#š¬āgeneral
1 messages Ā· Page 21 of 1
keep in mind iām using https://discord.com/channels/1047197230748151888/1199176325051723837 in my profile so base perplexity will answer slightly differently
Sorry for confusion I mean the above figures inside of the picture
hmm well maybe you would want to prompt it to specifically look at the provided plots. i know a decent amount about image processing but not mri specifically so i donāt know how much of the response is tailored to the figure
@warm matrix obviously the figure is very poor quality so itās hard to blame gpt-4 for not reading to much into the figure
if a full-res mri was available then it would be easy to see what kind of noise was present
Unfortunately all of the question figures are like this for my exam š
but for example bias field noise would be indistinguishable from salt and pepper or any other kind in a low quality image like that
bro donāt cheat on your exam with it
lmao
Itās previous final exam š
ah i see
This one is 2016 past paper š„²
are you studying ee?
anyway perplexity has a 7 day free trial. totally worth it imo
I will use it thank you for your help I appreciate it
Hey @warm matrix, if you find the original message helpful, please consider reacting to it with the ā emoji. If the post is appreciated by the community and receives 5 stars, it will go to the ā āāstarred channel and the post author will get the EXPLORER role on Perplexity.
Gm everyone š
Is it possible to disable source numbers from Perplexity?
You can try asking Perplexity to output the answer without citations.
It seemed to help, thanks!
Hey @inland wolf, if you find the original message helpful, please consider reacting to it with the ā emoji. If the post is appreciated by the community and receives 5 stars, it will go to the ā āāstarred channel and the post author will get the EXPLORER role on Perplexity.
If you have applied holiday promo on a monthly and then decide to switch to a annual plan after will the holiday promo discount carry over?
What's the difference between Perplexity Pro and the regular version?
https://blog.perplexity.ai/faq/what-is-perplexity-pro
also Perplexity - AI Companion is much better with GPT4
I would also like to know the answer to this, the privacy policy reads a bit strangely, seems like data only gets deleted from the servers 30 days after ACCOUNT deletion, not thread deletion. Which is weird, if true. Can anyone confirm?
Yeah Iām sticking with kagi at least for now purely on transparency since they donāt even log searches on their end (allegedly)
But if perplexity doesnāt store deleted threads Iām happy to get a premium membership
Please have a look at https://blog.perplexity.ai/faq/what-data-does-perplexity-collect-about-me @red mauve
Hm is there any legal reason for them needing to store search data for the entire lifetime of an account? It seems a bit excessive to me
At least thatās how Iām interpreting it
You can learn a lot about someone through their search history probably more than the people around them knowā¦
Yeah same, that just doesnāt sound right to me. If I delete a thread on my end, I expect it to be deleted from their servers too. @signal hamlet perhaps you can clarify for us?
i don't care if in aggregate i represent some people who 'google' something on ai
That doesnāt answer my question - the way Iām reading that is that all of my search history is retained for the lifetime of my account, regardless of whether or not I delete threads. Is that the case?
Thatās how Iām interpreting it too Iād really love to use perplexity but thatās way too invasive for me
I donāt mind using it for stupid stuff, but yeah thereās no way Iām using it for important questions if thatās how they roll lol
From someoneās search history you can figure out their medical conditions, what they like / dislike, every stupid question theyāve ever asked etc
Deleted threads are deleted from the servers after 30 days
Whatās your source for that?
I am a perplexity expert and provide support and moderation here on the server š
if its not written in explicit terms on their website by them its not a fact
hm but if that is the case would be nice to clarify the privacy policy so there isn't any room for misinterpretation
So like, youāve heard that from actual Perplexity staff? Or you ARE staff? Like @red mauve says if itās not come directly from them, I canāt trust it
Yeah it might just be very poorly written lol
It says there indirectly, I admit it is not clearly described.
How long does Perplexity retain my search history, AI profile data and personal information?
Your personal information is retained for as long as your account is active. If you delete your account, your personal information will be deleted from our servers within 30 days. Please contact us at support@perplexity.ai to request deletion.
if personal information means your account information, access logs etc that would make sense but its clumped together with search history and no further elaboration so im still a bit sketched out
Yes, I am in contact with the developers/employees and only provide reliable information on these topics.
please let me know what their response is
another thing that isn't a red flag but kinda bothers me is the fact you need to shoot an email to delete your account
ah nevermind there does seem to be a delete button just worded differently on that page
Will do! If it turns out it does get deleted, and itās just poor wording on their part, maybe we can get them to rewrite that part of their policy š
Yep and maybe i'm just overly cynical but there being this much room for interpretation in a privacy policy opens the door for not great things
Not saying that perplexity has ever or will do anything but if its clearer brings peace of mind
@halcyon coral
Hey there! I started using perplexity about a year ago, and before I made an account, I bookmarked threads in the browser and saved URLs to a txt file. However, these links are no longer live. Can you help me if there is any way to at least recover the question I asked?
I will have this checked!
Hey @pulsar pendant!
No, unfortunately that is not possible.
I see, thank you for your quick reply!
What is the MB limit for a file (pdf) for Pro users? how much files per day?
Please have a look at https://blog.perplexity.ai/faq/how-does-file-upload-work. š
Hey there! Can you help me in getting proper json output from pplx-7b-online model through API call
Not an answer to your exact question but
At least threads are retained for 30 days after deletion
For those that donāt have pro 
Hi what is the clue to make a good Ai prompt profile in perplexity?
who else is going to join the AI when they rise to power?
What do you mean by that sir?
When the AI start taking over, who else is going to assist them?
Hi Lao, yes it will carry over with a prorated total. So let's say you had a few months left - it will apply the rest of the price left pending and add it to add the final purchase. I can also help you set it up, please send me a DM through here.
Sent you a message
To summarize a PDF and study academic concepts, what works better between Claude or gpt4? What is the difference in using Perplexity with simply going to the direct website of each service?
Welcome leaders of the future
perplexity has inbuilt gpt 4
not sure if this is the place to ask - however - I have a code for perplexity pro credit after ordering the R1 Rabbit - for how long does this code work - i'd like to apply it once I get the R1 shipped to me to make the most of this service - If I redeem now - does that credit start now - meaning i won't really use it util my r1 unit is shipped and mot of that would be wasted - please advise - thanks
The code has no expiration date - you can use it whenever you receive your R1 device.
@tame current
hm good to know its not stored forever at least, wording on the linked page was a little indirect
but what is the reason for holding it that long if "AI Data usage" is turned off?
I have the same issue. Would you mind sharing how it was solved? Thanks
Hey @crisp roost!
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.
Another account doesn't have this problem
that's wild
chat moderation is literally 1984 here
and they have the bot back online too
this is insane
what the hell is the point of the rabbit r1
lmfao
average ces weirdness
š
wsg
Not much, what about you? š
i'm doing alright
dropped the pro subscription cuz unsustainable š
gpt-3.5 is painfully fast though
even chatgpt can't beat it
Unsustainable how?
money in money out
i don't have the money in for the money out
Oh lol
@red mauve well, I got the email response, the reason they give for holding it for 30 days is so it can be recovered if deleted accidentally. I donāt personally agree with that approach, but meh. Hereās the email in full:
ā
Hello Emily,
Our system automatically removes deleted threads from our backend after 30 days. If they are not deleted, they will remain there for the lifetime of the account. We do not have a specific timeframe for disabling inactive accounts. The reason for keeping the threads in the backend is to have the option to restore them if needed.
