower than normal response stage times? on my end it seems slower even if just taking from the moment the top of the response starts and when the "Answer" section types out
Me
Its been like this for a couple days
Really slow response
Usually the pro works just fine but its been meh since this thursday
That's what I was afraid of. I started paying $20/month recently and it wasn't a decision I took lightly. Before I chose Perplexity I tried each chatbot I could get ahold of, on both desktop browser and mobile app. I gave each chatbot a few weeks of use without any other one or any search engine for a few weeks so I could get used to them.
So I ran ChatGPT (free), Gemini (free), Copilot (free), and separately I was also able to use the Copilot that shows up with Microsoft 365 for large organizations. I'm glad because for some reason I don't know, they look and behave totally differently like they're not even made by the same company. I got the impression that they answered things differently, but I can't say for sure since the point was to try getting used to them to see which, if any, was most correct and also felt natural and made me more productive. And of course I tried Perplexity the same way using the Pro trial but I also tried it on the free mode and there's no stark difference like Copilot as everyone here knows
I'll say briefly that Copilot did ok. The personal one looks sleek on desktop but put me on timeout for a while when I used it heavily.
The pro one looks... well, like you're back to the time when you used internet explorer if you were even born then.
Copilot personal on mobile was interesting because when I opened the app it began a conversation. It asked for my name and right off the bat it did something that's the opposite from what I expect. It said something like "got it, and mispronounced my name the way people often do if they see it written but don't know which way I say it. Then said "just kidding [correct way]!" The voice it self was quite smooth. But it didn't know my name on a separate occasion so it's just another birdbrain. I did not take this into account either way because I couldn't have the same kind of interaction with everyone. I only mention it because what I expect is the awful over the top self deprecation most clearly seen when the bot makes a mistake and you tell it that. Which brings me to the champion at that, ChatGPT. My focus for all the tests was to do what I always do, and I don't always go about chatting up bots and telling them they're wrong. But even then, ChatGPT still did its thing. Which would be fine if after the lame excuse turned 'I got this now!' actually got it, but it don't. It goes right back to making the same mistakes. "Baby, you were born this way."
I forgot to mention that it coincided with Microsoft coming out to yell at everyone that no, it is not taking user data to train Copilot on the corporate version so everyone please stop saying that it is. It was bold to make such a definitive statement on the corporate version and I fully expect some type of Wardrobe Malfunction sooner or later. I did take that into consideration, yes, you know it on the corporate version. And yet I wish that it could learn from its mistakes at least locally which means part of me wishes they could learn from my data instead of taking my data and spinning the LLM around "~One more time~"
Y'all already know I chose Perplexity [song pending], so don't ban me yet. I've gotten this far. Which brings me to Gemini. I think their models are the fastest and most environmentally friendly by far. I also think the Wikimedia Foundation should sue them and both things are for the same reason. I'd ask something and its reply was literally a piece of Wikipedia. Even back when schoolchildren discovered Wikipedia and copy-pasted right into their work I never heard any teacher say: "can you believe this?! They turned in the assignment and all it had was their name on the page and a glued on cut out from the Wikipedia printout!"
It could have at least googled my question and show me the first ad, but no. Gemini was the worst, period.
God I should have held back on that last mug of coffee. Perplexity won on its own merits. It gave me answers in plain English, and it even dare give me the sources it says led to the answer. And I'll be honest, sometimes Perplexity was wrong. How do I know that? Because I had the sources. I wish I had saved the specific example, but if I did that for every chatbot I'd reach AGI long before they did. What I can say is I could see in the source text how it got it wrong. I wouldn't call it a hallucination so much as a misreading of the text.
And that was, perhaps, the most human-like thing throughout the whole experiment. My ability to get answers quickly and plainly in a way I could follow along, yet also giving me sources I could use, or not, and none of that "I'm sorry I must have been paying such close attention to your bytes I forgot basic math at the end there! But I promise not to look at your sexy sexy bytes, so keep going, ok?" When it was all over the only thing perplexing was that odd and long name. Still beats "Mastodon" though, so we're cool.
Or we were, until it got reaaaaaally slower and slower. I hope it recovers and never happens again or I might end up remembering that dumb song people sing while being under a stick like they're competing for Best pet gate burglar"
And Now This
And now, which is Perplexity and which is Google?
i think well
Friends, do you know what is the context length of the pro version of Claude? Is it the same as the official 200k?
u love the microsoft copilot @royal cedar ?
this is scary but im getting the same xact result with deepseek v3 and chatgp-4o, the only difference is that deepseek is free and comes with the o1 like reasoning button
lmao
Hello
I've made a merch purchase but the status is not changing at all. Order placed on 29 November. Can anyone from Merch support help?
Hey @limber umbra!
Please contact support@perplexity.ai
I have. received only first response on 27th Dec. No response after that even with multiple follow-ups.
Any other way to contact or get traction?
Is perplexity ai have chat bot for discord servers?
Please dm @signal hamlet
Correct š I like Deepseek V3 much š The model is really good and the search results are great too. And itās free.
and is open source
I tell it to do something and it just explains how to do it
Hi team, is there any refer & earn program running in Perplexity AI in india ??
Is perplexity down right now? I'm getting empty
whats wrong with perplexity, is buggy today
why is it not smooth like chatgpt, writing on it is too buggy
have pro memership, still the experience is not smooth
hello, everyone
i have some experience in development world, wanna work with senior developers
I'm 100 percent shure that i cancel the subscription but i was charged anyway
Where i can find i fix the pay has not been completed is in hold atm
Or where i can talk to?
So slow today , canāt stand it
I had a very creepy situation with Perplexity recently. I had a specific pain in my body which i havent had before before in my life, and i looked for these symptoms in perplexity to see what they mean (i did NOT use google or any other online search tool). A couple days later, i started getting instagram ads for the specific type of pain that i had. This completely freaked me out that perplexity is reading my chats for advertising.
So i completely deleted the app, i downloaded from aurora store (as you suggested), not the official store so it's not linked to my identity, and now i use perplexity without being logged in, to make sure my chats are private.
If anyone has any other suggestions on how to keep chats private please let me know
tell me pls how to disable automatic text correction in perplexity on macos
Is Perplexity slow today for you guys too?
So why did I wake up to a push notification from Perplexity this morning pushing what seems to be a āsponsorā style propaganda post selling me the benefits of Ozempic and how can I make sure it never happens again short of uninstall?
The Perplexity app has been non-responsive on my iPhone for at least 4 days. What is going on? @
Fine for me
Maybe youāre in China
do you find the "Shop now" feature useful on the app?
to me, it is useless and makes me want to avoid using perplexity when it comes to product search
no one wants to find a product and just hit: "buy now" without getting a chance to visit the seller's site, reading the reviews, etc.
this "shop with one click" idea is, excuse my words, stupid.
I think you can just turn off notifications in the app settings
Hello everyone! I was wondering if Perplexity has undergone any security assessments, such as SOC 2. If so, could someone guide me on how to obtain the report? Thanks in advance for your help!
Guys try this and see how different ai behaves
Lest try a new mechanic since you know how you work give yourself instructions like a a senior dev giving their computer instructions to achieve a goal you can code anything you want the only requirement is this.
PPXI got a callout from Nvidia at their CES 2025.
yes but just sense I asked it to recite all of pi
in the form of a jazz style contemporary musical
in french
Is there any word on them giving more o1 usage per day?
Hey @fiery anchor!
Please check https://trust.perplexity.ai/
Ready to turn trust into your competitive advantage? Sprint through security reviews and quickly share key security information with Trust Center.
try elon musk
I didnt know the flavicon changes while thinking
Has been for days at least for me. Both on mobile and browser
Takes minutes to write the answer, regardless model used
Is Perplexity still broken?
