#💬│general
1 messages · Page 74 of 1
A search engine that gaslights you
Lmao sometimes
But I think Perplexity, out of all the LLM and LLM based AIs I use is the one that hallucinates the least
It's very straight to the point
Yeah, it's fine if you're just searching for like general information or something that's not like a very new low SEO type thing. Like if you search for what is the latest update to x or y program it just freaks out.
I stick to Linguistic stuff most of the time. You should also ask other chat bots the same thing to verify information. I don't understand how people let A.I gaslight you.
Lol true!
I used to use it for assignments mostly
Typically I just use it for programming stuff like oh how do I do this or that in this language? And it works just fine for that. That's really all I need
Just saves me time from searching through stack overflow
Claude for summarizing lecture material, Perplexity for doing further research and ChatGPT for breaking the information down further and expanding on it
That was always my order
I find it to be rlly good with humanities stuff, but I don't know about coding
Not withing my usage case lmao
The extent of my programming knowledge are .bat viruses
"viruses"
Lmao more like a lil nasty script that would turn off your PC every time you turned it on
I had one that every time you'd close the window it would reopen itself unless you dug around and task manager to find it
In*
Lmao that's vile
But yeah that's the extent of my knowledge lol
I did want to learn how to code but I had this one teacher that made me hate coding
But now I usually just do more tame stuff like little toy projects in Python or Go
Lol
Doesn't mean I don't like tech stuff though
But only the conceptual part lmao
Don't make me write code
I can't claim to be a good programmer, but I'll have an idea at 2:00 a.m. and jump out of bed and brute Force my way to completing it
I think that's how most people are with any creative process
Mhm
And of course it happens at the worst time, when you need to sleep
Good thing ChatGPT is a thing. It always helps me brainstorm
I kind of use to expand on ideas I had
Or come up with new ones based on ideas I already had
That's THE best time to have ideas tho
That's when your brain is crunching everything you've learned
It enters fuse mode lol

Cus you're tipically very relaxed as well
That's why shower thoughts are thing
Me whenever I have a new idea
Bro i could never learn how to code. I would get so caught up on the little details I would ask chatgpt what does the / meand and ; every 5 seconds
Lmao
I feel you
Luckily I learned a majority of what I know before AI took off. If I started trying to learn now I'd probably use AI as a crutch too much and never get anywhere
Imo AI is just the evolution of going into Stack Overflow and copying some random block of code you don't even know does lol
I think you underestimate how lazy people are, " chatgpt, solve this code golf for me"
Lmao they're in for a shock then
Idk how Open AI o1 handles queries like that, but even ChatGPT 4o sucks at that
I was going to say it was hit or miss, but it's more "miss" than "hit" lol
I think ChatGPT is better for creative stuff
Yeah dude like maybe if I had a run down on ":;,*<_/' does within the code maybe I could figure it out.
The other day it made me an 80's playlist I loved
Yeah, usually my go-to test for new llms is to have it program a self-playing game of snake. The only ones that have done any good and actually been able to create it have been claude 3.5 sonnet, gpt4o, and now o1
Did it make up a song or two? Sure, but it was good.
Oh yeah Claude artifacts rlly come in handy for that
Lmao
Here is what o1 came up with (I have gotten better results with Claude 3.5 Sonnet but it took more prompting)
Oh damn that's actually crazy
Ai making snake "AI"
This was the process, it's pretty simple: https://chatgpt.com/share/66f03054-8c14-8004-bda0-c778976e0b43
💀 bro that looks so confusing
How do keep up with the syntax?
Must've took years to learn
Can you do linux?
That's actually not me at all. That's all AI generated code. I have llms make this self-playing snake game as a sort of test for them.
I don't actually know JavaScript that well at all
The languages I'm familiar with are Python and Go
Ohh
Is there a way to use Perplexity in VS Code?
Did you try asking Perplexity?
Combined perplexity with ChatGPT and created a bot called PerpGPT
so, every time i ask "what language model are you" it says perplexity ai. this is even with claude or gpt selected. the only time it doesnt say perplexity is when i use reasoning, in which case it states it is openai. this makes me suspicious. is perplexity even using these other ai models when you select them? is there any other ways to test this?
Hi, If i am looking to find the copy right of a list of organization, what is the best AI model to use?
Dear Perplexity Team,
I hope this message finds you well. My name is Iliou Theodoros, and I am in the process of writing a book that will include various educational resources and ai tools for readers. I am particularly impressed by the features and functionality of the Perplexity platform and tools, and I would like to include screenshots of your platform and tools in the book to showcase its capabilities.
As the book will be published for commercial purposes, I am writing to kindly request your permission to use screenshots of the Perplexity
tools in this publication. I believe that showcasing Perplexity
will be beneficial for my readers.
Thank you for your time and consideration. I look forward to hearing from you.
Best regards,
Iliou Theodoros
How can I generate images, I'm using the pro, prompts like generate an image doesn't create an image, instead seacrhes for ways on how to do it
Hi @calm phoenix, this guide will show you how to generate images using pplx https://discord.com/channels/1047197230748151888/1225557012532695060
Hi
What’s the context window for 3.5 sonnet
Hi - curious how does Perplexity decide if an answer is right? I ask because I got some wrong information regarding schools for my son
Does anyone know?
it can't
never trust Perplexity or any kind of so-called "answer engine" with fact-checking. Either you have to use specific prompts to force it to quote the exact content in the sources, or you have to double-check it yourself
Some of it is community driven with feedback but I assume it curates what search engines provide.
So it's working like a search engine of search engine results?
Always wondered how it determines the "right" answer. Note that I am referring to questions that are supposed to have a right answer, not some rhetorical political BS. 😂
Probably based on how many reputable websites say the same thing
Hello, How y'all doing!
hi
Is there any api of perplexity which can support RAG over internal websites. Can anyone point to api documentation about it.
can someone help me make a prompt to actually get the important things out of my readings?
I have tried variations and often some things are left out.
If i emphasize by giving examples of what to not miss, the model starts to focus too much on those when not even relevant
Hello good morning I had a question. There’s been a lot of complaints about the team of Perplexity not looking at their users feedback. Is that true?
I can confirm this sir, try to ask or request sir.
Thank you I was just making sure
do you guys think perplexity is good for academic searching?
I'm looking for someone to collaborate with me on my business.
No technical skills are required, but you must have a computer.
You should be based in Argentina, Brazil, Colombia, Egypt, the Philippines, India, the US, Canada, Australia, New Zealand, the UK, Mexico. If you're interested in this opportunity, please DM me.
GM
when will o1 model come to pplx
Anyone else noticed that responses are shorter. I have the subscription version and on the non-pro searches the extremely short and concise they used to be a lot longer.
you're not alone, must be some sort of site issue or maintenance
me 2
Also lost it now and can't sign in anymore as well
one week anniversary?
lmao the minute I purchased pro
my bad I broke it
@fluid wadi if you haven't use the downtime to install the complexity extension. improves the experience of using the site a ton.
Download Complexity for Firefox. ⚡ Supercharge your Perplexity.ai
is perplexity slow for any else today?
@tacit nimbus Site is having some issues the last few minutes
they should incorporate this extension in the official website lol
copy that tanks
Also having issues this morning. Pro not recognized on web. I've noticed a lot of updates getting pushed across iOS, iPadOS, and web recently. Along with it, a fair amount more bugs and quirks. My best guess is they've been doing a lot of work to lay the foundation for their ad product. I'd expect more messiness before things get cleaned up.
I’m having same issues with it, so probably systemwide
Perplexity just signed me out, said query rate limit exceeded, it had an option to sign in again so I did. and we are back to normal now,
small blip
carried on searching as normal, weird
Idk, wanted to dm but you weren't accepting dms or friend requests
Does anyone else get "Query rate limit exceeded. Please try Perplexity again later." ?
Ok I see, it ain't just me
lol this addon is so good
getting the same issue with rate limiting this morning. no unusual activity other than a few queries last night.