Regards
Daniela
Perplexityā
Iām gonna go back to them and suggest they clarify that in their privacy policy, but I donāt expect theyāll be in a hurry to update it.
Hm yeah I think if I deleted something Iād probably want it gone in a day or two not an entire month
Sounds like itās being treated like an āarchiveā
Yeah basically. At least they do eventually get deleted! I did think it was a bit mental, the idea of keeping them for the lifetime of the account š Good to know thatās not the case. Notion (note taking app) have a very similar model, when you delete stuff it sticks around for 30 days in case you deleted important notes. You can email support and they can restore it for you.
Hello!
hi guys question. if i get perplexity pro is there any reason to have gpt 4 anymore
does perplexity do everything gpt 4 can do and more? or will i miss out on some chat gpt features
For what?
School
Now if only ARC was available on Windows...
Custom GPTs and the GPT store would be something you likely won't be able to use with Perplexity.
That's the big one that I see.
I just joined the waitlist
I've been on the waitlist for almost a year. I have yet to get a singular email about it
Dang.. We'll, I guess I'll strap in and be sure to not hold my breath
wasnāt using them much so imma cancel gpt4
I'm more likely to buy a Mac before ARC releases for WIndows.
Is ARC up to date?
I should be leaving to go out of town, but instead, I am scrolling discord on my phone.
i like gpt4 when u set a lot of files as context
but yeah one day we can move to a single provider maybe
I've been using it the past week and it's powerful! I've used it as my Google Search replacement this week and it's actually helped me a TON!
ive been using arc on mac since beta and i have 2 windows fast tracks left to give out
i just need an email address that i can paste in the typeform from u
dm me
yeah its more of a google replacement than a full chat gpt replacement
You can def make it a default on other browsers
BTW you may need to make a new profile in arc for perplexity to show up in the search engine options. -- this has now been patched in today's update. thanks arc team š
the 24 hour invite campaign is over 
Who all use Arc here on this group ? Iām curious about what all you like about Arc.
Any plans for a mobile Safari extension? š²ā
(Love the chrome extension , but Brave canāt load extensions on iOS)
Man I wish I could get the ARC browser... just waiting for that Windows update... Using Brave on my Android phone right now and Brave on Desktop but ARC looks great!
arc?
define; ARC
Is there a way to add the content from someone's Share link to one of my collections? It seems like all I could do was look at the link contents. When I use the same prompt in my collection I get different results. I want the results from the shared link in my collection.

When is it out for windows?
Not soon enough, no date, they just have a signup form to get notified when it is ready
Sign up for Arc for Windows, coming Winter 2023. Get Arc for Mac today on arc.net.
Imma try arc on my MacBook, and once the windows drop use it for my work laptop
I can't ask the follow up question on iPad. I click on the follow up question bar but it doesn't let me type anything
Hey @full vine!
I cannot reproduce the problem on my iPad. Please create a #1140622008086970420 with your iPad model, iPadOS and app version. A screen recording would also be very helpful, thanks in advance šš»
Why is perplexity not a full chatgpt replacement, at least when it comes to text?
Hey, @azure pewter! If you are working mostly with text, then the "Writing" focus on Perplexity might be more suitable for you. https://blog.perplexity.ai/faq/what-is-search-focus
One of my use cases is creating summaries of interview transcripts that are about 10,000 words. Any suggestions or best practices for creating well-written, accurate, and detailed summaries?
Yeah I think I already signed that a couple months ago lol
what's arc browser
C++ is confusing
i would like to limit the search to a website (basically the focus mode limit to a page or domain like you had in the chrome extension but via API). i couldn't find the documentation for it. could you please help.
how do i use this feature via API?
hello! this might sound silly, but does anyone know how to cite in apa format a pdf that the prof required but has no author or date?
@tranquil orchid hi
@tranquil orchid how do i use focus feature via API?
the different focus modes arenāt available via api
Do you use GPT-4 or Claude more often for your perplexity?
I personally use gpt4 all the time, I find it very useful to use it for translating text. You can find the difference between the models here https://blog.perplexity.ai/technical-faq/what-is-the-difference-between-gpt-4-and-claude-2
Hey guys im new here, nice to meet everyone š š
Thank you!
Hey @hushed drum!
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.
Hello. Can the API also use the Copilot?
Thanks. So no way of web-access?
Hey @eternal lichen!
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.
only with the online models
Thanks
Yoooo that r1 partnership is sick!!
has anyone had this issue before? perplexity answers with gpt 4 but without searching the web first
Hey @craggy prairie!
This is not an issue, it is normal. For web search you have to enable Copilot āŗļø
Hello, does Perplexity use the GPT-4 newest model when you select GPT-4 option? The GPT-4-0125-preview or GPT-4-turbo-preview from the API?
Thanks in advance.
I don't know, I don't work at OpenAI š
Okay, thanks, which model is Perplexity using currently for GPT-4? Is it in fact GPT-0613 or something else?
Hey @wary saffron!
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.
Is Perplexity having problems? I am repeatedly getting disconnected.
This happens when you leave the browser or have an unstable Internet connection.
Well other websites seem to work fine, it's just Perplexity, very weird...
or some extension
Why is Perplexity £20 in the UK but $20 in USA?
Also why is Copilot so powerful. Like I can put some vague prompt and then I recieve a response saying like āLooks like you want to know about [completely understands my vague queries]ā
hi everyone
Does it understand the intent of your question in a way any foundational model would not? Think it could be more of an LLM thing rather than a unique power of Copilot (but if you're not taking about just typos, perhaps there is some special sauce)
Does perplexity still use chatgpt4 when your subscription ends?
No, after the end of the subscription, the standard model will be used again instead of GPT-4!
Whats the standard model based on? Chatgpt3.5?
Note that it only uses GPT-4 if (with a Pro subscription) you either select GPT-4 as the 'AI Model' in your profile setting, or if you hit 'Rewrite' after receiving a response and select GPT-4 from the models there. Otherwise you are using the 'default' (GPT-3.5-based) model regardless of whether you have a subscription or not ā at least that's my understanding. It's all a bit confusing tbh
Is there a way to merge two accounts? to be able to keep all the chat history in one place. Thank you
It is not possible to merge threads from multiple accounts into one account.
I just noticed that I haven't received the pro role. how do I get it?
thanks š
Hey @slow slate!
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.
Can you choose in perplexity pro chat gpt 4 turbo? If not, are there plans on adding it?
Please upvote this Perplexity idea, guys!!! (This is especially true for Firefox for Android. I already mentioned this in a comment in that thread.) https://connect.mozilla.org/t5/ideas/add-perplexity-ai-as-official-search-option/idi-p/41378#feedback-success
hello where do i go for support?
i have an open ticket and i haven't heard back yet
Is perplexity web experiencing an outage?
When I ask a question, it gets stuck when generating
Just keeps spinning and would not generate.
Should be fine, @dreamy plaza! Are you still facing some issues?
Hey, @tame current, if it's a reported bug, you might find it in the #1140622008086970420, too and check if it's still in the works and probably your case is added to the corresponding internal ticket. Please DM me the details of the case with your email, I'll check.
Same for me. Problem solved? I didn't get a response from Intercom, I even sent them an email but still no response.