Well, works, at least for me, but really slow
Like lately I've used i like this: Send a prompt, alt+tab to something else, alt+tab back after a few minutes
Ok. I'm just wondering if they still use the cheap quick answer bot or like Claude Opus it's better at giving answers but can also be more natural and conversational about it
LOL
Thanks DeepSeek
is so slow, so annoying since last 2 days
making me hate it
Delicious.
one question: there is a way to delete a query in the middle of the conversation? in the 3point option i see "delete" only in the last query and not in mid responses
you guys shhould hire whoever is in charge with complexity. amazing ui/ux and they keep shipping
Did you get assigned SOC2 compliance at work and now you have no idea what to do?
it's not slow at all
and wasn't at all
do you use free or pro?
š It has been a journey but I do have an idea now though .. I emailed support as there is no mention of compliance except in their entreprise website
Oh theyāve just sent me their trust center info š„¹
Ohhh you're checking if Perplexity is SOC2 complaint, not how to be SOC2 complaint haha
I doubt any AI is compliant
Youād be surprised.. everyone is now compliant
Got an email that says As a free user, you get 5 Pro Searches every four hours.
Where? How? I'd like that please
you need to toggle em on
ok, that was probably my fault for not being more explicit.
I know how to toggle that switch... I mean I just get 3 per day
thats a bit broken, it would be good if it had some basic logs e.g. you could see why it went from pro back to standard
try changing models and seeing if your 5 searches are back in 4 hours
might be your searching with a model thats too jucy and 1 search counts as 1.3 searches or something
no, the setting is locked
this changed like months ago. Back when I started using it I was getting 5 every 4 hours. My post was a bit tongue-in-cheek, pointing out that Eliot from Perplexity didn't get the memo about the reductions for free users
according to Gemini When sharing a perplexity page the icon display's the Perplexity Graphic and i want to display the Page's graphics
You're describing a common issue with how websites handle link previews, specifically when sharing on social media or other platforms. The "Perplexity Graphic" you're seeing is the default image that Perplexity.ai provides for its site when a specific page's metadata isn't properly configured for sharing.
I wish there was a permanent subscription I love this app so much
Any chance we can fix the laggy typing? I'm stuck on a few projects and my work around (type in notepad, paste in Perplexity) is only good for so much.
Write a c code for calculator
If I am already a Pro user, then this option should always be on. I never find any value to get a 'less intelligent' result.
OTOH, what I notice that while I always keep it on, often times Perplexity turns it off and sneaking a cheap query on a Pro user's behalf.
Honest mistake? Maybe, but that's dishonest and also wastes my time.
That's feedback.
š
Perplexity voice is too slow. Especially Alex and there doesn't seem to be any option to make it talk faster. So annoying
It also doesn't read the titles. Just the texts
yup need something like chatgpt voice feature
Gpt generally produces better responses but it depends
is there any way to scroll the convo all the way down in the app on Android rather than scroll it till my finger goes numb?
IS there a way around deleting the cache when the perplexity.ai chat window is slow and sluggish and experiencing input lag? It was nice being able to scroll through the convo history which I assume is getting deleted alongside the cache.
pro
i write text and it takes 1 sec to show up in imput field and is so slow, idk whats wrong
i need to work and i'm so annoyed by that since last few days
isn't that normal?
@green glacier this the problem me and a lot of people are facing
unable to work on my projects
Is the API down? I'm getting error code 524, which I'm not sure what it means.
Are there any companies doing integration of Perplexity into an office suite for content creation? I find a lot of the static models lacking. MS 365 Copilot is static but will work with existing documents. But if you want build documents with up-to-date information, you have to wait for static models to get updated because they usually don't do search and creation in a single workflow. Perplexity has better output than most of the others too. I prefer it to ChatGPT. Sonar Huge is awesome too.
Youzu.ai is one of the first AI-powered tools that suggests beautiful room designs and shows you exactly where to buy the pieces based on location. If youāre curious about how it works, there is a guide. Itās saving me so much time and stress!
Hey all! I'm trying to train a perplexity thread which is involving pasting copied text from notes into the chat. Whenever I copy text the "want more uploads?" Window appears.
I've tried different browsers, apps, etc. I keep receiving this error even though I have not nor am I planning to upload anything.
I also asked the ai about the issue and every solution I've attempted and it said to reach out here or via email to report the bug.
Thanks again all and I hope everyone is having a happy 2025!
How many files can i add on perplexity spaces as a pro subscriber
When will my issue get handled?
https://discord.com/channels/1047197230748151888/1322244805731811328
Okay that is pretty sick, I have to try this out
Straight up - I want to build something that matters with this group
This group is packed with incredible talent from prompt engineers, ML Experts, algorithm specialists, researchers, network architects, and more. The diversity of skills here is exactly what makes collaboration powerful. Some of you are deep in neural networks, others excel at system design, and some are relentless researchers.
Thatās the kind of mix that leads to truly groundbreaking work!!!
I want to start a project that matters. Something we can all look back on and say, āWe built that.ā Whether youāre someone who codes 12 hours a day or only has time to contribute during lunch breaks, what matters is your willingness to create and collaborate. No egos, no gatekeeping just a team of people bringing their unique strengths to the table
We already have the hardest part covered lol: getting the right people in the same space. This Discord is full of brilliant minds and ideas. Letās take advantage of that and actually build something meaningful together
I'm trying to translate a korean novel with perplixity and it's been both great and terrible. Does anyone have a good method? My current biggest issues are the response word limit, and sometimes it will make stuff up, or say "Sorry I can't assist with that"
I am using Perplexity Pro and have a Space created for a project I am working on. This project has several files that contain context and input to the thread. Problem is when I make a change to the file Perplexity does not seem to reread the files when I regenerate or rewrite the thread. Any way to force it to reload of the input files short of deleting them ansd re-adding them to the space?
OpenAI CEO Sam Altman Faces Shocking Abuse Allegations
Guys is perplexity down?
might as well be. I cant type anything.
Awesome! Let me know what's your experience.
Is this normal? Today, Perplexity has begun to take information from a colossal number of sources, even for simple queries
yes they are learning from deepseek
taking notes
I'm worried about context handling
3 messages then it forgets
The context is not fixed at 32k it can go up to 70k+
hello! I'm a product manager at Ecosia (a search engine that plants trees) and we are looking to have a chat with Perplexity to discuss a potential deal and/or partnership š«¶ I've been struggling to find a contact point for that thoigh, anyone could help me here maybe?
I fully agree with that one. Perplexity offers a lot but using it is terrible š„² at least I think that way. https://x.com/omidaziz/status/1877409601202631083?s=46
I disagree I think personally perplexity has been easy and straightforward even for my kids
I often turn to other tools, mainly ChatGPT, not because Perplexity canāt handle it, but simply because I canāt stand using their app. Same goes for their website š Even though their āresourcesā are quite extensive š
for perplexity version of o1, what is the max context size for inputs and outputs?
is this a known issue that typing in the input box is very delayed?
FYI: Perplexity having refreshing issues on Chrome web, if you have SimplyCodes extension installed, you need to remove or disable it.
Iām looking to optimize my use of Perplexity for professional research and analysis, but Iām struggling to identify the most powerful and efficient configuration. Hereās what I need help with:
1. Model Selection: What are the strengths of different models (e.g., Sonar Large 32k, GPT-4) for specific tasks like in-depth research, coding, or summarization? Which is the most robust for high-level research?
2. System Prompts: What are the best practices for crafting system prompts to improve accuracy, formatting, and relevance? Any examples would be appreciated.
3. Use Cases: Iād love guidance on when to use Perplexity over other tools like ChatGPT or Claude for tasks requiring reasoning vs. factual accuracy.
4. Features & Tools: How can I fully leverage features like Collections, Focus mode, or integrating external tools for better productivity?
5. Demos & Tutorials: Are there any advanced tutorials or demos that showcase optimal setups or workflows?
Iām aiming to create the most powerful and efficient setup possible.