I just lost my perplexity pro and it came back
yeah super laggy and timing out
took about 2 mins
lost it too
came back
and lost it again
yeah its back now, I needed to close re-open
i signed out and tried signing back in - still waiting for the auth email
ahhh and now i lost all my threads
wonder if the folks down at Perplexity are doing some maintenance
perplexity server: Live feed
it do be feeling like that
because i've been facing all kinds of bugs
its all back up for me now, only had bugs as of around 1h ago
looks like im back up now as well but still laggy
unrelated to the outage but I am glad the big tech companies with the largest AI models seem to be moving in unison for nuclear power either fission now or fusion later, imo nuclear stagnated because of the incentives for oil and gas etc, but it just aint enough juice
at a minimum it will get us back on track to using nuclear and stop building natural gas peaker plants
Uh, when did this recommended products stuff come up?
Can anyone tell me that limit of using Reasoning beta focused feature resets after how much time? A day or a week?
Nope does not work as intended
The search went to normal search and not to pro model @upbeat briar @unreal crater
Put the recommended products on the side bar in its own window, not in the inline results, it is distracting. And am I going to continue to see the same product recommendations when I go back to the history?
This is not cool, you guys are turning into search, and this is on Pro
I’ve never seen that while searching
This is in collections of all things
I don't mind shopping results, but they should be in a side bar window that I can call when I want
You got a good point
This is the kind of thing that motivates me more to spend time on a local LLM clone vs paying for it
when your model is recurring sub and your supposed differentiator to Google search is a moat to ads, this is hypocritical
if u have the specs to burn then so be it
u spend $240 on a llm chat interface for a year vs spending $5000 on a graphics card every year
I think you overestimate the costs of infrastructure, knowledge is the biggest gap in fact. And the perception of the learning curve is the moat any business banks on. And I gladly pay it. For $355 I can have a dual Xeon with 3x K40 GPUs with 72 GB of VRAM. Sure, my recurring is electricity and maintenance, but as it is, I have the hardware for other stuff anyhow
by gladly pay it, I mean I gladly pay Perplexity
Hey guys, first time chatter, I use Perplexity alot, but something keeps happening that drives me crazy. My session to the system doesn't time out, but if I have a tab open for too long and write a prompy, it will literally just eat my prompt and do nothing. Is there anything I can do to prevent this outside always loading a new browser tab every time I write a prompt? It's so frustrating to type a ton and then all gone.
I mean, the search did go through as a pro search. It just doesn't toggle the pro switch for followup questions.
This is probably something that needs to be addressed by perplexity themselves.
Ohh. Thanks man! It also works for now as this gets the job done
For now you just got to deal with toggling the switch for any follow up questions 🤷♂️
Although I do see that your browser allows for setting POST settings, which seems pretty unique and might be able to solve this. But I'm not at my PC right now so I can't really look into it unfortunately.
New Anthropic Model might drop tomorrow! 🔥
source?
Bro, what are you doing here?
Tell me how many subscriptions you got
I just finished up an exclusive interview going over a new, major AI model upgrade.
Can confirm, tomorrow will be a big day for developers.
Dropping the full conversation on X the second the embargo lifts 🔔
Thank you
You wish perplexity will get that model sir? in our dreams sir. 😂😂😂 sir.
so we have 300 pro searchs per day? if we dont enable pro then do we have unlimited other searches? and what is the difference of pro?
Pro gives a deeper web search of your prompt
Personalized Experience: Pro
Search offers a personalized search experience by considering user preferences and diving deeper into topics to deliver pinpoint results
Which me personally I still don’t understand
devs? so its probably a small model
betting on gemini flash again
or some gemini family model
yep
i was right
an improvement over gemini 1.5 pro ( again yea )
and gemini 1.5 flash
i dont think its gemini 2 since those big updates are usually announced in some big events
I'm looking for some one who can help my work.
they don't need any tech skills but have computers.
they should be based on Argentina, Brazil, Colombia, Egypt, Philippines, India, US, Canada, Australia, New Zealand, UK, Mexico and etc.
If you are interested in this opportunity, DM me.
Sorry to bother you, but I wanted to ask a question how many AI images can Perplexity make?
99/day
Wow, that’s great. Thank you.
Thank you so much for answering my question. I know it’s late. Sorry to bother you.
did perplexity remove pages?
Are you in the cheese mafia?
Does anyone here have a compass ai wearable?
I only have one sub and it's currently to Merlin. 🤷♂️
Does anyone know why the query per day limit has been reduced to 100+. I’ve seen 300+ and unlimited, but this is new to me.
you can check the current limits remaining here https://www.perplexity.ai/p/api/v1/user/settings
"gpt4_limit": 600,
"opus_limit": 50,
"o1_limit": 10,
"pages_limit": 100,
"upload_limit": 500,
"create_limit": 99,
"article_image_upload_limit": 500,
wow can perplexity use o1 ??
o1_mini*
Dang, that o1 limit is pretty low
Haha they need save budget
They probably have some kind of formula to calculate how much they should give to users
So they csn have X margin of gain
Taking into accounts multiple costs
But isnt 10 so low
Nope, I doubt that, at least for o1 the rate limit is super low currently, so you can’t give users more then 10 per day, even if you wanted to.
Hello! I am a new Perpleity user and I must say I am loving it! 😉 It has helped me with a couple of coding problems as well as in finding information faster than any Google search. 😄 The answers that come back from Perplexity are actually useful! 😄
greatest search engine in the world
That’s probably old because they don’t include Claude 3.5 Sonnet, and Sonar large, and Huge
So far, I believe that is true! 😄
i barely use google at this point
I have not used DuckDuckGo in several weeks. I go straight to Perplexity and always have a tab open to it. 😉
This bites me in the butt because sometimes it eats my query.
If you leave the tab open too long.
gpt4_limit is the global limits for all models except the default, opus 3.0 and o1 mini
yes 🔫 🧀
Did they remove the Pro search reasoning feature? i cant get it anymore since a few days? It was so good..
works for me \
likely ab testing
Is there any Perplexity Pro Trial? I would like to try it without immediately paying 22€
no free trial. but you could get 10usd off by using the referral code ^ from some one else, or explore the pplx's colabs (linkedin, backtoschool, etc) to get x months for free
||read the t&c first||
@mental shoal
It’s good to know. I usually don’t do more than 30 searches a day, but I never know if there’ll be days I need a whole lot more
Thank you
you can lowkey get a year for free with magenta activation code
What's that?
in germany theres a mobile data company that gives you a year free of perplexity when you buy their services
a lot of people sell their code for like 20 bucks
Where? eBay?

also 
really ?
for real
why im just saying a lot of people sell their magenta code cause they dont need it
well... im not the mod
but engaging with gray/black market on the official discord channel is... on the borderline
||since you already posted the link...||
Hyee everyone 👋👋
perplexity decided to give out yearly codes, im just telling this dude about it, not my problem that perplexity created the gray/black market themselves
Hey, I have a question. I have been using Perplexity on and off for a few months without an account. Recently I found out that I can still access the full text/conversation of my previous queries if I go to those URLs, some dating back to July etc. I had expected that those would be no longer publicly available at the very least even if they were trained internally
As I don't have an account I don't know if I can request these threads/links to be removed and I'm concerned about it
I had expected that those would be no longer publicly available
if you make the linke sharable, it is public to everyone
i dont know how you could unshare the anonymous link
but you could always write an email to support@perplexity.ai
||or dm @signal hamlet ||
I don't make them sharable, they are just searches I made while logged out, they were just previous searches, and I found out by accident today that the full URL from old searches will load with the exact conversation thread and responses vs disabling access to it or regenerating
i would definitely contact support
There is a small tag in the corner saying that it's over a month old. I understand there's a security by obscurity side to things as I'd have to share that link but I don't feel great about them being available in public, I don't think I shared anything private but it's possible I did
Ok, sending email now.
Ok email sent, here is an example of what I am talking about (the information on here is not real and from fake name generator) https://www.perplexity.ai/search/hello-my-name-is-mary-m-whitha-49roak7rRIieO3Dmoa.I7Q
Mary M. Whitham is not specifically mentioned in the search results, but there are several individuals with similar names. Here are some notable mentions:
1....
But you can see that the public link retains the data and would presumably do so indefinitely from what I've seen
Huh Merlin ?? what is that?