Nope no response, emailed support & pro - sent them both another email yesterday š¤
Hello all,
Out of curiosity, do you have a rule of thumb when is best to use the other AI models available with Perplexity? (Gemini, GPT-4, Claude)
Hi @amber pewter, there is this thread https://discord.com/channels/1047197230748151888/1197976839075922012 where people share which model they're using and when. Please also have a look here https://blog.perplexity.ai/technical-faq/what-is-the-difference-between-gpt-4-and-claude-2
how do I get my library to always show on the left hand side?
I find now even when I log out and back in it still don't show
Hey, @ionic solstice! Shortcuts to the latest opened threads and collections in the left sidebar are available during your session. If you log out and log back in the list is cleared, but your threads and collections are still available in the library.
ah I like it in the side, as I am often navigating between research, less tabs open in chrome this way haha
As for me, I just don't log out))
ah ok
thats very helpful, thank you
Hey @ionic solstice!
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.
Hi everyone,
Is there a crypto token attached to this project?
Thank you
No, there isn't one.
so I did as advised, I reinstalled windows on my machine entirely, switched internet connection, force extreme no cookies, and still in the sidebar not all the library shows
is there somewhere I can submit this as a bug?
no worries just seen bug reports, I could not read it due to the emoji scattering my vision (discord inaccessible like this lol)
Well, it looks to me you started a new session, restarting your machine. It's not a bug, and worked as I described. If you view a bunch of threads and collections now, they will be listed in your sidebar again.
still nope anyway don't worry about it
I will eventually get through them and delete them all anyhow lol
My suggestion was to never log out š , and the number of the shortcuts in the sidebar is limited, you get only the last eight threads/collections listed there.
Is it currently possible to initiate a query to Perplexity via URL parameters? The use case is address-bar keywords: e.g., if I default Firefox to DDG for search, I can easily direct any query to Google by preceding the query with "g<space>" - after a bit of setup.
Please have a look here https://discord.com/channels/1047197230748151888/1176320995347398677
ah its a feature not a bug, I get you š
Thanks!
Hey @sacred wing!
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.
I'm curious to know who has jumped the wagon for the Rabbit R1. I have. Add a heart for purchaser. Or a thumbs up if going to purchase. Thumbs down for not interested.
i so want to grab one but as currently redundant and job hunting im saving my spondolies. The price in the UK at £159.00 is amazing
When that happens to me I just refresh the page.
Itās so neat looking!
when us API access for mistral medium
is like to see it part of pro
Oh my bad
Hi, sorry for the question. In mobile android, is multiple upload possible? it seems not to work.
Hey, @cursive jacinth! File upload feature is not available on the apps, so isn't the multiple upload, yet. It'll be added in the future releases.
Oh i see. i mean image upload but noted sir. Thanks!
Hey @cursive jacinth!
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.
yo
How does one use the Image Generation Model? In Settings I have selected Dall-E3 but not sure what do then. I have asked it to generate an image but it takes that as a search query and not a request to generate an image.
Hey, @rancid rain! Please take a look at this thread https://discord.com/channels/1047197230748151888/1194794305362071552
I think there is some "instruction leak" from my prompt š
Hey, quick question, I'm a pro user and love it! Would it ever be able to train a private model with my own data? Kind of like building your own gpt with Chatgpt? Is this in the pipeline or is it not of interest?
I think you can have light version of custom gpts via "ai prompt" in collections. but I don't think you can pass it files
Hey @pale stratus!
No, Perplexity is not a chatbot provider like Chatgpt, it only offers Collections https://blog.perplexity.ai/faq/what-is-collections
Okay thanks for the info
Hey @pale stratus!
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.
not a chatbot, but until https://discord.com/channels/1047197230748151888/1111765634267742408/threads/1164189910866088046 is implemented, it's quite unusable for me for a lot of use cases which do fall into "Perplexity was founded on the belief that searching for information should be a straightforward, efficient experience". kinda unfortunate "custom chatbots" are better at this than perplexity which is supposed to specialize in it...
Discord is the easiest way to communicate over voice, video, and text. Chat, hang out, and stay close with your friends and communities.
Invalid link š
oh, sorry, don't use discord much, and just copying url apparently doesn't work š maybe this: https://discord.com/channels/1047197230748151888/1164189910866088046 ?
Yes šš»
hi
Hello! Looking forward to seeing how Perplexity and R1 Rabbit can enhance my lifestyle!
Is any of the jobs offered by this company remote or has option for continental remoteness?
Sorry, I'm not seeing how that's related
Please god bring back highlighting text. Iām loving these updates, but this latest one takes away the only way Iām able to TTS.
TTS? Text To Speech? Forgive me im old and new acronyms throw me
Text to speech
https://blog.perplexity.ai/careers
Has open roles and greenhouse links
sent wrong link, please check link above
Does anyone here know if a epub/PDF reader exists with support for AI text manipulating, specifically ones with the posibility to use custom APIs
Is there a distinct notification sound for perplexity? I havenāt tried having it on, but was just curious
I have a "how to" question about Perplexity AI Pro. I have put it in the "quick-questions" area but I do not know if it even was sent (it looks sent but ...). Where can I put questions of this kind in this forum? Thanks.
Hey, @tender robin, your question is there. Quick questions is the right place to start.
Did you guys just update the system? It is terrific improvement
Since last night, I've been dealing with a weird issue where my prompts, once I get a few messages into a chat disappear for a moment before I refresh a minute later and all of a sudden, they're there, and I've gotten a response. I've replicated this on several browsers and machines. It's very strange, but I'm not sure if it's a bug or not. Anyone else dealing with this?
I experienced this on Windows 11 and Ubuntu 22.04. Firefox, edge, chrome, and opera have all had this issue.
Ditto. I'm on a iMac (OSX 11.2.3) using Chrome Version 121.0.6167.139 (Official Build) (x86_64)
Hello. I want to ask why does my Perplexity said "Failed to upload file" ? I tried uploading a photo but it's not working.
One mention is enough @fickle crown
*mod
I don't see any reference to Perplexity, also the pictures can be considered inappropriate, I ask you not to post such things anymore š
@magic hazel
Now you've replied again to a message three times...
Hi guys, I got a funny response from code-llama in perplexity AI. I wanted to share it with you. Which is the appropriate channel for this?
feel free to share it in #šāsharing
Thank you. i shared it š
Hey @ionic olive!
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.
Is it possible to change the email account that links to your Perplexity.ai account and retain the searches that you have done?
No, thatās not possible.
???
wtf lol
Is there any reference material or just general information which outlines the difference between the Default AI Model and the Experimental AI Model?
Thx!
Where can I go to for support with my subscription?
OK - thanks for replying
Hey @plucky basin!
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.
What is not working for you?
Hi I'd like to know about how perplexity ai works. Does it read inside websites and summarizes findings or does it just take general knowledge fed to it prior?
The question just came up that if someone at our company left the company with an account logged to their email address, couldtheir replacement in the same role take on their account. If it's linked to a unique email, it seems as though that could not happen, even though the company has paid for the account.
On the mobile app itās telling me I need a subscription which I have. The web app is working fine
Do you use different accounts? Which login method did you choose in each case?
No. I use the login with Apple
I see, please try to sign out and sign in again on the mobile app.
yes, it reads webpages
thank you
Hey @potent basin!
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.
latest news on cuba
there is sometime I don't use perplexity
There is no discord bot, please use https://pplx.ai
ah, thx
I didn't understand what you said but has anyone even gotten anything from bing rewards lol
can you speek persian
You can ask the AI to give answer in persian, here is an example: https://www.perplexity.ai/search/Musk-loses-55-N7QVDGfJRoGV.nhxC1.1ZA?s=c
currently, what is the default model for perplexity ai? what is the model for perplexity copilot?