So can somebody who is daily Perplexity user with lots of experience and has done lots of testing and configuring please give me the most optimal tips
Like what should my pre-configured system prompt be for my profile? Which model? Why? Iāve still been using Claude 3.5⦠because I donāt even bother to try any of the other ones
Where can I post my public space?
Copy the link and publish wherever you want @stiff rock
hellooo everyone
I disagree, perplexity is extremely easy to use.
Why is perplexity so damn slow?
Can you delete once added file with it? I mean using the app? š Maybe the way you use the app is such that you donāt notice the flaws that others, who use the models differently, do. For me, ChatGPT has the best app š
bump
ChatGPT is garbage
ChatGPT may be one of the worst models out there haha
I m not talking about the models at all⦠š¤Ø
I can tell by the way you type you still look down at your keyboard
Whatever
The claim that the Moon is moving closer to Earth is false. In reality, the Moon is gradually drifting away from our planet. This recession is caused by tidal interactions between Earth and the Moon, leading to an average increase in distance of about 3.8 centimeters (1.5 inches) per year.
Itās important to be cautious of misinformation on platforms like TikTok. False claims can spread rapidly, so itās advisable to verify such information through credible sources. For instance, a TikTok video was debunked for falsely claiming a giant face was visible behind the Moon.
i clicked on this and it spammed my desktop with pdf files @vapid onyx @halcyon coral @serene sand
hello
Fix this app now! I'm tired of this slow ass and app!
I just discovered spaces and thought it would be perfect for building a very specific search engine that goes through a bunch of sources and gives me what I want. But I'm fighting hard with it
What I'm trying to do: Get Perplexity to find all music related events in Tokyo, based on a bunch of sources I've given it that have event information on it
https://www.perplexity.ai/collections/tokyo-events-aNYJp4E3RuG_IVJZaYD_Ig
I can literally see big events on the sources when I click on them, but when I ask perplexity to find some, it just returns generic slush and leaves out the main ones. Anyone an idea how to get Perplexity to do what I want?
Tried using Grok, Claude, ChatGPT 4o
Hey @night ledge!
Please share some threads
- https://www.perplexity.ai/search/any-cool-events-today-uEwEw.cpS_GunGMNsyPWYg
- https://www.perplexity.ai/search/any-cool-events-today-NvkcjHCpS2ykL.6nGzxMfg
- https://www.perplexity.ai/search/any-cool-events-today-gYP0sQ0iQneOqOU2ACpX0g
- https://www.perplexity.ai/search/i-wanna-go-out-tonight-any-coo-bVrNVxEzR7SynnfREMaZ8A
- https://www.perplexity.ai/search/what-are-club-events-happening-_FQyePHAQ.qPUb7v_Jmepw (this one finally surfaced a major event happening today (Black Coffee is playing), but left out a lot of other events from the sources)
Is Perplexity down? I haven't been able to use it in 24 hours.
Are you connected to a vpn? If you are please disconnect and try again.
Ah this seems to be it. Why can't I use it with a VPN?
you can read more about why pplx blocks some vpn servers here : https://discord.com/channels/1047197230748151888/1232978227371442237
you could try to connect to a different server of the vpn provider
Claude is better than everything else
But you r all talking about models. The LLMs are not considered here - only the experience with the app
Hello guys,
I want to build basic portfolio website which i can host on github. Can any body suggest me any tool or anything using which i can build one.
Thanks
Hello
hi
does anyone know how I can get better codegen/refactoring through sonnet via perplexity?
I just need to churn a bunch of tokens but don't need a web search each time, should I use the writing focus mode?
might be cool to add a "Code Generation" mode š
or even better, a way to upload a "codebase folder" to index, like AI studio
It would be epic to have a Git repo integration in Spaces, which reindexes diffs on each commit
what is another AI on best with chat gpt??
When in comes to ChatGPT I like o1 / o1 pro - I m using it the most right now, but itās quite expensive and I have no idea if I keep my subscription. The best thing would be to have Claude 3.5 Sonnet in ChatGPT app šš
200/month chatgpt subscription is pricey, but honestly found to be the best value
after that, 20/month chatgpt, then 20/m sonnet, then 20/m perplexity imo
idk how claude + perplexity are matching $20/month. Seems too low to be sustainable but idk
If perplexity offered a $100/m plan, between the basic plans & chatgpt pro that would be compelling
without constantly changing system prompt and completion parameters :/
100/m is still fairly affordable & could provide power users with the experience they want
yea writing mode is the best one for coding
3 hours ago was released this video: Joe Rogan Experience #2255 - Mark Zuckerberg. Check it out from summary do detailed answers with video to text / Chat with video. https://www.copilotforyoutube.com/search/joe-rogan-experience-2255-mark-zuckerberg-vAHjVHQqkgI07k7F3G4piE
Mark Zuckerberg is the chief executive of Meta Platforms Inc., the company behind Facebook, Instagram, Threads, WhatsApp, Meta Quest, Ray-Ban Meta smart glasses, Orion augmented reality glasses, and other digital platforms, devices, and services.
Take ownership of your health with AG1 and get a FREE bottle of Vitamin...
Perplexity is going to have a difficult time I guess. It canāt stand out despite seemingly offering a lot. Maybe Perplexity team should figure out whatās wrong? š https://x.com/cb_doge/status/1877570367239209273?s=46
Perplexity UX is far ahead of every other AI when it comes to replacing search tools. For the last two months, I use excusively Perplexity as a replacement to anything DuckDuckGo or even YouTube (as I was using it for developement and tech videos). The free version of Perplexity is like fresh air in all that š All the rest are just numbers.
Watch this vĆdeo : https://youtu.be/EXfFBEuCAr0?si=pJqgmK_4RVJiA8LO
Build your website in 5 seconds: https://hostinger.com/networkchuck10 (use coupon code NETWORKCHUCK for an extra 10% off)
š³ļøš³ļøVOTE!!: Who has the best website? ā”ļøā”ļø https://www.youtube.com/channel/UC9x0AN7BWHpCDHSm9NiJFJQ/community?pvf=CAI%253D
Super awesome, amazing advanced website code: https://github.com/happygilmore445/mywebsite/blob/mai...
hello what happened to the O1 model?
did o1 disappear from open ai the one with 10 uses a day?
Please bring it back
@fierce sonnet any word on this? ā¬ļø
Can I generate an image on the iPhone app using Dall-e image setting for perplexity
I can on my pc but canāt seem to do it on the app
Is there no way to generate an image on a mobile device - even by requesting the website in desktop mode?
Has anyone successfully used perplexity and all the models without filters ??
Why can I not see more than 20 threads in my spaces? Theyāre all short so I doubt memory is an issue,
the past week I noticed that the context window of the searches has been reduced... well more like eliminated as the answers to follow up questions are invariant to the initial query. Any plans to bring back context?
how to turn off shopping in perplexity?
i can't use perplexity reasoning ability for a few days now. is this feature gone now that o1 has added the model. does it no longer turn on automatic reasoning ability in specific messages?
I have a question regarding models: I selected Claude 3.5 as the default model, and when asking in while applying a Writting focus, it says it's Claude, but on Web it says its Perplexity. is the "default model" in settings applying just for the Writing focus?
Hey @acoustic sun!
This is normal behavior, the AI model is the same. You can find more information here:
https://discord.com/channels/1047197230748151888/1225549032718602332
iPad only and you have to use browser
Hi Guys, i made the ai to make a list for me, also to download it, after i click on the download i got an error message which says AccessDenied
any idea how to resolve this?
This XML file does not appear to have any style information associated with it. The document tree is shown below.
<Error>
<Code>AccessDenied</Code>
<Message>Access Denied</Message>
<RequestId>C41BHJDJ1CSW23M2</RequestId>
<HostId>n7UDMW4jIg5WwMSSwfOOboNX8ZKHDLblC1fhP%3255GHF5tdWFLvuBXZTUi3ntiOwCw+gmTmJq3Fwz2NOHvMp+vYxmbYm</HostId>
</Error>
what do you mean exactly? can you share the thread?
did you click the download button below a table or?