/imagine dj
@tame current Merlin is an extension. It lets you bring up a sidebar in your browser and chat with various LLM models, including premium models. It can read the site you're on, including YouTube videos. It has a code interpreter option. Their main draw is that they boast unlimited uses of premium LLMs. I tried it out for a month and didn't run into any limits in terms of the quantity of queries. But they have obvious constraints on response length and a limited context. It's a pretty "black box" program in that they don't tell you exactly what settings the models are operating on in the background. Personally, I opt for HARPA AI and got their lifetime license, which has a similar functionality to the Merlin sidebar but allows for creating elaborate custom commands. It has built-in support for using ChatGPT, Claude, and Gemini web connections via your existing subscription. They support using your own open router API key with any model available on their site, letting you tweak the settings to your liking.
Hello! I mainly use iOS and the web,
but I haven't been able to log in to the web using my Apple ID since this morning
Can you let me know if there are any related issues?
I tried using Chrome and Microsoft Edge to check
if it was a browser issue, and I also cleared the cache
However, the situation is the same...
Logging in with email and Google works,
but I can't log in with my Apple ID. Anyone can help me?
That sounds like a bug you need to write this on the bug-reports
how do i turn off this new autocomplete thing? its kinda distracting
Why hasn't the usage of o1 recovered?
After about 16 hours, only one access was restored.
So sad
Hi
What's the optional ai prompt text limit? Does it count markdown as additional text ie 2 for 1?
Got harpa lifetime as well, great buy imo
I personally use Merlin through their site and the context/output length is far more than the mobile version. I don't know how it compares to the extension though since I use Firefox and it's not available on there.
But you are correct that it is very nondescript when it comes to how it works behind the curtain.
I am gonna look into that other tool you mentioned as it seems to be pretty positively spoken about.
Hey everyone! I'm trying to replace Google with perplexity, just went pro... The main challenge I am having is you know when sometimes you actually do just want the rubbish Google style top result being the link to the company or thing you searched...
I.e. I search discord on perplexity
It's not easy to just find the link to discord quickly clearly near the top, if perp is going to replace Google for me and probably many others it will need to dedicate a tiny bit of screen real estate to the likely most relevant link direct to the subject in questions official site!
Any of you having the same problem? How to solve now I'm a "pro user" 🙏🏼😁😁
When will Gemini Pro 002 be supported?
Hey sounds cool! Is it similar to sider.ai? I'm very close to just buying that for a month for the side bar functionality and on screen page Summaries etc .... You tried that?
Sorry if this has been asked. But I was wondering if there was a way to see Links at the end of question or search, and then when you ask a next question like carrying on the conversation you get a new set of links for that question. But the links for first question seems to Vanish. Really been enjoying the Perplexity Experience it's helped me a lot.
After about 16 hours
each quota will be replenished in ~24hrs after use
Never sir, Perplexity hates Gemini sir.
For real, why is that?
History check sir, they had a feud before sir, perplexity competing google that is why sir. Before they told us sir that they removed it due to users not using it more often sir and peformance issue sir but was a pure lie sir. Dont believe everything here sir.
Ok thank you sir have a good night
welcome sir, nowadays people choose to lie often sir than to tell the truth sir.
If I read the word "sir" one more time...
sar pls sar
Can any of you say that the new ChatGPT advanced voice mode will follow my instructions of asking it to enunciate Spanish slowly?
Is there a special prompt that will do that better?
low but not that low actually
o1-mini is cheaper than gpt4o iirc
that's whats publicly known, most of the time you have custom aggreements on limits and pricing. So we can only guess what the actual limit for pplx is
Hi
likely higher. Bc it can choose not to use a custom account if lower.
Does that mean….perplexity will actually have the real o1???
nope, I doubt that will come in the near future, it's way too slow for pplx.
You wish sir, Perplexity is so really strict on their budger sir. Believe me sir. Dont hope sir.
Gentlemen, can you read the contents of the file?
Gentlemen, can you read the file contents of per?
why cant i see how many messages i have left with opus on my pro subscriptiion. did they remove the limit (let me dream)?
Yikes, sir. That’s crazy, sir.
So if I wanted o1 preview with web, you.com is the way to go?
But of my knowledge, I think you have 20 messages
But if you want, I could double check it
20 messages per day?
Sorry about that. I answered that incorrectly. It does not have web access.
Ok thanks for the clarification. So it has web access for other models so it’s basically perplexity but also with o1 preview right?
But no web access for o1 preview
No 1o mini is only with the reasoning search the other models do not provide o1 mini
The o1 preview too expensive
Would u recommend I unsubscribe from perplexity and try out you.com premium for a bit?
I mean I’ve had you.com for a little bit but what are you looking at an AI for? Are you looking for research up-to-date information?
Yes that and also when/if a model has web access for o1 preview with web search capabilities
can someone give me the best run down on perplexity...its been some time since i last used it, but i recenly just subscriped again.
-best system prompt
-tips
- best model within
etc, etc
Better yet I’m gonna give you a video. That’s very educational of that.
ugh
OK, fine
Ok
@calm compass if you use the Complexity extension you can use O1-mini with web search
credible? @tame current
Give me a minute I’m slow
That’s the best model I’m gonna give you the reason why
also for school stuff ?
Yes
I use it for my school stuff. It is pretty good. Little wonky on math though but it’s pretty good on other things.
i didnt even see the guide section. Thats pretty dope
⚡Join Ai Foundations: https://swiy.co/aif-0
In today's video I'm going to be teaching you everything you need to know about Perplexity. This ai tool brings the best flagship models from all over the place and curates them into one interface that has search capabilities and advanced reasoning.
Learn everything you need to know about Perplexity ...
do you have any posts within that section that have been beneifcal or stand out?
Thanks man
Let me know if you have any more information I got you
i wish you could sort by populatority in discord. That would be a game changer
I wanted to tell you I was wrong o1 does have access to the web search, but I was right about it only has 20 messages
On you.com
Ooo awesome thank you
I’ll go ahead and try it out 🙂 thanks
I m a UI / UX designer and Frontend developer.
Familiar with Figma, Adobe XD, Adobe Illustrator for various genre website and smart phone apps.
and also I have rich experience with Next.js, React.js and use GSAP and other js libraries for animation effects.
if you have trouble in UI / UX designing or frontned, I would like to collaborate with you.
Is here anyone who still got Reasoning in pro search? since a few days my pro search went from Genius to brainless. Im only getting 2 Reasearches in line and not like in reasoning minimum 4-5. The anwsers got really much worse for me. Anyone expirienced same?
It is frustrating because I have made several web queries in Perplexity and in Cursor, and even Cursor, which is actually intended for coding, unfortunately provides me with better answers than Perplexity.
will perplexity have o1 preview?
What ai model does the genius mode use?
hi
What’s with perplexity not remembering the context of the initial question after I asked a follow up directly after it?
I swear it didn’t use to do this as frequently as it does now
Yep i noticed this also. I dont know why but it seems like Plex is going back since a few days.. Still a very usefull platform and im still using it everyday but it definitly got worse. Maybe they need some money after burning
I dont see on perplexity
There is also Merlin(.ai) that has o1 with web access, it also doesn't seem to have a daily limit, just a rate limit of one message per ~5 ish minutes.
Anyone having trouble signing into the website version with their AppleID? I’ve tried 3 different browsers and it will act like it’s logging in then take me back to the non-logged in homepage.
Please take a look at this thread: https://discord.com/channels/1047197230748151888/1288313500099280926
ive been using chatgpt -4 and 4o on the official website and i havent experienced a limitation so far besides the o1
question
what model is pplx using in free version
when pro ain't available due to usage
is it still using gpt 3.5? or is it using 4o mini?
or is it confidential information
Terms of Use in Perplexity says text/image pictures are not for commercial use (only personal). If I work with Perplexity to create new programming code - is my code commercial or only personal?
Iam not seeing option to use o1 mini in perplexity pro
There isn’t an option for it in the ai model list. You have to use the reasoning focus (beta) which does use o1 mini.
what's the best way to submit formal feature requests
How to see beta than now
How to create graphs like this for real-time trends
Create a new thread and select reasoning. It is only available on the website.
For this graph
I'm trying to paste a big text into it on the perplexity website but it keeps putting it into a text file
how do I get it to just paste the plain text
Hi guys new to perplexity pro, here from Xfinity Rewards
Fatigued by the daily task of clicking the Cloudflare verification countless times
Aren't there any more user-friendly verification methods?
Continuously repeat the verification without end
Hey everyone! I’ve been working with Perplexity for just two days. Has anyone worked with the Make.com service and integrating Perplexity into it? I really need some advice, I can’t figure out why Perplexity is ignoring my prompt.