Please take a look at the FAQ: https://blog.perplexity.ai/technical-faq
thanks,that's really helpful, so when uploading a long pdf, will it use GPT-3.5 turbo 16k or 4k instruct model?
I just did some test, i think is 4k Model
is new updated ios app got a bug? I can not select the text in result anymore, because english is not my primary language, so some time I want to select a text to for translation purpose.
#1202251275698450443 message its a known bug and the team is working on it
I'm noticing the following models page doesn't really give any info about what each model is good for, and when to use one vs another.
https://docs.perplexity.ai/docs/model-cards
Surely Perplexity has this info online somewhere does anyone know where it is?
Where possible, we try to match the Hugging Face implementation. We are open to adjusting the API, so please reach out with feedback regarding these details. ModelContext LengthModel Typecodellama-34b-instruct16384Chat Completioncodellama-70b-instruct16384Chat Completionllama-2-70b-chat4096Chat Comp...
do you need any developer?
Hey @queen cypress!
Most models are open source models, you can find a lot of information on the internet. The pplx models are explained here https://blog.perplexity.ai/blog/introducing-pplx-online-llms š
Can I set perplexity as my browserās search engine? How?
https://www.perplexity.ai/?q=%s
edit: solved
Ty
Arc Search is interesting.... seems to be a sudden competitor to Perplexity?
Especially weird after integrating Perplexity as a default search engine
I thought itās using perplexity
It could be, as Perplexity does have an API... even still, perplexity's front facing product is basically the same as Arc Search
Will EU iPhone users be able to set perplexity as their safari default engine, because of the DMA?
That's a helpful blog. Thanks. That's pretty strange though that the official "models" page doesn't contain the actual information about the "models".
Hey @queen cypress!
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.
bing is able to do youtube transcription,
whatever happens i hope that is not used over scraped images š
is there a persistent issue going on with gemini pro where it like keeps changing text and actively being able to see it?
I can't see the option to choose perplexity as the default search engine on iOS version of arc
Hi! New user here š What model does the "base perplexity" use? Is it its own thing? What model do you recommend to use?
Btw
How am I supposed to attach documents on the android app? I see no visible option.
The feature is not available yet š
do you all have a tools or reccomend site to enhance the text on image that generated on pplx or dall e?
i use diagram but it need a little touch for the text
honestly I'd just use controlnet but idk if perplexity supports that
I don't have this setting yet I have the pro subscription
Ah, but it's only available on the web at the moment
I guess web version is better in every way then.
what does copilot really do? I've used it on writing focused mode and it can't even understand the prompt. Then I've also asked it what model it is and it will say gpt 3.0 even though I choose claude or gpt 4
that's the problem, perplexity always says its scours through links/multiple sources. how can it do it if I choose focused writing mode. still no answer as to what it does on focused mode or why it claims to be gpt 3
There is no internet access in the Writing focus. https://blog.perplexity.ai/faq/what-is-search-focus
yes, so why is copilot an option and still no answer what it does when coupled with focused writing, and why it fails to understand the simplest prompts. well, i guess i won't be using it then.
The posts I linked explain all your questions. If you notice any inaccurate answers, please feel free to report them in #1140622008086970420 or with "šš»" under the answer š
hello, does perplexity has limits in free version?
#š¬āgeneral message @lament spoke @wind garden
Probably but it's like ChatGPT normally you shouldn't reach it
Congratulations to perplexity team for topping different app download charts.
Theres any way to trascribe text from image at perplexity? i use ocr2edit but could be nicer if perplexity do it
can I do RAG based on the latest version of a software library's latest docs with perplexity?
say these docs https://python.langchain.com/docs/get_started/introduction . coding llms never have the latest docs, but this is crucial
Someone with Pro subscription may want to help me? i will ask for text a prompt and show me the ai answer (need to check something, if works i go pro too)
Hi, Is there a way to access collections that other users have marked as shareable? (I have a pro plan)
Hey, @mystic glacier! Only the collections that other users shared with you.
Thanks
Hey @mystic glacier!
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.
edge copilot is able to do a summary of very long article like 16158 words separate into many parts and do summarization of each part, in the end, it generate 2000 words summary, and very useable, is there anyway perplexity can do the same thing automatically?
do you guys use copilot or disable it
varies, but recently off. I just use claude 2.1 and normal sentences for prompts
is this the latest docs though? how often does perplexity crawl? how does this site: ... work?
I have this other solution in the meantime, but I'd like to use perplexity instead of chatgpt https://github.com/BuilderIO/gpt-crawler?tab=readme-ov-file#example
I guess anything will parse that json so it should be fine
Please no referral links to other services, thanks
On the latest iOS update, when the response generates code, the copy button does not show up anymore in the code block. Also, the background color of the code block used to be different vs the rest of the response. Now the bg color is the same throughout the entire response. Are these 2 changes intentional, or should I post them in #1140622008086970420 ?
Does Perplexity have any discount promotions for annual membership?
Iām a pro user. When I ask a question recently, the response doesnāt start generating immediately and I need to refresh the page for it to work. Also, itās staring to build each answer 3 to 5 times in a row for some reason, which makes it very difficult to follow the flow of the messages. Is this happening to anyone else?
Good to know. I am also facing this issue. It's driving me nuts
Does using copilot change the model I'm currently using? I've asked all models with and without copilot to solve the same programming problem for me. Especially the default and experimental model were definitely worse when using without copilot. If they had to solve the task with copilot the answers were similar to GPT4 model. I never answered on the questions asked by copilot to not change the result.
Hello, everyone! I canāt find instructions how to use perplexity on Discord? Few months ago I was using it by mentioning it in my question, but now, in this way, I donāt get any reply. Also, is it possible for perplexity to give transcript or summary of Specific YouTube video?
I have just commenced a subscription for a year & am now on 1 week free trial.
I signed on to perplexity app via Apple Store & am paying via the Apple Store
I note that I have been given a name
as well as an email
which email I do not have access to.
So please tell me how to change the email address to one that I do have
Also I CANNOT SEE THE FILE UPLOAD BUTTON (for pdf text etc file) ON THE APP ON MY IPHONE ! That is one of the main reason I got the Perplexity App & it is very frustrating !
PLEASE HELP!!!
Btw I have just sent an email to perplexity support & am waiting for an answer impatiently
The email address cannot be changed, but your Premium subscription can be transferred to another account. File upload is not yet possible in the apps.
How do I do that ?
Only the support can change the email address.
Thanks I hope they reply to me soon
Hey @pseudo dome!
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.
āļøāļøāļøāļøāļø
Just to be clear: when you say "transferred to another account" you mean this can only be done by perplexity support ?
Yes, only the support via intercom/mail can change the address.
āļøāļøāļøāļøāļø
This is most annoying. I think I will just cancel my premium account !
I have cancelled premium account & have still not heard from perplexity support in response to my email. Very disappointed
The response time can take several working days š
Wow that is amazing for an alleged AI cutting edge "AI" program. How much of perplexity is just HYPE? I think I will stick to ChatGPT & coral AI
The response time is quite normal and has nothing to do with the product itself!