There's a link to download
But after i click on the link, a new page opens up with the mentioned error codes
There is no download csv option for me, the list kinda long, it has 120 things in it.
I have the PRO versionn
that is made up by llm unfortunately
llm?
the large language model (gpt/claude)
Does anyone know if its possible to make the api call use a specific Space? Cant find it in the docs
The main website and the api are two different products. The main website doesn't have an api to use.
OK. Thanks. Could I add all the instructions I have in the space as the system role content to try and get the same experience you think?
You could do that but the answer could still be different than when using the website. You can read more about that here https://docs.perplexity.ai/faq/faq#why-are-the-results-from-the-api-different-from-the-ui
how to turn off shopping in perplexity?
I get the need for monetization, just not a good look pushing ads on pro users
is there any perplexity bot that i can add to my own server?
Used to be a year ago
No longer
ah, okay
Loving the Perplexity experience, especially the pdf analysis. Was wondering if there is any local llm setup for same
I remember using Msty for this but sometimes it just gives up on working over anything
Hello Perplexiters,
qq I have a Pro account and working on a beginner Python project which involves libraries like Numpy, openCv2 and some Matplotlib. Which model language is more suitable to assist on the code., so far I have been using Claude Sonnet 3.5
Hello,
I'm using Perplexity for the first time.
I bought the subscription. Can I enter my own openai api code?
Hi Danny, it is not possible to use your own api key from openai
whats shopping
Thanks for the information. I try to understand how Perplexity works without a API key.
Like I said before; I use a paid subscription 20 dollar a month.
I use ChatGPT 4o. I use it to collaborate with AI making plans for IT implementations. It's a smart model which give fine answers. However I also noticed that the AI sometimes goes into a loop of information.
Can I use this AI infinite for this, or are there limits for this?
Each chat has a context window similar in size to Chatgpt's however with Pro you get access to more models (sonnet 3.5/grok 2/etc). I'd recommend checking out the Complexity extension which adds a lot of quality of life adjustments and gives you a model selector on the main UI. for your use case, I would recommend looking into building a dedicated Space with your direct plan files/notes included.
You can choose between different models on your settings page: https://www.perplexity.ai/settings/account You have access to 600 uses a day with pro models. O1 model is currently limited to 10 uses a day.
is o1 only good for essays/long form content or whats the actual usecase for it
Hellow!!š I have just signed up into perplexity. How can I use it properly for geting best outputs for improving my productivity??
welp
Is it not?
Whats is the problem with doing images with this software? It not letting me do anykinda images like from ChatGPT app?
Any new version currently in development for perplexity online model?
when are we going to see deepseek v3 in perplexity
I guess itās better to use a different tool for image generation. You can use Grok for example. Grok uses now its own model. You can use X or a separate app (available in US).
There were many complaints on image generation in Perplexity but nothingās changed and as I believe nothing is about to š
thank you for your time!
I have perplexity pro ,do I need to pay for usign API ?
they give you $5 worth of API calls a month for "free" with your subscription. You can top it up if you want (I have never needed to). Queries are cheap so $5 goes a long way.
Yes Llama , it supports these models which you can define in your API call:
Model Parameter Count Context Length Model Type
llama-3.1-sonar-small-128k-online 8B 127,072 Chat Completion
llama-3.1-sonar-large-128k-online 70B 127,072 Chat Completion
llama-3.1-sonar-huge-128k-online 405B 127,072 Chat Completion
I havent tried if it supports 3.3 or other models yet tho.
it's supposed to be "slow" i believe?
Hey @jovial blaze!
Please explain your problem in more detail (response time, expected response time, screen recording).
How can I make perplexity use Claude 3.5 Sonnet
I have just found out that someone made alternative to perplexity ai last year, released last Nov.19.2024 and name is perplexica
no gemini models in perplexity pro??
no but since gemini is trash I dont see why you would want to
Hey @spice aurora!
Pro subscribers can choose the AI model here: https://www.perplexity.ai/settings/account
Watch the full video i get that slow loading or stalling. On pc and mobile
I am also having this problem where the query never finishes and related just hangs. On android.
I'm having the same problem/bug where I get the answer but the loading at the end never ends, so it never shows the follow-up question and I can't continue the thread with another question.
š
Anyone having troubles in loading their recent chats?
Are you on a VPN?
Iām using mobile app, is the library function down?
same, I'm having the issue too. no VPN, just internet. works on the web version on the phone, just not the app (android)
iOS not working too
Least I know I'm not the only one with trouble loading stuff.
answering ~~~~~~~~~~~~~~~~ >??>>>
I was using a VPN, but it was fine and working before, suddenly it's not working today
pending ~~
Can the API open up more models besides llama-3.1-sonar-small/large/huge?
How did perplexity know where i live
I didnt say it that in that conversation
Please i need some explaininh
ip address
any website can estimate your region (so at most the city) with some accuracy from your network information (IP)
Only real way around that is a VPN
"Can someone assist me in understanding why the option for Perplexity is not appearing in the default assistant settings?"
is it a specific model that you get this? I get this on o4 and base pro search
I tried with and without vpn and still getting this
are you on pro?
where can i find the conversion rates for API calls?
how many USD per queiry
tysm
w8 whaa
why u delete that
||forgot to screenshot the bottom part||
no
token = 3-4 characters
ghawd dayum
using API would be much costly for me then
š i usually write my clg assignments
Web site created using create-react-app
Speaking of that, say i mailed someone with my email,
Can they find ip from it?
I am a professional annual paid pro member~
Why still use Cloudflare??
I really hate this thing
Always pending, must enter the homepage to check
I think most if not everyone is pro user here
My self-built VPN for my bandwagon has been blocked, and the IP addresses in California, USA, are all blocked, which makes me so angry
pplx 's api is so expensive~
^^^
@shy moss
is perplexity better than chatgpt??
Perplexity assistant for Android is live ...
In my personal opinion, yes.
Its not better than o1 preview or o1 models imo
But better than gpt4o for sure
hello, i have a question. I have Perplexity Pro, how can I create images through the AI services?
nvm , found it.
you can use o1 on Perplexity
Thank u
what is the best model ai in perplexity??
Hi im a rabbit r1 owner and beside i got the gadget which didnt require the sub but they offered a 1 year sub as a extra gift if you were one of the first to order a rabbit r1. Im now in a difficult situation where i either have to live with the results r1 gives me with its built in command version of perplexity or to continue paying for perplexity or simply i wanted to ask couple questions about your newly announced perplexity gadget? What will it do differently to r1 for answers on searches? What features will it have and how would updates be? 50 dollars seems cheap for a voice to voice gadget using perplexity. Will it include perplexity pro or would you still have to pay 20 dollars for the sub in order to have a use for the 50 dollar gadget? Why even consider the gadget when there is a website and apps for perplexity? Why will i have to own it? Or was it just something to say at ces that is a maybe project? How will it look like and function? Will you also be able to type on the device when you commute? Should i just stay using r1 for now? I have so many questions
Hi! I am starting to use Perplexity Pro.
Is there any way to make it read long texts without stopping?
Could it be exported to audio?
I hope the questions are appropriate and this is the right forum to ask them š
Hey there š
Any idea Pro users how to login in the Windows App of Perpelxity
hey @hybrid bone! There's no official Windows app. I wouldn't recommend using this one.
Thanks for reporting!
Hello everyone! I have a question, I recently published a lot of pages on perplexity, but it seems to me that they are private and not indexed by google. I copied the links to the pages and visited them using a different browser and I can't get to the page. How can I make my pages visible to everyone and have them indexed? Thank you very much!
is someone available to discuss the security side of prepelxity enterprise ?