Cloudflare's indiscriminate verification is intended to prevent abuse by free users, but it ends up placing the burden on Pro users instead
😡
@signal hamlet 👍
Hi and good morning.
Hey @errant shore!
Longer texts are automatically uploaded as text files.
Hey Guys, new to the Server. I was wondering if I could somewhere find more information regarding the new reasoning beta feature? Thank you in advance! 🙂
That frustrated me too from the beginning of my subscription, is that even Pro users are subjected to this Cloudflare verification curse over and over again. 😖
Yes, I just left the page open maybe for a while and ignored it and then Cloudflare came on, too much of an exercise for the mind
OpenAI or Claude also check for abuse, but none of them do it as often as perplexity. Obviously perplexity logs cookies, also requires a logged in account to use it, and also sets a cap on the number of times it can be used per day, so why does it have to check over and over again if you're human or not?
Every time I open the page to face the CF ordeal, I wonder if the subscription is worth it. Maybe a good product needs to be more considerate to the paying subscribers.
I am using Perplexity for a single-player tabletop campaign. How do you stop the AI from ending it with a narrative summary or asking me "What will you do", or that "this choice will have far reaching consequences" every single message.
I give prompts like:
[Player's note: Reminder to avoid narrative summary or ask me what I will do next.]
It does, once, and then it repeats itself... again. Drives me insane.
probably every example in the training set includes a question or summary
it just does api calling for gpt o1 model
Thank you for responding to my question! Is the limit of 10 messages for one day, or like a week or so?
10/day
Use the Complexity extension, there is an option to prevent making long text into a file
anyone have a referal code
@bright trellis ^
bruh it's right there
Perplexity your daily source of knowledge and memes 😅
btw Plexi is good for fast code snippet/corrections, i should try it early
i had troubles with new implementation of langchain library, other AI struggle(they had outdated training data) but Perplexity found solution in recent github issues
make a full for fun Eddie Murphy type of lyrics about strawberries and they party all the time. make it a really catchy song in the best viral 80s way possible
Why wont this prompt no longer work??
All i get is responses saying sorry i apologize bla bla bla
Hate this
I used to use such prompts to make lyrics for suno ai
To make fun music and i cant no longer
How much more do we need of this ai compliance to suck
Gonna use my rabbit r1 for this for now because perplexity is apparently not the tool for this anymore
why cant i see more than 5?
seems like a waste of space.... maybe have a collection folder system when you add to a collection. #premium sub here lol what can I do to help?
how to set up my own channel?
Can anyone teach me how to generate image please??
I have a friend who just signed up with his student email. How long until he's able to get upgraded to pro?
Should have been upgraded already, please send the account details to support@perplexity.ai, so we can check and fix it.
Hey @bleak oxide, which model do you use?
Doesnt matter what model i use it refuses to make such lyrics for any prompt i try to give it
Like any of the models does this
It didnt do this hours ago
Not sure if you are doing something similar, but the model doesn't refuse:
https://www.perplexity.ai/search/make-a-full-for-fun-eddie-murp-D_iyqLmcRG.x8z7kTIZYmQ
@signal hamlet Is it possible to transfer my pro subscription to my friend's account? Or change the email address of the current account?
Changing the account email is not possible, but transferring Pro to your other email is. https://www.perplexity.ai/hub/faq/log-in-account
Please send all the details to our support@perplexity.ai email.
Thx. For some reason it will become quite hard for me to use pplx in the future. I'd like to deliver this subscription to one of my friends.
So I send all the information and the subscription will be transferred to another account with the same ending date? Should we first register the account, associated with that email, before the transfer?
You see here for some reasons it wont do it
Depends on the model, tbh, Sonar can do it with Pro and search on, too:
https://www.perplexity.ai/search/make-a-full-for-fun-eddie-murp-Hsg2T9mOTSyQ358qh4A54Q
Thank you Ill use sonar then
Thanks a lot
I got this PROMO code a few weeks ago when I signed up for Replit for four months of Perplexity pro. Had to wait for my ongoing subscription through Google finished to redeem it. Now it says redeemed on the top of Perplexity page but I got no pro and if I try to subscribe I have to pay right away and the code can't be typed in there. How do I do?
Worked first time: https://www.perplexity.ai/search/make-a-full-for-fun-eddie-murp-1blreaWcRn.p7NMRtCN3rA
Good day everyone, my name is Joseph, i am a passionate advocate for AI, i post alot about AI on my linkeldn account, please connect with me as i share valuable tips. Thank you all
Have any of you guys invested in any AI or tech companies recently? Are there any under the radar AI startups that should be on my radar? Potential to explode LT?
Yup 3 French startup
same, can't click anything
Europe?
yep
Hm.... We got the answer I guess
Me neither
Hmmmm.
Cleared cache and everything, the UI just doesn't respond to clicks. Weirdly enough it seems to respond to keyboard input. I can start a new thread with the cmd + k keyboard combo.
But cant enter anything after that
Seems to be a desktop web app issue, just tested inside the mobile android app and it works fine there
when does Perplexity come out with internship apps?
Interesting. I turned off uBlock Origin for this site and the UI is working fine now
@cosmic mantle @sage skiff
Creating a mobile app to recommend content from popular streaming services like Netflix... but the api is not as robust as web version. Any suggestion?
hi i signed up wiht my school email but didnt get a pro subscription
How can I get this promotion, please help me.
@signal hamlet Can we access to latest news using perplexity api. I am trying to search for latest Gen AI news and it is giving me the news which is from last year. How can I access the news
Same! Seeing some net::ERR_BLOCKED_BY_CLIENT errors in the console.
Workarounds mentioned here: https://discord.com/channels/1047197230748151888/1289006934090842182
Same
why when I enable nordvpn the website just don't work
Is there any admin from the team that I can DM to?
Any plan on adding support for LLama3.2 Vision90B via api?
when I come back to perplexity tab in chrome and start a converstation, most of the time it stays blanc, and I need to close the browser and open again for perplexity to work, what is going on?
I need freelancing platform account like Upwork or Freelancing.
Site stopped working/interacting today. Desktop and mobile browser.
🧭 Perplexity, guide of the lost,
🤨 Your pick has raised my brow,
🛠️ This instrument, at what cost?
🤔 Seems not right, somehow.
Hey @worldly perch, are you facing the same issue? https://discord.com/channels/1047197230748151888/1289006934090842182
Im on firefox, I downloaded the omnibox extension but on the website it still pastes it
does it need to be on edge or another browser? How do I use the extension
sounds like it
the android app works though
same problem here @signal hamlet
Llama 3.2 added yet?
Hi, using Perplexity on the Web; fairly often if I come back to Perplexity after some time, click on it and enter a prompt, it just doesn't respond at all. I have to go to a SECOND new prompt for it to work. Any ideas why this is happening? Thanks!
oh so everyone else is having the same issuee... It seems like a div overlay over the entire site, but there isnt one from what I've seen so idk
according to @fathom narwhal 's post in pro- lounge and https://discord.com/channels/1047197230748151888/1289006934090842182
the content blocker caused this issue
Nope. I’m wondering if they’re tuning it for Sonar instead of just adding the regular one to the labs.
There is a Firefox version of the extension as well
Search for "Complexity"
How come each sub search only gets at most 7 sources?
I mean sure, the search methodology is improved by breaking up the search into multiple tasks, but each search only a few sources, which reduces the validity of the results. Will perplexity increase the amount of sources in the future?
I would love to see 100 total sources
I would think you would need a lot larger of a context window than 32k tokens for that amount of input
You mean for the ai synopsis? How doe san am reading stuff require tokens?
thats sad
Who is elymc? How much of these news articles are written by them?
Please guide me how to buy Pro at a discounted price, not through a referral link ($10 off)
Reasoning focus disappeared several days ago
ab testing
Hi, im using pro version and get $5 api, next month, if i cancel subscription, will this $5 gone
free api credit is none accumulative
where are from this screenshot? tell please
nope, no idea what you're referring to but I'd love to. Can you explain?
Hello, does anyone know how to generate word documents? I tried all. It says to me it generates it and will send me, but no download. after it, it says i cant generate word i have never generated
Hi @past basalt!
Perplexity can't create word docs.