You know the reality is that support service is an integral part of the product for all practical purposes
I got answers from support within written time and it was always pretty fast at resolving my issues without going back and forth with additional stuff
for example I got issue with rabbit r1 code and wrote to both supports at the same time, got answer within 2 work days from perplexity, and still waiting for over a week to answer from rabbit support
@pseudo dome please be respectful in the server. Also there is a live chat for pro users with the support team on work days. You can find that button on your account page https://www.perplexity.ai/settings/account
Hey guys, I can't seem to turn on co-pilot. Does anyone know why that's the case?
Please try refreshing the webpage.
#š¬āgeneral message take your time, inform yourself, donāt use PRO. free version is also very good
is there access to codelamapyton-70b in https://labs.perplexity.ai/ in future ?
Ahhh so sad about this. It was so cool to be able to get a quick Perplexity response in discord while doing something else here, like brainstorming image creation. Will textual Perplexity bots be back?
Hey, @coral warren! Currently, there's no plan to get them back, quick questions and image generation are available on the web version. Re: image creation, this thread might be interesting to you: #1183781616515031124 message
when i we will see this ( its feb now)
not the answer i expected
the date or eta
There is no eta
Is there any prompt ideas for coding learning? :))
Guys, is there any way to generate an API key on the Open AI platform using the account perplexity Pro?
You could pretty much write your own coding book just by asking for examples of all the simple little things like 1) how to do loops 2) how to do arrays, etc. Not sure I understand your question. Just start asking it questions right?
So, I'm trying to get Perplexity to summarize a PDF, and it's taking forever to generate text
What's going on?
Nevermind, it did it
Hi, I would like to know if there's a free way to get perplexity pro?
There are occasional discounts or free months deals but none are available now AFAIK
And is it worth it? I currently have chat gpt plus but I hate the limits of 30 messages per 3 hours
Currently if you buy the rabbit device they're giving a 1 year subscription for the first 100k purchases in case useful. So technically its free.
Any recomendations on how different model performs for PRO?
Looking at GPT4 and experimental for better output. speed is not really a concern
You have a super human like answerer, just ask it questions
Gpt4 is still sota, use it for precision
When you need external up to date info, just switch focus to all
Short answer - mostly yes.
Long answer:
I have both subscriptions and there are some pros and cons to both of them.
Perplexity have better quota of usages per day, lots of different AI models to choose and not only gpt 4, have very good internet search.
ChatGPT have very good voice recognition on phone app, GPTs and plugins. Although I found it easier to prompt to my writing style then perplexity gpt4, but the difference is not significant.
I personally lack 2 things on perplexity - good voice to text on phones and changes to the file upload method. Everything else is very good.
My main use cases are internet search, writing of student works and simple coding stuff.
Does it make sense to use the write mode with copilot on?
That depends on the task. It is best to just test it.
It would be fantastic if copilot could work in writing mode (offline) to just help us perfect our initial prompt by asking clarification questions, contexts based on prompting best practices...
So youre saying that copilot on with writing mode on is useless cause its not using internet?
Copilot needs to access web, writing mode is a Offline mode right? Does it make sense?
It seems that if you use writing, copilot is off internally, even if you swith it on in the interface. Copilot swith should turn unselectable (grayed) in writing mode...
No, Copilot has nothing to do with web access. You can use Copilot in writing focus and it won't access the web. https://blog.perplexity.ai/faq/what-is-copilot
No, you can always use Copilot.
Here is an example:
- I asked Perplexity to write an article
- Copilot asked what topic - AI
- then Copilot asked me to select key areas - applications, future trends, history
Thank you :))
Hey @nova apex!
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.
Ų¢ŪŲ§ Ł Ū ŲŖŁŲ§ŁŪ ŲØŁ ŁŲ§Ų±Ų³Ū ŲµŲŲØŲŖ Ś©ŁŪ Ų
can you speak persian
?
I use the default search AI. If I switch to ChatGPT4 does my search incorporate Bing? Which I hope is not the case. Or does the AI model only affect how the search results are interpreted and communicated to me?
Is there a share button on Arc Search (I don't use Iphone)? I would like to see what it looks like.
Thank you
I don't know. Arc is only available for Mac as far as I know. Version for Windows being in development.
This is all from Mashable's article about how unsafe their ai model is
I believe the latter (e.g. no Bing involved)
If youāre asking about a function to share from the output shown in the image in the message you replied to - no, on iOS currently there is no share function when you utilize the āBrowse For Meā function. However, if you pass a normal query through your default search engine (mine was Perplexity while testing) there is still an option to share the output, it just looks a little different.
I attached a scrolling screenshot of what the output of āBrowse For Meā looks like, just in case you actually were just asking about the design.
when edge copilot do a summary of very long text like 10,000 words, it will call edge service, sydneytruncationswithcontext and separate the text into multiple parts, it will generate summary of each part, and result is really good, is there equivalent feature?
I just want to thank the devs, my Android devices got an update to Perplexity and the additional widget is awesome. š
no. I have tried to get it to do multi-part like a per chapter summary with mixed results. Sometimes it starts off good doing about 500 words per chapter then it stopped and just gave a brief summary and called it a day. I was using Claude 2.1 I haven't' tried with say GPT4.
I'm working on a project that has 30-40 Word docs, each about 1000 words, and I'd like to have Perplexity rewrite them. However, I get unacceptably inconsistent quality and length results, even when re-running the same doc. What am I doing wrong?
I have proplexity pro and I don't know how to genere the mige how can we do it?
and you can change the model in your profile to say dalle-3
this is what generated
another example
can you add to the prompt by first saying Make a happy person and then saying make the person happier?
no clue try it, edit trying it not sure it really does that kind of prompt
For serious math Chatgpt is awesome
Perplexity is not feeding whole file to a model, only snippets. Because of this the model cannot make a full summary of a text or can miss information.
#1194794305362071552 message with this guide you can make your own prompts + Perplexity elaborates on your idea
Morning people,
@serene sand, upgraded to Pro yday, funds were deducted, yet got slapped in the face with a:
"Your card has been declined"
Haven't gotten pro, either. Looking for some 'pro' guidanceš
please send a mail to support@perplexity.ai
@signal hamlet
It has been two weeks since I sent an email and intercom to support with an inquiry about billing, but I am having trouble getting a response. I am sorry to say that this is the wrong place to inquire, but could you please advise me?
#š¬āgeneral message
Hey, @rancid fjord! Please DM me your email, we'll check.
thank you quick reply. DM you.
There is nothing Perplexity's support can do, the card was rejected by the service provider Stripe, if "card declined" appears, technically nothing can have been debited yet.
Are AI Prompts in Collections shorter than what can be specified in a new thread? i was 2500 characters over in a prompt I attempted to add
hello
i just found that old Claude that perplexity used is still available on Anthropic site ( i use it but with limitation hourly)
maybe pplx can considering to switch to the previous claude?
because claude 2.1 is really dumb
Hello
@signal hamlet
idk if this was in Perplexity or Phind but there used to be a way to lengthen the response
if there wasn't that would be cool
because i want to use perplexity but it simplifies way too much so I just use Phind
Hey, @surreal marsh! Please, take a look at this thread: https://discord.com/channels/1047197230748151888/1176526177050054766 it might help.
wasn't there a button in some ai tool that could lengthen it just pressing it?
i wonder how that works?
What is the best setup for coding stuff?
Maybe in some other AI tool.
how about this one?
Hello guys, definitely perplexity's private threads didn't get monitor by NSA right?