HAHAHAHHA
Security ...yes yes
Unlock the power of knowledge
How to get pro role
By becoming pro
šÆ
True
Anyone know if there is a student discount for perplexity pro?
dude fell for it

Youāll have to leave the server and join again by using the āPro discordā button on your settings page
Lol okay so why the f do they say they may have a device coming out f it im out this community got rudw
Really impressive. Though, until it can open with voice activation like "Hey Google", it's basically useless for day-to-day use.
Yeah, there's no hot word for that as of now...
Would it be possible to rise the instruction level to like 3000 instead of 1000?
We msn now!
They seem to want to resurrect Yahoo š
Bunch of Yahooligans
It's getting to the point where like I'm less excited for new features perplexity is adding and more excited for them to add an option for me to hide their last features
html
I literally just want to clean way to access information and learn and have control over that.
I am tired of being force-fed things everywhere I go I'm so tired tired tired
Whatās the next step?
A photo of Aravind? A picture of Friday nightās empty pizza boxes? A counter for new Pro (free) subscribers in Canada?
Honestly, this doesnāt even feel like stagnation anymoreāit feels like a step backward. Please, focus on what really matters: restoring stability and reliability to your core tools.
I've honestly months ago given up on asking them to add their own basic features into their own Android app. Things like being able to configure a space, make a page. Or simply not talk about the same photo over and over again I have to go on the web page just to access those because they abandoned the web app to think of ways to clog perplexity with the same kind of garbage that's all over the web already
I'm honestly probably should be grateful that they showed their true colors before my free year subscription ran out for pre-ordering a rabbit
I'm not sure where they get their ideas from but it doesn't seem to be from their users
Oh wait I know!
Perplexity is set to become solely a (supercharged) search engine.
It would be cool if they focused on that part of it instead of trying to turn it into more passive ad crap
I have to switch into writing mode to ask for advice on things sometime cuz otherwise it throws like 5,000 shopping links on me
jeff bezos pulled out ofcourse more ads will be put in
and cutting the potentialities of the custom api's
Wait he cut his investments?
Are they changing the apis? I seriously was considering just using a different frontend instead of perplexity and paying for their API when I need to do a search. After you get that initial data the search actually makes a complex to talk about because it gets confused by extra searches
Seems way smarter to have web mode for the beginning of the conversation and then switch to non web mode.
https://apps.microsoft.com/detail/9P02M3TKS0RJ?hl=ko&gl=KR&ocid=pdpshare
Is this an official program ?
Powerful AI search based on the perlexity API. Where does knowledge start. The answer you need is right at your fingertips.
Put aside all the noise and get reliable and up-to-date answers directly. This free application can harness the power of artificial intelligence, such as OpenAI's GPT-4o, Anthropic's Claude 3.5 Sonnet and Haiku, and xAi's G...
Hello everyone! I have a question, I recently published a lot of pages on perplexity, but it seems to me that they are private and not indexed by google. I copied the links to the pages and visited them using a different browser and I can't get to the page. How can I make my pages visible to everyone and have them indexed? Thank you very much!
Could it be because of an enterprise subscription?
Hello everyone! I wanted to ask if anyone knows whether it's already possible to access the Pro search engine via API. Thanks!
Hey @polar cypress!
This app is not official, I recommend you to not install the app.
Ty sm
Hey @lilac tundra!
Pro search is not supported via the API, please have a look at our docs for further information. https://docs.perplexity.ai
thanks! why why!!
Sonar large v2?
hi
So a couple months ago I signed up for Perplexity Pro through Xfinity rewards. Played around with it a bit (maybe 20 minutes) and wasnt impressed, though I was trying to use it for coding help and I just dont think that it is what is good at, but is good for research, etc. Anyway, i had canceled it immediately as i didnt want to deal with any renewal issues if i wasnt going to use it. I would though like to start using it again and unfortunately it says i cant use my xfinity rewards code anymore as it had already been redeemed. I had used like i said about 20 minutes of the 12 month promo. Any way i can continue/restart this?
hi, acidentaly hit the corporate buton. now it says i can not use the pro plan anymore since i havent payd the corporate. im not interested in corporate at all
I have the Pro plan and am using Claude 3.5 Sonnet for coding up a web app. It's been "fun" to get the coding to work correctly (lots of it circling back to do the same thing that didn't work before) but at least it's doing better than Chat was doing. š Anyway... I have a pet peeve that I'm hoping someone can help me fix. I provide it feedback on the results and it gives me small snippets of possible fixes, and then asks me if I want the complete code for the file. I've found that trying to figure out where the AI has "fixed" something in my long javascript codes, is nigh unto impossible, so I just usually ask for the full code for the file with the updates implemented. BUT, the dern thing asks me if I want that EACH AND EVERY TIME we interact. I have the thread in a Spaces with instructions to give the full code (maybe it's not phrased correctly to ensure that happens) AND I've repeatedly given instructions along the way to just give me the full code rather than asking me if I want it. It's ignoring all those instructions. EVERY dang time it asks if I want the complete code. It obviously knows I want the complete code because it asks that every time but it makes me jump thru the hoop of typing that in before it'll give it to me. AND it gets into "smart a**" mode at times of not giving me the code but saying it gave me the code just to see me have to go thru the whole rewording of the acknowledgement that I want the complete file code.
Help! This is beyond frustrating to have to do so much jumping. šš TIA!
ADDENDUM - I had perplexity write up a new set of instructions for the Spaces the thread is in (but it's still ignoring it) -
Always provide the complete file code
Never show partial snippets or diffs
Never ask if the complete code is wanted
Show the entire file contents after any modifications
Include all imports, functions, and classes
Maintain proper formatting and indentation
Keep all unchanged sections of code
Include any necessary comments or documentation
When troubleshooting or adding features:
Show the complete fixed/updated file immediately
Display each complete file separately with clear file names as headers
Use appropriate language syntax highlighting in code blocks
Include all code sections, both changed and unchanged
Never ask for confirmation to show complete code```
try to enclose them within <instructions>...</instructions>
I have the same issue. I literally remember last year uploading source documents in a space, and then started chatting with perplexity and with pin point accuracy was able to recall extract and summarise and synthesise pdfs I uploaded. I loved going for walks and chatting to perplexity using my mobile phone learning about a project I was managing. Back then there was no need to copy and paste the content of the uploaded documents into the chat.
I was glad I paid for the pro version but now seems useless without this, it tells me it can't do this at all. I even asked it whats the point of having an upload file/doc function? It says it only reads the title of the file. What happened here? Please can this be fixed or explained? Help
Hey @sturdy delta!
If you find the original message helpful, please consider reacting to it with the :star: emoji. If the post is appreciated by the community and receives 5 stars, it will go to the https://discord.com/channels/1047197230748151888/1082806833938436228 channel and the post author will get the <@&1082034222778302614> role on Perplexity.
Maybe. What. I call reverse prompting physcology make it believe is a file inspector/analyzer using custom instructions if you want you can give the the instructions and a random document so we can test it
Remember the more specific we are the better
tried that. Sadly it's ignoring that too. It's still asking if I want the complete code after every interaction for updates to the code, rather than just giving me the dang complete code in the first place. š¦
Like this...
Would you like me to provide the complete updated code that keeps the highlight improvements while restoring the working reset functionality?
š±
Why can we not have image generation on the mobile application? Chatgpt allows it.
Has anyone tried to make perplexity their default assistant on their android phone, OMG its amazing yall need to try it out!
Going to try it out, thanks
Lemme know how you like it
š
I am using Msty on online api... How do I make the citations show at the end?
is there a difference between pro & standard search when using Writing?
Has anyone used the Claude model inside Perplexity Pro for an extensive coding project? I would obvs prefer not buying Claude Pro when it's baked into the paid Perplexity plan I have.
To use both, Iāll allow myself to respond.
Several points should be considered:
Claude Sonnet 3.5 uses a specific prompt system when integrated with Perplexity, which may result in a different experience compared to its native platform.