Thank you, How long can the different ki models analyze documents (pages/character) and the answer how long
bruhhh yuu kno dis how we duu aint nobdy finna c me slippn out chea wylin lk dat nahh we stay 1000 frfr keepin it lokkey fa dem boiis who cnt evn kno wha im sayin yaddamean
Ngl perplexity fell off, there are much better options these days for free even compared to perplexity pro
i use chatgpt chat interface official openai site
Ye but I feel like that’s a different purpose yknow
I def use that too for solving things and debugging very specific issues but for up to date knowledge and current events? Ai search engines are p good
now what if one site combined normal chat ui, but also had a good ai search engine and didnt suck
and then also other use cases...
all in one subscription for better models, instead of 10 subscriptions for 10 things...
tru
might be slowly working on that
you can explicitly pick when to search too
perplexity is sorta weird
there's a writing mode, but then also still a basically useless pro toggle
i dont think that ux makes sense
youd rather just have a familliar chat ui instead, for non-web searches
Felo ai free works better honestly I’m not gonna lie
hmm
i mean isnt this almost too many options
whats the difference between
social, reddit/X
and is document just web, but adding "filetype:pdf"??
again, sorta confusing
hm
Only 4o-mini for free though there
I believe
And 8k tokens on free plan context window
@signal hamlet can you please explain how the RAG works, does it take context every time and the files get sent off as part of the system prompt? Or does it only get sent once and then retrieved after? How does it affect the 32k context window?
That isn’t perplexity..
no, it isnt
??
why jump to assumptions
i do know... it's literally my own site
Well you still did not tell me what it was so I’m right
well i don't want to be banned for advertising by posting links
Is there any reason for the search results to become extremely short and disappointing lately?
@neat elk @wheat granite can you please explain how the RAG works, does it take context every time and the files get sent off as part of the system prompt? Or does it only get sent once and then retrieved after? How does it affect the 32k context window? Or do just the relevant snippets get parsed and used in the context?
what's with the reasoning thing
Hey guys, have you noticed that sometimes the response quality of Sonnet 3.5 and GPT4o on pplx will significantly decline? I tried asking them previously correctly answered questions again, but no matter how many attempts I made, they were unable to provide the correct answers.
I'm wondering if pplx occasionally switches to a less intelligent and cheaper model in the backend to control costs.
Can a code interpreter that runs programs like Python be added to writing?
In the 'discover' category, I can't search?
And is there a limit to the number of messages here?
they never answer questions about context window
Sad day, just want to know how files get parsed
You are trusting a third party provider.. This can happen
Any 1o updates?
is that even possible
hey, can i get help here with some problems i have with perplexity on make.com?
new here 👋
Llama 3.2 coming or no?
Just for info sir, dont ask something like that here sir cause no one answers that kind of question sir.
sir
Sir, sir.
Sir, sir, sir.
কিরে কেউ আছস রে বাড়া খান চুইষা দেওনের লিগা
ওই কুনো মাইয়্যা চুইষা দেওনের তাকলে রিপ্লাই দুচাইস। বাড়া দিয়া পকপক ঠাপামু
Sir, sir, sir, sir.
Aight so luk RAG be out here snatchin up da info it need from dem docz n basez real quick throwin it in da convo like BOOM It aint finna send all dat whole file errytime nah jus grabbin da hottes bitz n tossin em rite in when yall talkin Errytime u aks sumthin it slide thru pullin up da freshes deetz keepin it movin Az fo dat 32k context window RAG be playin it smart only slippin in dem lil snippetz dat matta not da whole joint ya feel Keepin dat window breezy no overload jus da good stuff str8 up
“guys my waifu actually took the yappanese test last week and… she passed with the highest grade”
Perplexity might be cooked 🫡 /j
Microsoft Copilot is making a return... wtf
Not my screenshot
This is neat
Do you know of a better ai assisted search engine?
hey does anyone know how to claim the free year of Pro for US univeristy students? I'm not seeing any options to activate it
referring to this link https://www.perplexity.ai/backtoschool
felo
How is it better?
Just better results when searching the same thing on both platforms
What search engine does felo use? Does it give more results, or just more relevant results?
i dunno lol. i would ask the devs
Perplexity is one of the best option when it comes to info. It seems that Google's censorship is crossing the boundaries of absurdity. Gemini's filters block information about the conflict related to Israel but not about Ukraine. I'm not even mentioning other types of blocks, even innocent questions are blocked. In this case, I was only interested in the historical background...
What do you mean by a theoretical background?
Are you talking about the multi-lingual capabilities? Maybe I just haven’t searched anything relevant to outside the us, but I just tried the free version of felo, and i only get 5 free uses per day, where as perplexity gives me 5 free uses every 4 hours, and perplexity has a much better search methodology. I do like the mind map feature, but I would think that that is more relevant for ideas and maybe history rather than my regular searches. Is the multi-lingual option important to you?
I do also like that every site opens in a pop up window rather than taking over the current one
I meant historical ofc 😄
Do you think that felo is better than perplexity for this use since it can search websites from Israel, Gaza, and Russia and Ukraine?
Have no idea what you are talking about
kappa fungus
The ai search engine that was just being discussed by dord
I m not interested in Felo right now. the post I wrote was about Gemini and its stupid censorship
Can anyone think of a question relevant to daily life that perplexity’s reasoning focus could answer?
But as for Felo, I use it much more when I'm looking for scientific literature. Perplexity Pro isn't able to find what I get from Felo in the free version, without using pro.
So how would you compare perplexity pro to free felo?
And how is felo better for scientific literature? Is it because of hte multi-lingual feature?
Felo is able to find papers that Perplexity Pro can't, even in Focus academic - I mostly get nonsense there. Perplexity Pro in focus Web has started to lose context more and more often and I have to write prompts more precisely 🤔 However, when I want some general info on a topic I ask Perplexity more often than Felo. I'm rather satisfied with the answers. In Perplexity I also use the writing mode, for coding and working with texts
interesting takes here and in other messages
And I don't use Felo with Pro at all
maybe i should make like a SearchEval
set of like 100 questions regarding various types of searches and topics
and then some scoring system also based on depth of response, accuracy and formatting
Felo works great in the browser version on PC. When you hover over a source with your mouse, you get a preview of that source with highlighted/marked text fragments. Perplexity doesn't have this feature. Here, you simply get a preview of the source but without all the rest.
do you think that’s a good feature?
also what do you think of like the mindmap feature in felo
do you think it’s any good / useful?
Is there a plugin or something that tells me the context of a running convo in tokens so I know when I reach close to limit of 32k?
I don't use mindmaps but I do use those highlighted fragments because I can immediately tell whether the paper is ok or not.
It seems to only give this type of mind map where everything is just branches and branches, sort of like an evolution chart
Hi, I have been a subscriber for a while, but I cancelled a few months ago. Now I can't access my invoices or pay receipts anymore. Could you please let me know how I can get access to them?
just wanted to ask does pro inckude chat gpt 4?
mindmap useless asf
idk why tf they put it in
Why don’t you guys introduce cooler background themes for perplexity. I’m getting super boated of the basic white and black theme that we have had for a while..
@shuyobiccesup_encrypted warned
Reason: Spam, please avoid sending single word messages
Duration: 7 days
Just for the record... I was not phishing when I recieved a timeout. I was fishing as when Jesus said "follow me and I will make you fishers of men". I was simply trying to give this website to someone that could bring it to life. I still love you guys and that can never change, it is all I have. Boodrow

Is there any support here? Email support with no reply for a week. No reply on github forum. No reply on #🧪│api-general . Can anyone point me to the correct channel for support?
Wait until you see Alex online and ping him.
in Green nickname?
I wouldn’t spam him but if you see him in chat, he’s usually fast to fix or at least get things moved forward. If it’s platform or bug related get a thread out in #1140622008086970420 since their team monitor it.
reply to what in https://discord.com/channels/1047197230748151888/1161802929053909012 ? please link your post, not entire channel
wassup guys just bought the yearly subscription
api team used to reply within 2 days before they moved to github
japan going to dissapear because nomore babies
Based on the available data and projections, the claim that "Japan is going to disappear because no more babies" is an exaggeration. While Japan is facing significant demographic challenges, including a shrinking and aging population, the country is not at risk of literally disappearing.
https://www.perplexity.ai/search/query-fact-check-japan-going-t-PbsDai54RiKwKDB3kdU2Tg
try to change the temp and topk/topp
and/or use some prompts like
the reply must begin with
<!DOCTYPE html><html
Ican already imagine jeff bezos head found in amazon roilet skibidi
It's poetically fitting
What should the temp and took/topp be? I did prompt "please output html only" . I also did "please output markdown only". I still get mixed results. Sometimes exactly like how I want it the some times it output '''html making everything go wrong when I automate it
Why can't the API give consistent results like the web UI or even playground?