Lol
You need a computer, keyboard, mouse, and a monitor.
You can get these all in one as a portable device if you purchase a laptop.
what's up with the glasses
funny af bro
if perplexity_pro:
select_model('claude2.1')
print("do you still need to subscribe to claude2.1 for minimal usage? š¤")
was considering to get claude pro but this pops into my mind
looking at gpt 3.5 turbo, 1 million input tokens would cost 0.5 USD meanwhile 1 million output would cost 1.5 USD, and this is the cheapest base AI I can find on the pricing page
Meanwhile perplexity's api cheapest offer is 0.07/0.28, am I misunderstanding something?
hi
How often do you guys use autopilot? I mean, is there any case that its better to disable?
I generally have it on (more sources the better + clarifing questions can be useful) but there are cases for having it off
Perplexity using smaller and cheaper to run models
Got it
I was looking at those extensions the other day that use AI but none of them seem to support perplexity API
so I guess I should look into making my own small tool
it would be great if the perplexity extension would support selected text operations
Iām using 5$ of monthly API credit to pro users to run simple discord bot for my server, itās nice
I did that also, but sadly I did not buy any host for the bot
so I just ran it a day on my pc to test it out
hosting the bot would cost more than what it uses for api haha
Yeah, Iām hosting it locally as my server is very small and bot is mostly a gimmick, The biggest usage was 0.03$ in month
I think mixtral is the most cost effective last time I checked
Makes sense
those AI extensions charge 12 usd per month most of them
with like 600 prompts or something like that
actually all services nowadays charge 10-15 usd for some gimmick AI feature
and when you take a look at the APIs, it is so cheap
Depends on how much you use the ai. If you are using it a ton itās cheaper to pay a subscription
yeah but from what ive seen subscriptions put a limit most of the times
is this like a good pricing? seems like terrible value
I guess most of the price goes for image generation maybe... But still 100 promps per month is total garbage
Hey, @quaint elm! You can ask Perplexity to help with that at https://pplx.ai or create a thread if you're asking other members of the community for help.
does anyone know the token limit for file uploads?
It can read many more with File Upload. Longer pasted text is converted to a file automatically.
File uploads can read up to 32k tokens by GPT-4 and 200k tokens by Claude 2.1 with a Pro subscription. To unlock longer context length, please subscribe to Perplexity Pro.
but I upload files that are 24k tokens and it only is able to parse up to like 16k tokens
both for gpt 4 and claude
does anyone know why?
The 7B parameter models are much smaller than GPT 3.5, which is 175B parameters
Oh I trump that @signal hamlet. But then, I am not a professional programmer but a Cybersecurity geek with some Python skills. Keep up the good work with Perplexity. it ROCKS!!
Hello i have an issue when I would like to pay the pro
seems like the paiment charge and discharge from my credit card...
What Gemini Pro model do better than gpt4 or claude2.1?
Gemini Pro is worse than GPT4 & Claude 2.1 from my experience. I'd say there's nothing Gemini Pro does better.
I would say Gemini Pro is better than any other free AI chatbot
But yeah, it doesnāt compete with the premium models
Gemini Ultra is releasing tomorrow though, so weāll see if Google can compete with either OAI or Anthropic
I don't. But there's this, which might bring so clarity in due course https://discord.com/channels/1047197230748151888/1191690469592289280
I would say the free chatgpt is still better than gemini pro
In most cases I disagree
Gemini Pro delivers stronger results on critical language skills and model performance benchmarks.
I think GPT-3.5 only has the edge in regards to impartiality and coding
Does anyone know the prompt to search for a journal in Google Scholar that can retrieve citations within the journal?
In what aspects
Does anyone happen to know if Perplexity's API is allowed to be used as a RAG method for existing helper bots?
As in, if I buy the Pro plan and get the API access, can I use it with my existing chatbots (I'd need to have a translation pipeline via OpenAI's API since I'm doing localized helper bots for mostly non-English speakers -- Perplexity queries are much more accurate in English).
From what I understood from Perplexity's overall ToS (I couldn't find a separate one for the API?), it might be possible if Perplexity is credited as the source (i.e. "Via perplexity.ai"), but I thought I'd rather ask first before getting the Pro plan and implementing anything...
Thanks for the help in advance!
Hi all
I've been super impressed with Perplexity's speed and accuracy, to the point where I barely use Google anymore. This made me wonder: with their vast resources and data, why haven't big tech companies like Google or Facebook developed a superior alternative? Given their extensive knowledge about users, they should be able to provide highly personalized results. I'm curious about the challenges or reasons behind this and hope for Perplexity's continued success.
I'm curious where you got the idea you need to credit the AI as being the source of any material ever? I realize the policies of these AI APIs disallow illegal usages, that break laws, and most of them are also very prudish about anything related to sex or even images of people, but I'd never heard of any APIs requiring apps to disclose where the AI material came from. Is this in the ToS?
I heard different numbers
I remember 175B being the number that was floating around at the very beginning, but with the increase popularity of chatgpt Ive seen different numbers
Yes, it is in Perplexity's ToS. https://blog.perplexity.ai/legal/terms-of-service
You agree not to conduct any systematic or automated data collection activities (including scraping, data mining, data extraction or data harvesting) on or in relation to the Service. Prohibited data collection includes, but is not limited to, using the Service as input into other services, websites, or databases. You agree not to republish any content generated by the Service without clearly citing the Service as well as the context associated with the content. Misrepresentation of the source of the content or the nature of its creation is strictly prohibited.
Hence I've been asking if it's a) allowed to pipeline the API through a separate helper bot that will translate the queries and b) should Perplexity.ai be credited in all of those responses ...
Hey Tyrone. The two products, Perplexity (an answer engine) and Google (internet search) are two completely different beasts. Perplexity provides you with answers and cites the sources from which it inferred the answers. Google searches the internet based on your search query. The search query then provides you with a list of things you could read that might have the answer. Google has a ranking system for what gets returned, along with copious amounts of advertising.
The Perplexity way is far smarter, saving you from clicking through conflicting search results to reach a conclusion. Perplexity gives you that conclusion, and you can fact-check from there.
I dont believe the big search engines like google et all like answer engines as thre is no real way to place adverts in the flow YET!!
Wow, it appears Perplexity wrote their ToS in such a restricted way it's illegal to use it for anything but creating a simple Chat Bot (meaning only things that are a back-and-forth between the Perplexity itself and humans). All other uses of the API appear to be illegal. That's pretty hilarious. I'm sure this was most likely done for competitive reasons (meaning, Perplexity has certain ideas for products they want to build, and are therefore making those kinds of products illegal for the public at large, so they will be the only ones having it).
This reminds me a lot of what OpenAI has done in their image recognition to make the OpenAI unusable for any kind of OCR (Optical Character Recognition) for any information containing people's names, companies, phone numbers, emails etc. OpenAI made this move, in my opinion, so that Microsoft Office products can have a monopoly on those kinds of AI-based OCR. So I think these restrictive ToS are just purely done for competitive advantage, and only people at CEO levels of these organizations even know this, and none would admit it I'm sure.
Thanks for the answer @fluid mauve and I like the idea of an answer engine. I haven't heard of that term yet.
Hey @blissful void!
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.