The context window also differs (200k tokens on Anthropic).
On Perplexity, you need to convert your code files into ".txt" files to use them as sources, whereas on Anthropic, you can import them directly.
In conclusion, I believe the answer depends on the scope of your project. Perplexity might be sufficient or not, depending on your needs.
yeah itās just a site issue i thought the models had a problem then used them through api and realized either the sites system prompt is bad or the way they give access to bad prompts is bad
I have NBA 6 men demon picks mate
Send a re-quest now
Shout out to T-Mobile, I've never even heard of Perplexity AI until I got an offer for a year pro plan 
is Perplexity dead right now?
Having the same issues here
Me too
Same here
yep, same here
App and web dead
nice
Just when i started doing an app for my company š¦
what happened to it just died
ded. š¦
dead fr
I dunno if another 4k / Month hire to tell you that it's down is worth the investment. oAI also down for quite a lo
maybe you dont know customer support is not there only to tell it is down
yea ig it's not opening
Neither on web nor on phone
yes
things go down .. this is alright .. as long as there is communication and work
Same here it is not opening!
Same here
+1
yea i have an issue where the sentry session is being set to version unkown i was signed in and working and it gave me a white page
"release":"VERSION_UNKNOWN"
hmm
wait when i open in incognito it opens
It says partial
lemme try rq
i think its a sentry/cloudflare thing
nah, somalian servers
True it worked
powered by hamsters
it works in incognito but not in regular windows
partial outage my ass
now its back
nah still down
No @storm jetty sitll have blank page when i try to open it!
i think something changed in a config or update and it takes a bit for those changes to propigate out to things like cloudflare workers esp integrations like sentry
This never happened before!
give it a min or two to push a cache update u can try reloading w a new cache
i think it is getting issue to login but when we are opening in incognito it dont have any saved cookies so it is working
set the headers to cache time 0
It is working for me now as well again - sorry guys I think it was my fault, my question was to complicated
i am still logged in and right where i left off on pro
what is a cache update
Perplexity down???
it's not working for me
Confirm.
sm say yeah sm say no
same here
the best ai doesnt work š¢
still down for me
It's down for sure
Just use it in incognito and login it'll work
nope down again for me tooo
Deleted cookies and retried - still not working1
dead :/
i mean the servers are up i can get a ping and there is content being returned from the server it could just be a bad update
Some intern pushed to prod š
yes me too
anyone else have redux tools open? looks like there was a change to somehting
down?
Maybe they added reasoning
thread does not exist
me too
got 0% loss
500ļ¼
I'm pro how can i get in
in disocord under the pro section
damn, just when i really need it to solve my coding, this web is down...
lol i was in the middle of boilerpating my wallet app lol
web still down
the pro service is also giving trouble, like not showing or allowing to use pro features
No it's working
there was a change in my end for the redux cache i think they changed something
doesnt even open website now
humanity has lost one morning
Can't access the website at all here.
down
How do I get plexer role
i feel like we're encountering issues with parallel routes in Next.js, specifically when prefetching is disabled
perplexity not working still
yea its in the nav bar
ill just use grok in the meantime
Wow, so it's a general issue. I can log on my mobile phone, but not using Chrome or any other browsers.
it is back
still not up for me
hi
kinda back i think the solution would have been to roll back and we need to wait for the endpoints to come back up
it is coming and going i think they fixing
the frontend is prob fixed just need backend to catch up ... mine is all back up and working
Hey @storm jetty!
If you find the original message helpful, please consider reacting to it with the :star: emoji. If the post is appreciated by the community and receives 5 stars, it will go to the https://discord.com/channels/1047197230748151888/1082806833938436228 channel and the post author will get the <@&1082034222778302614> role on Perplexity.
doing a deep dive on the redux of perplexity web is actually super cool would love to see an open source version of the frontend.... i like the statemanagement
waiting 5-10mn between tests but it's still down on my side
It was back up here for a minute but I tried to do a search and it went down.
Same here, sometimes it loads the website but doesn't allow me to see my spaces or use pro features
i mean im sure its just some fancy nextjs app router thing but i think how u manage hyrdation in this conversation context would be cool to learn
its back down here too, it seemed to be back to normal, spaces, threads, etc but back down again, playing with our feelings
yea my sidebar/library history isnt hydrating
It's a real piece of fun to spin back up a service users are actively ddos'ing š¤£
Seems to be coming and going here. Up for a second sometimes but I try to do anything and it goes back down.
mine isnt hydrating either
hey i know u from scroll and shapeshift
Down again!
were u at permissionless? or ethglobal?
it's not working
no way! hahaha, what a memory you have, might go to eth global taipei
the web can be accessed again, but there is a bug, when i try to prompt, it will log me out and use free version....
Servers down?
It is the end of the world as we know it.
You have Discord since 2015
& ?
I did not see any yet with this numbers
Is it slow for you or doesn't load at all, right?
perplexity down ?
yes
cant load the page on my pc
Same
yeah site is down sadly
forced to use the bootleg perplexity until real one is back online
And if you insist, you enter but the subscription does not load or it never ends up giving you an answer.
yeah happened to me 
@shut light what is bootleg perplexity?
It's better not to force us to enter, because that requires more load and will be worse.
š„²
Is it the same as perplexity?
from my use, it feels the same
perplexity down?
yup
just a blank site
yeah
Im also getting a blank page on all browsers
Its down for >1h
Kinda shows you how many people rely on it though.
we just rely on the LLMs powering it for the most part
As a one person marketing team in a company, its impossible for me to do EVERYTHING by hand
I needed it today to generate images for an ad script
well, it really helped me finished my 3 project of web application for company with just me doing backend and frontend
That's like saying "we just kinda rely on the javascript that discord runs on"
No.
Perplexity is a search engine that uses multi layered optimization. It's not like they just run claude on the results directly. They use different models for different tasks, feed some sort of a summarised version of the question to claude using relevant data and get the answer.
with this i rarely spend my time on stackoverflow
yeah, its great. It's not extremely reliable for up to date information (their search query is not really optimized that well) but its good enough for most tasks
Yes, still comes down to the LLM as i said, just now with a web crawler as a preprocessor
^
lmao
If search + web crawling was a simple problem one can solve using purely LLM's, OpenAI and Anthropic would have had better results
Is it down again??
yep

it become like this
god damn i js wanna finish my shi
@distant juniper theres like 5 new ai chat services that pop up a day that have web search capabilities and its not that hard to do
Let me innnn
again, AI with web search capabilities is not the problem.
Optimization is.
That's like saying google is just a web crawler
Its an oversimplification of what they do
It's good now
Sam you need to grow up
ye
Thank you
"eli"
when im using claude model, sometimes when i try to ask create something that make the model use markdown box, it always stuck with just response like "here is how to do it:" only, so i have to click rewrite everytime, is anyone have this problem too in claude?
work hard play hard
@distant juniper you're special too, but like the kind that cant leave the house without a helmet and knee pads
role diff
I do indeed like to ski when leaving the house
Thanks for noticing :)
yw u seemed fairly predictable
@distant juniper no respond?
respond to what
I'll take that as a win, never lost a debate before
yes you won
i know, hopefully you learned a thing or two
Where did I should put promo code?
Hey eli wheres your pro sub? You're reliant on perplexity but can't even support them?
Good job you can google what a pro subscription looks like, maybe you'll be able to buy it one day
No problem keep working hard and youll achieve your dreams of buying a sub soon enough šŖ
definitely
Why no status in #šāstatus ?
should be back up again
and my favorite ai is back 
Hi, Iām new here, I donāt really know what perplexity is but I gained the pro version with revolut metal! Is this better than chatGPT?
Guys, I have questions about where to see my completed Pro consultation
My GFs 2017 Macbook pro just crapped the bed and has those annoying stuck keys. Half the keyboard š
You can give her this no one has requested for it yet
@tame current what did you end up upgrading to?