Even when asking for citations with URL, it gives me only title, or only URL or sometimes no citations
via trial and error/grid search
Why can't the API give consistent results like the web UI or even playground?
would be nice if api team provide all the parameters used for reference
@fast meteor who's in charge of the API team?
The #🧪│api-general is just full of the same inconsistent output companies be it json,html,markdown, etc and nobody gets an answer
#🧪│api-general is actually unmanaged #🧪│api-general message
Aarash Heydari is the person who used to answer all api related questions before the discussion forum moved to github
The GitHub forum is a ghost town. Email support don't answer for a week
Which is better, to upload a document, and prompt the AI to reference it now and then, or just paste it in convo?
curl --request POST \
--url https://api.perplexity.ai/chat/completions \
--header 'Authorization: Bearer pplx-XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX' \
--header 'Content-Type: application/json' \
--data '{
"model": "llama-3.1-sonar-large-128k-online",
"messages": [
{
"content": "You are a helpful assistant. When you receive a USER_QUERY, use the output to **FORMAT THE REPLY** to the orginal USER_QUERY.",
"role": "system"
},
{
"role": "user",
"content": "Given the following USER_QUERY below, please reply in a **RAW VALID STRUCTURED HTML FORMAT** that best answers the given text. Reply starts with \"<!DOCTYPE html><html lang=\" and ends with \"</html>\". \\n\\n USER_QUERY: REPLACE_WITH_ACTUAL_QUERY"
}
],
"temperature": 0.2,
"top_p": 0.9
}'
query: a working raytracing toy example implemented in js
reply:
"role": "assistant",
"content": "\`\`\`html\n<!DOCTYPE html>\n<html lang=\"en\">\n<head>\n <meta charset=\"UTF-8\">\n <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\">\n <title>Simple Ray Tracer in JavaScript</title>\n <style>\n body {\n margin: 0;\n display: flex;\n justify-content: center;\n align-items: center;\n height: 100vh;\n background-color: #f0f0f0;\n }\n canvas {\n border: 1px solid #000;\n }\n </style>\n</head>\n<body>\n <canvas id=\"canvas\" width=\"250\" height=\"250\"></canvas>\n\n <script>\n // Get the canvas and its context\n const canvas = document.getElementById('canvas');\n const ctx = canvas.getContext('2d');\n\n // Define the scene\n const scene = {\n camera: { x: 0, y: 0, z: 0 },\n objects: [\n { type: 'sphere', center: { x: 0, y: 0, z: -5 }, radius: 1, ...
(modified from hf post)
I asked a question in pplx with search on web and got an answer with references to sources. The same prompt but limiting pplx to academic papers yielded an answer with no references at alls. (tried it twice).
Is this a feature of a bug?
when 1o preview
do any of you guys know this type of formula? or do you guys know where is it originated from?
(la-)tex
$$g_k = \nabla f(x_k, \xi_k)$$
$$s_{k+1} = s_k + \lambda_k g_k$$
$$z_{k+1} = x_0 - \frac{1}{\sqrt{\sum_{i=0}^k (i + 1)^{1/2}\gamma_i g_i^2}} \circ \sum_{i=0}^k (i + 1)^{1/2}\gamma_i g_i$$
$$x_{k+1} = (1 - c_{k+1})x_k + c_{k+1}z_{k+1}$$
$$
\begin{aligned}
g_k &= \nabla f(x_k, \xi_k) \
s_{k+1} &= s_k + \lambda_k g_k \
z_{k+1} &= x_0 - \frac{1}{\sqrt{\sum_{i=0}^k (i+1)^{1/2} \gamma_i g_i^2}} \circ \sum_{i=0}^k (i+1)^{1/2} \gamma_i g_i \
x_{k+1} &= (1 - c_{k+1}) x_k + c_{k+1} z_{k+1}
\end{aligned}
$$
here the latex, the thing is i already ask for a evaluation from different model, like claude opus and o1 preview. the intresting part is how o1 preview say it isnt a standart formula for MadGrad even tho opus gave me that when i ask about it.
Nvm, got the paper.
Hello there I am a regular user of Perplixity and love the pro version. I think it has many capabilities which make it worth the subscription. If you are looking to try it out here is the link: https://perplexity.ai/pro?referral_code=1MI14NS6 you can use to get $10 off so you can try it and feel its awesomeness for yourself
hello
is it me or chatgpt down?
Hey, I remember that the app would sometimes ask me for more information before answering the prompt, have someone triggered that on purpose?
That was the previous Pro search experience before the agenetic search was rolled out
Can I incognito prompt on my iOS app?
I am a passionate dev, so far attended various kinds of projects.
so if you have some recommendations or looking for extra devs, I'd love to collaborate together. 🤝
Best language learning prompt there is for vocabulary.
Provide a comprehensive phrase for the Mandarin word (insert word) that captures its meaning, nuances, and precise differences from similar words or synonyms, and briefly describe typical contexts where the word is used.
Use it for any language.
You're welcome
How did they know I was pro
when does o1 gonna be released? if i not wrong its in october right.
Guys, Who knows if gemini pro uses 002?
Hello good day, I am looking for support. I entered my card information for the API and purchased 3 credits. However, I could not create the API key. It did not create automatically. What is the problem? How can I get a solution?
https://www.perplexity.ai/settings/api
so this generate button is hidden from you?
I was about to report an issue (in #1140622008086970420) that dark mode looked weird in Perplexity Labs and made the text unreadable (low contrast between colors), but I don't see it occurring in Firefox or Edge.
I had this problem in an older version of Chrome yesterday, but I don't have any pictures to provide, just wanted to inform this.
Thought it was already fixed.
hii
I found it now, thank you.
I'm wondering if perplexity might be well suited for a kids' intro for AI because since it forces web search on everything it keeps it from deviating too far into weird territory. The voice conversations are also pretty good
I wanted to get my nephew a chat bot but I just not sure how safe I feel not seeing what it's telling him
In general I feel like even when I bring up dark topics it finds a way to have pretty constructive conversation on things it doesn't tend to bring up anything weird on its own
Generally uses your thoughts as a as a starting point so I feel like if a kid is using perplexity it's just going to keep the conversation really around what the kids interested in
I usually use o1-Mini and GPT-4o for large boilerplate code when starting projects, Llama doesn't help me as much, I usually have short deadlines that I need to hurry but I can't catch up myself without these language models.
Hopefully there won't be more unrealistic expectations from what people can do in such a short time.
These models do present me some interesting ideas sometimes I wouldn't have thought of myself though, but there are a lot of bugs.
I really like the Chain of Thought (CoT) reasoning in o1 and I hope to see this in Perplexity.
Hello there I am a regular user of Perplexity and love the pro version. I think it has many capabilities which make it worth the subscription. If you are looking to try it out here is the link: https://perplexity.ai/pro?referral_code=1MI14NS6 you can use to get $10 off so you can try it and feel its awesomeness for yourself
With such a large document, you may not want to rely on Perplexity since the context on Perplexity is limited to 32,000 tokens. I'd recommend perhaps NotebookLM (https://notebooklm.google.com/) or Google AI Studio (https://ai.google.dev/aistudio). Both are free to use, and you can take advantage of Gemini Pro's 2 million token context window.
Cases where you need a very large context window is where Gemini has the advantage currently, and the model should be well good enough to go through some potential questions with you.
hi everyone, im to the point where i think im going to stop subscribing to perplexity. im using perplexity in a public library and im on vpn. im on vpn because im on a open unprotected network. i get frequent disconnects while using perplexity and cloudflare constantly re-verifying my identity.
im losing time waiting for cloudflare.
im losing time just typing this complaint.
im very much sick of cloudflare's invasive protection racket.
hi
Is asking for rate limit increases a thing?
Could really use something more than 20
on API?
where did my message go
wait is anyone having the problem of not being able to log into gpt?
either someone deleted my message b4 or something
Yes
What focus do I use for coding yall
Llama 3.2 soon?
hey, is there a way to download our chat from ppx in pdf?
chungus
Is it a single chat you’re looking to save as a PDF?
ok
one full converstation session, but it's okey.
i already got it done.