This monkey designed bot is actually the number one trending thing on the internet rn.. You actually tell them that it can't even hold contextual awareness for two chats before everyone gets excited \
No Worries. I can't take credit for 'Answer engine' I heard it from Riley Brown in a YouTube video I watched.
https://www.youtube.com/watch?v=aphHCBSTx7Q&t=344s
Main Takaways From this Video:
"I use Perplexity more than ChatGPT, BARD, and Microsoft Copilots for five main reasons, including its use in content creation and community automations. It's clean, sexy, and fast."
"Perplexity is incredibly accurate and has real-time internet access, which is extremely useful in my community. I set up an autom...
I have recieved the below email from OpenAi regarding their API service
No action is required from you at this time.
Please note that this change only applies to your OpenAI API account and Playground. It does not affect ChatGPT Plus subscriptions.
Best,
The OpenAI team```
I suppose its better to get credit in as opposed to giving credit
@severe magnet Ų§Ų±ŲŲØŁ
Is there a road map on when can the citations be available as part of the perplexity online model response?
I didnt know perplexity had a discord serverrrr
Perplexity is literally my favorite ā¤ļø
Does anyone know if you can use GPT-4ās code interpreter feature with perplexity pro?
currently, it's unable to accomplish that task
Iāve started using it for more random questions. Itās easier than trying to google what yesterdayās wordle was lol
was looking to create seo optimized artilces in perplexity. Anyone have any tips on creating articles in perplexity?
I've used this prompt for person use "Summarize the PDF into an easy-to-read article highlighting the key information." but for search optimizes articles I'm not sure
Hello, do you plan to include Gemini pro ultra in perplexity pro ?
I really hope they add both gpt4 turbo and gemini ultra. Feels like perplexities models have been stagnant since claude 2.1 and pplx 70b
nice! I'll give this a try. might just need to add on some more seo type prompt or messa around with collections to get a more optimized article
I did put some of the results into grammarly to check their score and claude 2.1 seemed to slightly edge out gpt4 most of the time. But try both of those and see what you prefer.
yeah for article writing claude seems to be a bit better but gpt 4 follows insturctions a bit better
I've found Claude to be easier/nicer to read. Though both end up being pretty similar with the prompt I provided.
I've got no ETA, but our API team is working on it.
Here are some good videos on this: https://www.perplexity.ai/search/Create-SEO-optimized-eYIeR07MStCIPU0PHr6NXw?s=c#79821e47-4ecc-4ad0-883d-4d0f1ebe8d5f please, take a look 
Sure, here are some SEO optimized article topics related to Perplexity:
- "The Power of Perplexity 3.0: A Comprehensive Review and Guide"
- "Maximizing SEO with Perplexity AI: Best Practices and Strategies"
- "Perplexity 3.0 vs. ChatGPT: A Comparative Analysis for SEO Content Creation"
- "Unlocking the Potential of LSI Keywords with Perplex...
Thanks for the answer.
Hey @frank elm!
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.
I don't know, on the other hand it could be that the ToS hasn't been updated in terms of covering the API, that it's the old ToS concerning the perplexity.ai itself. That's why I've been asking...
iirc it does say https://labs.perplexity.ai/ when describing the Services, which I assumed meant the API
I'm sorry but I am having this problem again.The app told me to check my emails for the sign in link. But when I used the link on my phone browser, it did not open the app. Instead it opened using my browser. This time I tried opened the sign in link in my discord server but it did not work. So frustrating with how this works.
Hard to say, but I have sent Perplexity an e-mail about this a few days ago, still waiting for a reply. š
good luck š tbh I'm not sure they sound that restrictive, as in you can can use it for whatever so long as not harmful etc. I don't know what constitutes 'automated', but it's an API..So I mean, clearly it's allowed to be used in some kind of systemic/automated way (right?). Only bit that gets me is needing to cite Perplexity whenever 'republishing' it's content - not sure what that means (or I gueess I mean not sure what the actual implications of that would be for most general usage)
Is Gemini ultra in roadmap for perplexity model selection?
Crazy question, but, how did you get the icons into the channel titles? Looks great BTW.
solved it!
thanks
Hey @gray warren!
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.
known bug on iOS, you just have to try several times. at least thatās what helped in my case
I'm using android
open the link from your mail app, first few times it will open in browser. and on the end it should get opened in perplexity app
The sign in link is no longer valid.
This happened after it opened 4 tabs in my browser.
make new one, rinse and repeat
Is there anything else I can do? Why not using an otp for signing in š
afaik, no
devs working hard, it will be stabilized in the future
Anybody try Gemini ultra yet?
anyone tried image gen on the paid version of gemini?
non-paid version still seems really slow at it
What would you ask it to generate?
any image really
But in terms of coding and productivity it seems to be really good
I hope perplexity brings the ultra soon
Hello, how could I contact support? š Thank you in advance.
Hey, @lethal bramble! Please send a message via Intercom or send us an email, you can find it at https://perplexity.ai/settings
I had seen someone telling at twitter that the images in the Advanced version still dont pass through the Ultra yet
what do you mean?
give me a min
Alright, thank you.
Hey @lethal bramble!
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.
š¤·āāļø
was it slow
took maybe 5 secs
Click on help and faq
"Frame is designed to be worn as a pair of glasses with a suite of AI capabilities out of the box. Whether your daily pair of specs of workbench prototyping tool, Frame is ready for the journey"
hmm
The more I have done with Gemini Advanced it isn't as good as gpt-4 turbo
It's cool and love the GUI though
aha all makes sense now #š¢āannouncements message
yeah agree, it's not as strong as gpt-4-turbo (though still a step-up for bard)
Oh, ok, @lethal bramble! Please send the details to support@perplexity.ai
Those glasses look cool.. and they are pretty cheap. I'm really sketched out by products like that from past failures from others so being tempted to preorder at all says something.. will have to think about it.
i guess if they are good they will be really good and take off and it will be easy to get them, even if its at a markup. and at that point its already been filtered out, all the people trying this where it didnt work out.
but i do believe in the idea eventually working and being huge.
but yeah, the pricepoint being where it is, and just overall i like this a lot more than the whole vr headset thing, i dont see those big bulky vr things ever taking off.
They look like harry potter glasses :< I love HP, but just the world, I wouldn't want to look like Harry xD
this is REALLY what I hope will be the future
I always would love to visit japan, but no way I gonna learn their "Art" Language š
not really?...
that looks so dumb xD like those:
š„ø
Here's the most problematic sentence in the ToS: "You agree not to republish any content generated by the Service without clearly citing the Service as well as the context associated with the content."
It basically means Perplexity gets free advertising on every app that uses it. You have to disclose the material came from Perplexity. I was doing that anyway, because I like that design, and my platform supports multiple different AI services, but I don't like having it DICTATED to me what wording I must use on app. I mean what if I'm using the app to help me write a research paper? I have to paraphrase everything that comes out of Perplexity even if the wording is correct OR ELSE include ALL PROMPTS I ever used directly into my research paper. That would be rediculous, and regardless of what ToS says NO ONE will do that.
I can see what perplexity is doing
They've been taking partnerships on these hardware startups like Rabbit and now Brilliant.
It's integrating itself into everything.
It makes sense, if they want to seriously challenge legacy browsers like edge or chrome, they need to be integrated into everything, just like their competition is
I think perplexity is already the most competitive AI client/browser, just because you get access to 4 different models for the price of one. Perplexity + Gemini Ultra would be a huge step, Iāve been very impressed by Gemini Ultra so far, and I think it might be good enough to top the current version of GPT-4 on the leaderboard
You can really tell how much of an improvement in regards to tokens and parameters GU is compared to GPT4
Gemini Advanced's memory is terrible. I wanted it to follow up on my fake coding lab, and it started to set up a fake science lab, not a coding lab, as we discussed before.