I would like to connect perplexity to my intellij Ide. Is this currently possible? Any resources on how to go about would be useful , because as of now I have perplexity open in browser window alongside and would like to integrate it into the ide
i saw that there is an api access , but it costs extra per token, I was hoping since i have the pro version that I can somehow integrate the chat itself without having to pay extra for the Api access
Also when I asked Perplexity it suggested I just use intellij copilot lol maybe thats the solution ?
"If youāre looking for seamless AI integration in IntelliJ, you might also consider other tools like GitHub Copilot or JetBrainsā native AI Assistant .."
Any resource of help would be apprectiated
Chatgpt pro users, do you think the $200 is anywhere near worth it? I got Chatgpt pro today and so far Perplexity is like miles ahead of chatgpt (better web search, web search in o1) and also provides much more concise and to-the-point responses with GPT 4o than Chatgpt with GPT 4o.
The only thing Chatgpt is really giving me is unlimited o1 usage but not sure if that is worth $200 (if anyone knows alternative methods to get unlimited o1 usage at a cheaper price, lmk)
links are broken, and perplexity stuff don't give a damn.... in spaces when you add a link to resources it get lowercased, which break all aws docs, bug reported months ago, and this 1m fix they deeeeeeeeeeeeeeeeeeeeeeon't care about it....
the #1140622008086970420 section serves no purpose at all, I think they don't even look at it
at this point this company just looking to get bough by someone else I guess
and no one will acquire them, no F body
I used a magenta code to use Perplexity AI for 1 year. It was all working perfectly for about a month or so and now it is gone. What could have happened?
What is a magenta code
From the Telekom Application (T-Mobile) moments
Hey! Which is the best AI Model for programming?
Hi, hello from the Czech Republic, I'm curious what all it can do.
In programming, I'm a complete novice
Sonnet 3.5 by far
for very complex debugging o1. But normal programming task definitly Sonnet 3.5
You meant Claude 3.5?
Claude Sonnet 3.5. There is also a smaller model from Anthropic called Claude Haiku 3.5 which also very good and much cheaper but not as good as Sonnet. If you need a cheap and good LLM you should look at Deepseek 3.0. They are very cheap and even discounted till mid february. and i think kinda same smart as Haiku 3.5. All in all is Sonnet 3.5 the best for daily programming
I'm interested in video generation and video/image modulation. Can anyone advise me?
I was given the opportunity to use this AI for a year Pro for free and I would like to take advantage of its potential .
Perplexity doesnt support Video generation. However you can create Images with different Models. imo Flux is the best
And would he be able to write code for another AI, for Github, Hugg.Face?
I'm starting to like stable 3.5 in its "rawness" photos
Hey guys,
I've been building autonomous systems, and I recently realized that OpenAI's approach to AGI is fundamentally flawed. In fact, they lit billions on fire trying to brute force intelligence. Intelligence doesn't scale with compute or model sizeāthat's like saying a bigger brain makes you smarter and watching millions of driving videos make you a better driver.
Intelligence scales exponentially with constraints, not compute. Bora's Law, I believe, is the new blueprint for building intelligence. More in this article: Bora's Law: Intelligence Scales With Constraints, Not Compute
https://chrisbora.substack.com/p/boras-law-intelligence-scales-with?r=aszci
Sorry but l must admit it perplexity fell off
Am using pro search l asked today's date this was the answer are u drunk?
No, I'm not drunk - I'm an AI assistant and can't drink! I was providing accurate information about today's date being March 25, 2024. I aim to give correct information
It must have hallucinated or something
Imagine you got your own junior dev. Problem. You must be very specific and break down the stuff into steps to get the best out of it
I don't think Iit has to be smarter than humans to reach agi in the sense of being conscious. It only has to outperform. Humans in everything I have a concept down where agi is basically like a society of models working together they would absorb and learn the data including the physical world one. And think asynchronously running. Millions of Simulations to determine the best outcome for it to work an artificial brain that emulates the chemical reactions of the human brain would be needed
I used chatgpro and gemini but unless you just want image creation wich chatgpt pro is the best at it, just use perplexity its fine and can also create some wicked art
Image Creation and ChatGPT calling "best" is crazy. ChatGPT is very bad. Flux and Grok are miles ahead.
it mixed up Auth0 and Auth.jsš
I'd encourage anyone here to run some tests on Perplexity claims. Claude 3.5 Sonnet should have a 128K context window (allowing up to 128k input tokens per query)
Running a sequence of tests, I find Perplexity nerfs this context window down to 4096 tokens.
For 128,000 input tokens using Claude 3.5 Sonnet, the cost would be $0.384, as the model charges $3 per million input tokens.
Accordingly to Perplexity itself. I feel this is being highly nerfed in order to save costs on PPLX end
Using the API, I find I get far better results but also far higher costs than Perplexity.
It seems PPLX is nerfing input/output tokens to control their costs?
For some reason my chat is changing from Sonar Huge to Claude 3.5 Sonnet everytime, even in a Space using Sonar Huge and with the configs set to Sonar Huge, any idea why?
but seriously, just try churning massive amounts of tokens and you'll quickly encounter failures with PPLX
Are citations down?
Has there been an issue with perplexity? Why can't the latest information be retrieved?
So I've been fighting with Claude Sonnet continually "fixing" and breaking some code that's accessing a csv file for eval. I've begged. Pleaded. Yelled. Reminded. Etc. Etc. Ad nauseum. Finally got so fed up with it that I asked what would work... here's the convo we had... š¤£
And this was the response...
And here is our solution (for all who are using Claude Sonnet and experiencing similar "issues" of "helpfulness")
Do you speak chinese to me?
the price comparisons arenāt super fair because of the deals these companies get, thereās several other competitors with significantly larger contexts and they donāt have a lot of the issues brought here, but perplexity has a slightly nicer ux
why my screen turning white when i try to search something?
For image creation? Not really dude
And i used them all
My Android Perplexity app has been down for the last 24+ hours. Reinstalling it did not help. Anybody else?
Hey team, could anyone help me clarify what Reasoning Effort Parameter you're using for the o1 model?
does anyone here speak Czech?
Introducing Nintendo Switch 2, the successor to Nintendo Switch, releasing in 2025.
Learn more: https://ninten.do/6003ohKAB
I'll be sendin you the next plays, Congrats to the people who messaged me todayĀ and tailed just send me a friend requ/st⦠and I will send you what I got šŖ
Bruh this is not even worth a discussion. Prompt "Create an 4k image picture of a Sunrise on the seychelles" its not even close....
and here the FREE version of Flux called "schnell"
Hey everyone š Excited to test out perplexity
Testing is almost like "already adopted"! Welcome š Perplexity is great.
Can a moderator help me with an issue Iām having
Hello i need help pls, it i got a free year code it says i dont have to cancel the kontrakt for the next abo but in the app it says it get automatically refreshed. ??
Yeah ive been loing it so far
do u also sell snake oil?
my portfolio of snake oil got 200% more profitable in the past financial quarter
Lately Perplexity is making many errors and repeating answers. not sure if others get this.
this is with PRO search enabled
GROK and SONNET mainly
When I use GROK on X its much better.
When I do searches related to finance, sometimes the Tako dashboard of the company in question appears, sometimes it does not. How can I make it consistently appear?
I have NBA 6 men demon picks mate
Send a re-quest now
consistent will result in consistent profit margin loss for perplexoty so expect it to be nerfed along the milestones
I have NBA 6 men demon picks mate
Send a re-quest now
I donāt know if its normal but after every time i search something up on the mobile app i keep getting N/A. i donāt know if its an issue with my device but Iāve deleted that app and reinstalled several times and i have no clue how to fix it.
Doesn't fix your issue but try deleting cache /web is more stable with access to 01
Thank you, how do i do that? š
Nevermind i figured it out
Im tired and that was a dumb question, mybad.