Would love to see Lquid LFM 40B integrated with web or to regular PPXI search.
It's snappy fast compared to llama but it is widely inaccurate.
Therds
Does the back to school still work?
from what i gather it doesn't have a fixed context window either and can efficiently do up to 1 million tokens or something, so the expanded context window would likely help with the inaccuracies i would think
Hey @rotund geyser!
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.
They always get models out sooner but do a bunz job with it

that site is a rip off
laughs in mr arvind
English please
Are there any news about the possibility of audio reading for responses on web?
do people use this feature a lot?
Hmmm. I don't know about people, but I use it frequently, espesically when working and there is a long reply that is easier to listen to instead of reading.
Even if tts is having some issues with pronounciation, it's very valuable feature to me.
I noticed that too, like half of the articles in trending and top pages are written by them lmao
what's even weird is that their articles aren't even widely circulated on twitter
And Felo wins over pplx too
only use chat interface
examples ?
yes
example of o1 answer vs sonnet?
wtf. u did erp on perfplexity?
yes it was worth the $30,000
Which platform?
300 o1 per day?
is that a real legit 01 preview in felo.AI 300 per day sir? how about the limit context sir?
open
what are you going to do with 300 o1 per day
Hello everybody, I have a perplexity pro subscription but :
1- I cannot download more than 4 files in a collection either at a time or individually. The number of files is limited to 4. it does not want to download more than 4 files.
2- Downloaded files (limited to 4) do not attach to my collection. When I leave the library and then return, my collection is empty
It's really frustrating and I'm considering canceling my subscription, can you help me?
so how was the virtual reality experience
better than reality
i doubt. u just need to find the vorrect voice chat dude
is there a prompt which tells the llm to change its personality until i like it?
hwo did u get the best llm friend?
? on average i need to talk to 200 humans till i find 1 person i like
Get a better vpn. I use vpn with no issues.
Hi, the security problem I messaged about a while back on here and was advised to email support was never resonded to
The email was on the 24th
upload* you "download" files from the internet. you are "uploading" them. attachments are maxed at 4.
How do I get an official response from Perplexity on the matter? Kind of bummed because I potentiall could have become a paying customer given their response but between this security issue and the lack of response I don't have a great view on the product/company
I tried to use it, it is LAGGY, almost crashed my computer, MY MOUSE WASN'T EVEN MOVING.
I'm extremely frustrated, the output of all models is as stupid as a weak AI, and it's completely different from the official model's API. Moreover, the output quality is extremely unstable. As soon as Collection or Pro is applied, the output becomes stupid. For example, the following prompt "The strongest large model prompt framework, sorted by performance, top 10"😡
Even if only using writing mode, without any additional configuration, the quality of the output for the same question is extremely unstable. Sometimes it gives high-quality answers, and sometimes it gives low-quality answers. Such performance instability makes the product unusable.
I wanted to use the Claude model, but I've been getting GPT-3.5 and per-language models instead.
whats the best api to use for writting a indepth research essay finding case studys
how did you ghet this
Why does the interface of my web app look different from everyone else's? There's no thread list, no settings, nothing except a question input box. (I'm using Microsoft Edge). Can anyone help me?
maybe its zoomed in ... try zooming out or try doing complete reload... ctrl + shift + r
#1223058316662538331 message > watch GIF, give that route a try and see what you come up with.
yeah, you ships way too much js and weirdly laggy sluggish UI
so does felo
Why does perplexity think there was a 2015 Nissan maxima made?
Hi, after installing the latest update for the iOS app on my iphone, I am still missing the option to select the latest reasoning model within the focus section. When using the web version it is there. I am a pro subscriber?
When will it be available in the app?
atleast u got it all back, right?
some read reddit and can tell this taken from there, then cropped and sharpened.. please RP in your private space =]
woww..
is that 29k or 29$
how to delete it?
If you click on a page and then click on the three dots in the top right you'll see a delete page button.
Thank 😊
persplexity dead i dont hear about this on X !
so i came to here from people who use regularly
what does that mean?
Learn how to take your actions to the next level with Nelima's brand-new scheduling feature! In this video, I’ll walk you through how to use Nelima’s powerful ability to schedule tasks effortlessly, from simple reminders to advanced, recurring actions.
Whether it's sending live game scores, tracking stock prices, or creating personalized newsl...
$29,000 sir
I just installed complexity extension on chrome for perplexity. It let me choose Gemini Pro model. Is it really in the pro package or not?
do anyone of you know the prompt perplexity have for their "Writing" mode? i wanna try it on my local model.
GASPS
did you just buy 1000 usd api credit by accident?
what is the status on https://www.perplexity.ai/settings/api ?
write an email to support@perplexity.ai asap and contact @signal hamlet
there should be a confirmation
Llama 3.2 when?
neat, but man, I wish we could get a bit more control
the very low temperature remains my main gripe with writing mode
it's just not very useful for... writing
That's right. I need a refund
Accidentally purchased, 1000 API points, need refund. Thank you
@toxic moss@wheat granite@signal hamlet
Hi. I’m completely new and only did two searches so far. My question is, it created a long answer for me which looks great, but was incomplete. Perhaps it has a character limit. Is it possible to get longer replies? Or is it possible to ask it to expand on what it’s already said? Can you take a reply and work on it/build on it in a separate area (like Jasper). I'm interested in understanding the capabilities available, and if there's a fee associated with this, that's perfectly fine—I'm simply looking for clarity on what's possible. I have an assignment to do (urgent) and I think I want to use this program.
I just scrolled up and I’m seeing some photos of people in an editor. I haven’t gotten to see that area. Can someone clarify if it writes text to you directly in the editor or you can only make queries from the chat? For my purposes, I need the editor to have access to the internet as well and write text in the editor area.
hi
Hey
question why is sonnet 3.5 so bad lately
can't even pull direct quotes from provided URL's
its like it has almost no idea how to properly read and assess requested information from articles even when i tell them what to look for and give them the link
and if i send it multiple links with clear instructions, it sometimes doesnt even attempt to read it
Hi guys. How does perplexity pulls the news in the discover feed?
Thank you , that is awesome .
help
The Boeing 777-300ER is an extended range variant of the 777 family, offering impressive capabilities for long-haul flights. Here are the full specifications of the aircraft:
Dimensions
- Length: 242'4" (73.86 m)[1][4]
- Wingspan: 212'7" (64.8 m)[1][4]
- Height: 60'8" (18.5 m)[1][4]
- Wing area: 4,605 ft² (427.8 m²)[3]
Weight
- Empty weight: 370,000 lb (167,800 kg)[1]
- Maximum takeoff weight (MTOW): 775,000 lb (351,530 kg)[2][4]
- Maximum landing weight: 554,000 lb (251,290 kg)[5]
- Maximum zero fuel weight: 524,000 lb (237,680 kg)[3]
- Maximum payload: 151,000 lb (68,500 kg)[5]
Performance
- Maximum speed: 590 mph (950 km/h)[3]
- Cruising speed: 560 mph (905 km/h)[4]
- Maximum operating Mach number (Mmo): Mach 0.84[3]
- Range: 7,370 nautical miles (13,649 km)[4]
- Service ceiling: 43,100 ft (13,137 m)[3]
- Takeoff distance: 10,000 ft (3,048 m)[2]
- Landing distance: 6,233 ft (1,900 m)[5]
Powerplant
- Engines: 2 × General Electric GE90-115B turbofans[3]
- Thrust per engine: 115,540 lbf (513.9 kN)[3]
Capacity
- Passengers: 340-396 in a typical three-class configuration[1]
- Maximum seating: 550 passengers in a single-class layout[3]
Cabin
- Width: 19'3" (5.87 m)[5]
- Economy Class seat pitch: 31" (79 cm)[4]
- Economy Comfort seat pitch: 34" (86 cm)[4]
- Premium Comfort seat pitch: 38" (97 cm)[4]
- World Business Class: Full-flat seats[4]
Additional Features
- Fuel capacity: 47,890 US gallons (181,200 liters)[4]
- In-flight entertainment: Available
- Wi-Fi: Available on some aircraft[4]
- Power supply: Available at seats[4]
The Boeing 777-300ER is renowned for its exceptional range, capacity, and efficiency, making it a popular choice for airlines operating long-haul routes.
Citations:
[1] https://www.dimensions.com/element/boeing-777-300er
[2] https://skybrary.aero/aircraft/b77w
[3] http://www.flugzeuginfo.net/acdata_php/acdata_7773_en.php
[4] https://www.klm.com/information/travel-class-extra-options/aircraft-types/boeing-777-300er
[5] https://www.airlines-inform.com/commercial-aircraft/boeing-777-300.html
[6] https://en.wikipedia.org/wiki/Boeing_777
[7] https://www.modernairliners.com/boeing-777
[8] https://www.boeing.com/commercial/777
The Boeing 777-300ER stands as an extended range variant of the 777 family, blending efficiency with long-haul capabilities. Its raked wingtips and elongated fuselage distinguish it from earlier models. The aircraft employs powerful GE90-115B engines, the world's most powerful jet engine. The Boeing 777-300ER has an overall length of 242’4” (73....
The 777-300ER ("ER" for Extended Range) is the B-market version of the 777-300. It features raked and extended wingtips, a new main landing gear, reinforced nose gear, and extra fuel tanks. The -300ER also has a strengthened fuselage, wings, empennage (tail assembly), and engine attachments. The maximum range is 7,930 nm, made possible due to a ...
Flugzeuginfo.net is the comprehensive civil and military aircraft encyclopedia with photogallery, airport-codes, airline-codes, aircraft-codes, country-codes, NATO-codes, aviation museums and much more.
Boeing 777-300 commercial aircraft. Boeing 777-300 pictures, specifications, cabin configuration. Boeing 777-300 reviews and traveller comments.
Why even ask this
Make a new account and you get free searches
idk why u asked someone here
Slower response time, no message history, poor formatting, and just like, why?
Hi- my perplexity PRO doesn work. It just hangs. I logged out and tried to log back in but error. Is there a support channel here?
write an email to
support@perplexity.ai
and to
api@perplexity.ai
describing the incident details
Didn’t know perplexity was so popular, had it for like two years now, thought it was secret 😂
It’s definitely my favorite ai for asking questions
I appreciate you
🤝
I'm new to AI and I've been enjoying Perplexity so far.
Can it create images and what not? When I try, it just tells me how to make the image using image editing software.
oh thanks, no wonder @blissful raptor is not making any bread

thank you so much!
deep dive ➤ #💬│general message
so cluttered
am i able to upload images and have it transformed into cartoon art for example?
how do i copy items from perplexity to ms word
looking for team member for solana radar hackathon.. if you have knowledge of djagno because i mostly write my backend in django i m available for you anytime i have a good idea for hackathon that could be a great startup too. if you think that you can create a complete prototype of a nft store kind of website in 4 days we can team up 🙂 feel free to dm
? sp500/nasdaq is not down
never forgetti 😔
can pplx read a 35 page file in enterirety and then answer very specific questions about it? or is this too long, and it'll just read random parts and intuit the rest?
would also getting a claude subscription be better for this use case?
nvm, perpelxity pro search combined with claude 3.5 sonnet better than any of these other options for getting info from textbooks
or just general info
like it's just more right
what is 01 mini
When is llama 3.2 coming?
When I search on perplexity, each query has a set of recommended questions but they vanish after I ask a follow up questions. Is there any way to get those suggested questions again?
remember and never forget 
can u get complextiy on safari
isnt claude 3.5 sonnet better than opus
max verstappen being a part of the opus 50 movement is something i did not expect to see today
hmm we got Claude 4 opus here
Which llm leader board is more accurate: lmsys or livebench?
hi
Wussup?
Hi, anyone having issues with perplexity app on iOS18? It's instantly crashing on my iphone 14pro, app version 2.30.00
Please refund me $1000, thank you.
Please refund me $1000, thank you.
@wheat granite@toxic moss
I sent you an email, but you didn't reply to me.
Please refund me, thank you.
@flat agate
Unlimited as in we have a qty of 50 that refills, or unlimited like the Perplexity Default Model?
I see 50 qty that counts down at:
https://www.perplexity.ai/rest/user/settings
or
https://www.perplexity.ai/p/api/v1/user/settings
its an old message, not relevant
Lol, I see that now
This may well be the most complex reply I've gotten from perplexity so far
It's helping me develop a offline AI searchable database for my personal life
ah thats nice what you used
Try obsidian and Msty or Lm studio there's a lot of cool things you can do
They even have perplexity support
So not only does it know like all the relevant information it needs to know about you it can also pull information through perplexity if needed without you having to put massive archive of yourself online
You just switch to sonar huge and then toggle your vault data off to use the chat contexts. So leverage is different AI systems for different part of the task
too late, claude Opus is no longer in the top 3 🤷♂️. What's next free RC soda for all Pro users lol
.
forwarded message form months ago
has anyone tried poe? I heard it's really good.
is opus still 50 messages per day?
No only. "opus_limit":10
They have made it only 10 per day. This is absolutely ridiculous
I hope it's not permanently limited to 10. Apart from Opus, the only other decent model for long-form, high-quality responses is o1, which is also 10. Apart from standard use cases (searching for basic things), the platform is now useless. You can forget doing thorough research or writing quality content. Sonnet will never respond with as much detail as Opus, so it's not a 'replacement'.
I'm seeking a partnership with someone who can help me.
If you are interested in exploring this opportunity, please feel free to reach out to me for further discussion.
this opportunity is for you to get extra income.
Check you DM!
Llama 3.2 when?????
A.I has basically taken over the translation industry
how to set the ollama api url
How to fill the blank with correct url in obsidian text generator?
#ollama
When did opus get limited.The good models are all getting limited
So... not only you use our money to make donation to something totally useless to get taxes reduction instead of improving your services
You ignore the hundreds of users begging for an increased temperature for month
And now you limit Opus to 10 messages per day ??
F you ! I'm so close to canceling my sub and advising my 120+ referrals to do the same...
Just let them do what they want sir and we will see sir. 😂😂😂
Sarcasm from the CEO sir. Must be a lot of friday Pizza or maybe they turned into thousands of team sir.
I don't find their sarcasm funny in any sense but infact feel that it reflects the CEOs mentality given the lack of progression in PPLX, refusing to listen to the user base, and limiting their services without warning
Well let us hope he knows what he is doing sir. He has Phd sir.
Phd my a** sir
I feel bad for the moderators here who have zero control over anything happening from the higher ups but are the ones catching the verbal shrapnel
What if sir… They also want that sir. 😂😂😂
For all I know it could be
before their actions really piss me off sir.
But i am really tired of their Bullsh*t sir.
moment i saw Opus turned to 10 sir, just made me laugh sir. 😂😂😂
is this something else
or is that opus, I have premium
its o1 mini
this is the non-bulk cost for PTU for sonnet 3.5
3.5 sonnet is ($3 + $15)/2 = $9/Mtoken average
opus is 5x that
so we can just say one opus ptu unit is
$321,000 /month
they dont say anywhere what one PTU actually is
but i'd imagine at perplexity scale, they would have enough users for 10s or even 100s of PTU of opus
id belive they dropped say $N million(s) on opus PTU 6 months ago, when it launched with "unlimited" before they realised it's not enough for their growing userbase as they kept buying millions of users from partnerships - hence opus 50
and its just expired recently or something, so they decided to strageically renew for less, since they would be more profitable that way and opus isnt really a "super big prioritity" in their eyes
sure the top 5-10% of their users might complain about opus being 10 and not 50, but the rest dont care
so in their eyes, it's a win anyway (at least given their attitude about everything so far...)
...perplexity now = driven by the startup money, and becoming profitable, not always a focus on the
...perplexity shopping and this and what else
Reducing opus to 10 without informing people is such a disgusting move. I hope there'll be an explanation
You wish sir.
cause they dont give a F sir.
whenever i saw a notice in announcement sir, 99.9 percent regarding to promos and giveaways and non related to service improvements sir. so trash sir.
In settings
Yup, was my favourite one to use
Im one of the 10%
I genuinely dont care as i have enough access to other models
But i get why ppl seem unhappy
on perplexity, or like generally?
on perplexity and gpt
Tbh i use perplexity more than gpt
mainly bc of the web accessibility
fair enough
WHATS HAPPEN???