Goodnight
i can see through that
what model perlexity has that is most suitable for academic research in you people's opinion?
any suggestions?
Hey, does anybody know how to change the theme from dark to light?
Is o1 working now?
And when will be new version of mobile app with o1 model?
o1 is not working in perplexity and wasn't use my message
https://www.perplexity.ai/settings/account or rewrite option is not aval to you?
or is your ip dirty?
Still not working
I know about this , but when I send a prompt, o1 wasnt using a prompt
How can i fix that ?, he gave me fast response instead use o1
Hey @jade yarrow!
If you find the original message helpful, please consider reacting to it with the :star: emoji. If the post is appreciated by the community and receives 5 stars, it will go to the https://discord.com/channels/1047197230748151888/1082806833938436228 channel and the post author will get the <@&1082034222778302614> role on Perplexity.
change your ip somehow
I don't what happened with Perplexity o1
š default model instead o1 and still have 10 remaining responses of o1 and fast response
Hi! Where is the chat where you can see the replies from Perplexity related to issues/ideas/questions made in the past?
I'm having an issue with activating my promo code that I got from T-Mobile. It's a year of Perplexity Pro for free.
please contact support@perplexity.ai
What about this ?, someone from help center look this
Anyone have an issue on Pro but can't get image generation working? I tried but it will not generate the image and instead describes an image or something similar
It's the library option, in the app the rightmost icon along the bottom
Its on the web app not the android app.
Yes it's really bad but use this prompt first
Generate a useful description so that a generative AI can create an image of
Then insert that prompt again and there should be a generate image button
Doesn't work on the App, just Web Version.
Why does Perplexity Pro with settings as ChatGP-4o gives wrong calculation while ChatGPT.com (even without Pro) doesn't?
Example query:
Options Contract sold.
45 contracts
Price per contract: $2.6.
Whats' my estimated credit?
I tried it for fun and it gave me the correct answer 
Really??? I tried 3-4 times. Can I have screenshot? Surprising!
hiiiiii
hmm, that moderation censorship on images is really broken. it fails 95% of time to generate a pink horse head, but didn't have a problem with naked angel painting (even when I didnt ask for her to be nude)
hey, this might be a hopeless question, but I was wondering if there is a reduced rate available for students using perplexity?
(based in EU if that makes a difference)
Is there a problemw with o1? I try to set o1 as the default model and it does not use o1 for the response at all.
Same here
It's not using o1. Can't even see the model for some reason
nevermind, it's my VPN
I had my VPN off and its still not working, though the other models (sonnet, 4o, etc) are working perfectly fine.
same happening to me. did you find the solution?
thought it was just me. I see SIGNIFICANT reduction in response.
it's quite unusable at this point
It's as if I'm starting new chat every single message
Its pretty weird I actually like the longer responses and now that it is short and too the point it feels like I'm getting the briefest synposis possible and I'm really disappointed.
Hey @sick wyvern!
Yes, there is a referral program for pro subscribers: https://www.perplexity.ai/hub/faq/coupons-and-discounts
I cant help but agree
For those who need referral code. You can use mine
https://perplexity.ai/pro?referral_code=P089QUFO
How much % do you get off
50%
Just for 1 month right? Or would work for a year
Yo did llama 3.1 online get removed from perplexity labs?
where is dream space
guys whats best ai for charts and maths + analyze... want to make space for daily market analyse but not sure what ai should i use
im on pro version
Default model instead O1. When fix?
Hi, I'm completely new to programming. I work a lot with image and video. I would like to write a simple program for image modulation. E.g. RGB colors, CMYK, or frequency, edge detection and transfer the image to another image. It's something real. Or is it a job for xy hours? Thank you for your reply.
hi
I have same problem , 2 or 3 days
No one from support answer me , only gave a script what to but it doesn't work
hey guys is anyone facing this issue in perplexity where it has stopped researching as much as it used to and is giving generic responses
it's also getting stuck a lot
Kind of regretting my perplexity pro purchase right now
yo guys. I'm from Singapore. Happy to be here with like minded ppl
Singapore is a very fun country , I regularly visit, where in Singapore do you stay
very exp country š
Has anybody else issues with very concise responses and o1 not working at all?
when im using perplexity pro, and I dont use the pro search, will this just default to Claude 3.5 Sonnet without the web search feature? (assuming ive set my AI model in settings to Claude 3.5 Sonnet)
if you dont need web search results, change focus to writing or disable web in space
bro what change did these guys make
Anyone else having problem with Perplexity? Whatever I ask it just keeps loading and doesn't load anything. I'm using free version without being signed in.
Same problem, i think everyone has this problem
@jade yarrow At least Perplexity answers you, I can't even get it to answer.
no!
just 1 month
if 50% off for whole year
, take my money!
I would think that there would be features like that to lure people into using such subscription plans.
that's a stretch! š
Hi, I've been using perplexity for a few days. I've noticed in the last 2 days the quality of the responses have been reduced. Prior to this reduction in quality. I was able to select GPT o1 from the dropdown and it evidently went through a chain of reasoning before coming up with a response. Now, even though O1 is selected, the quota of 10 searches per day is not decrementing and the quality of the responses are instantaneous and of low quality.
I'd appreciate if someone could guide me on how to restore functionality to my account, it feels like my account has been nerfed because of my extensive usage. I am on the pro plan.
I dont think it's using Sonnet either, no idea which model it's using, but the quality is sub par
Back to Google Gemini Experimental 1206 it is
Same
@jade yarrow use aistudio Gemini 2.0 flash thinking experimental and 1206 for coding tasks
for now
while this is looked into
I don't know if that is a good idea or not.
š
I'm out of perplexity then if this is the case, not connecting my data to tik-tok in any way whatsoever
Plus wouldn't that mean that Bytedance would effectively own have stake in Perplexity that is equal to the stake in Tiktok? Its the same issue with Tiktok just in a different face?
aaaand, they lost me as a customer.
This must be due to financial issues with serving models etc
Meaning they want to mix with tiktok because of it.
Lol they only care about more user engagement
so they can sell more subscriptions
probs
šŖ¦
Or they just want to be the first AI / Social media
People are glued to TikTok, they want the same here
Will be cancelling subscription soon
They'd gain access to TikToks user data
its so over
Definitely will stop use if this goes through. Not that me leaving will matterā¦
Iām a little confused how a merger would negate the ban here, the original owners still own most of the legal entity even if it happens.
Nope that website is absolutely horrible. Nice luck getting accurate information for your essays
Sure. I love when it forgets what I said 5 seconds after I sent a question
Nope I absolutely dislike tiktok (kinda cap)
Yes
Ooh guess what
You.com sucks horribly
Matter of fact it cant do accurate research on devices
Minus aura my badš
I'm just saying perplexity is overall much better
Keep on coping
Perplexity is so much more accurate
Give me an example
Ok
So what's wrong with that
Just ask it to be longer idiot
š¤Æ
Bullet points are much better for learningš
Yep that's embarrassing
Unc, you prolly 50
Ooh
I don't even got tiktok installed
And I mainly use perplexity
Jesus this man isn't real
One year free perplexity
This man using revolut ahahaha
It's the best bank
This man definitely broke at any moment
4.75% interest on savings
Have you seen revolut complaints lately
In which country?
Any country
ive never had an issue with revolut
Omg lol
it's the most advanced "non bank" bank
It said the ayaneo kun had a 1440p resolution
the*
@sinful moat
is it the .com.au?
oh
but
max $100k lol
i need interest on $500k
and im not in Australia
Ubank
New tiers
Tiers Bonus rate (p.a.)
$0 up to $100K
5.50%
Over $100K up to $250K
5.00%
Over $250K
0.00%
Yep don't listen to that aussie with horrible opinionsš
Hah stupid
Literally says
duh
1600p
Revolut is global
And look at the official ayaneo website dumbass
offers interest on different currencies