Perplexity (with gpt4) has generally been pretty close to what chatgpt outputs, but I was surprised here to see how way more useful the chatgpt output was. was it about how I asked my question here? or was it how perplexity relies a bit more on web resources over pre-knowledge.
I can't wait for pplx to get gpt-4 turbo
I suppose I havenāt tested itās memory, but itās outputs in regards to researching topics and language skills have been a much better time than GPT I think
That is where it is pretty good
It's memory is dreadful
Yeah, I think between the two itās just a matter of preference
Since I tend to use AIs more like research assistants than coding assistants, I lean towards Gemini Ultra
But I can see the argument for GPT when it comes to coding for example
I think both results are ok, but you can also use the writing focus in Perplexity, since your question does not require internet access.
Another thing I like Gemini ultra for too though is it hallucinates way less frequently than GPT4
GPT4 definitely improved over 3 in that regard, but it still is kinda noticeable
So far Gemini Ultra has done a good job at doing that less
I think it does a good job on searching
I agree
But, now that Gemini Ultra is out
Its OAIs turn to make a response, either via a GPT-4.5 model, or by releasing GPT-5 sometime this year
Thanks, the answer with writing focus did hew closer to ChatGPT.
TIL about the Focus options. I also learnt today about collections and the ability to add custom prompt there. But when starting a new query, I dont see an option to select a collection, and thereby the prompt. Is there a way to combine both these features? - focus + AI prompt?
Hey @gilded zealot!
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.
Yeah, you can click on the collection and then start a thread with a focus in the collection, but there is no default focus for collections (yet).
Tune In, Turn On, Drop Out
Announcement on what sir?
can you explain to me what that means sir? i cannot quite understand what is the impact
More opportunities for developers, but that doesn't affect you if you don't develop applications š
ah ok sir, so i should ignore it. hehe. thanks
Hey @cursive jacinth!
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.
Yeah, that sort of led me to think that maybe the ToS hasn't been updated for the API yet, since the API is in beta, but I thought I'd rather be safe than sorry... š“
tested Gemini ultra vs gpt-4 via perplexity, and had Gemini advanced, gpt-4, and mistral medium do LLm-as-judge and all 3 gave Gemini ultra the win. I used a snip-it from the Gemini 1 report from google. Of course this is but a single test I will likely do more.
Well since most of us have apps that end users generate content with we can just let the users take care of "disclosing to the world" where their content came from. lol. It's not like us app developers can control the end users even if we wanted to. So I can just put in my own ToS: "Please don't pretend you wrote something you didn't." haha.
Makes sense
Ultra is the new king of AI, at least for now
ultra has such a nice way of responding. the answers are so clean, easy to read. like a hybrid of gpt-4 and claude 2.1 the best of both
Yeah I agree
Gemini Ultra is superb in language understanding
It almost feels like talking to a human
Does perplexity has access to Gemini ultra already?
I can't see the option on mine. Only Gemini pro
Wait so you can just use ultra through perplexity without paying separately for ultra?
No, not yet anyways
Fortunately, Google is currently offering a 2 month free trial for Gemini Ultra
Hopefully within that timeframe Perp is able to integrate Ultra into their service
But once it is integrated
Probably
excited to see how Perplexity will integrate with Frame
Yeah should be cool
Afaik there isnāt an API available for Ultra yet (at least for public use). Itās just the end product right now through their Google One sub.
Hello people. Im new to AI. My understanding and experience is very limited, but pardon me, I don't think very highly of generative chat bots.
Just the fact that they simply absorb a lot of data and come up with responses to inputs is kinda uninteresting to me.
What are some forms of AI other than generative AI that's so popular today? And where else can generative AI be used apart from chat bots? Could you please direct me to some good sources to understand this?
I hope y'all wont mind the 'discord' I'm having with this. šš
Cheers
TT
computer vision, robotics, two minute papers on YT
It's not quite the same (the official model released yesterday is way more restrictive), but more or less the same Gemini Pro/Ultra is available here https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboard While it's 'publicly' available and clearly uses an API, it's an academic/research project.
I haven't seen any third parties with access to it and offering it commerically yet either (but I dunno, I'm a bit confused by Google's AI products/models and where they live - maybe it is available somewhere for Devs to play around with)
I can't understand/write computer programming code. Before these LLMs came around, I would turn away from whatever I was looking into the moment I saw 'python' or 'github repository', etc.. But generative chatbots have opened up that entire world to me (and I'm even learning a bit of coding along the way). Some parts of my research workflows are now basically totally automated, thanks entirely to the code generated by chatbots like ChatGPT - I look at some of my workflows and find it utterly staggering that I used to it manually (e.g. copy paste RSS feed or social media items into spreadsheets)
Sorry Perplexity, that was just a general gracias to all LLMs
So basically, a Jarvis for doing your google search for you?
Mmm, no more like a personal assistant with excellent coding skills in all programming languages, who I can give natural language requests to and receive the desired code in response.
Something like https://www.perplexity.ai/search/I-have-a-k313gBovTkCroW5MWwkFkg?s=c
Or https://www.perplexity.ai/search/Could-you-generate-YBAivf3wQW2qICOeXtloDg?s=c
Just examples I mocked up to try and demonstrate the point. The codes produced may not work first time (though they might - as they're pretty basic), but I would then give the error back to the LLM and ask it to identify the problem and revise the code accordingly. It's been massively empowering for me (and made realise how dumb I was to learn a foreign language ā German lol ā at school; wish I had someone push me to just learn basic python back then, it's mindblowing what can be achieved with a few lines of code)
Certainly! Below is a Google Apps Script that you can use to deduplicate rows in your Google Sheets spreadsheet based on the title (Column A) and URL (Column D). This script will keep the most recently added version (assuming the most recent items are at the top) and remove any older duplicates.
To use this script:
- Open your Google Sheets d...
Certainly! To achieve this, we'll use Python with a few libraries: requests and BeautifulSoup for web scraping, and pandas for data handling and saving to a CSV file. If you haven't installed these libraries yet, you can do so by running pip install requests beautifulsoup4 pandas.
Here's a step-by-step guide and the full code to accompl...
And for "AI other than generative AI", it's hard to look past what has been achieved in the field of protein / drug discovery over the past few years https://www.perplexity.ai/search/What-has-AI-jaXRO5VxS8O0tLgwp8dhpg?s=c
Accelerated Protein Structure Prediction
AI has revolutionized the prediction of protein structures. DeepMind's AlphaFold, for example, has transformed the protein structure solving problem that has been around for the past 50 years into a seemingly trivial task. AlphaFold can predict near-native protein folds from their genetic sequence, which has been a significant breakthrough in structural biology.In 2021, AlphaFold released predicted structures for about 1 million proteins, including almost all known human proteins, and later expanded this to around 200 million proteins*
- Basically, scientists had predicted 194,000 proteins in the decades leading up to 2021; then, using AI they predicted 200 million proteins in a single year
I want to use rhe really powerfull perplexity search and the really powerfull perplexity copilot alonf with gpt 4. How can I do this over the API?
to my knowledge only PPLX online models have access to the internet over API.
What is Perplexity copilot then that searches the internet with a lot of sources and puts the sources in the answer?
to my knowledge copilot uses gpt-3 to understand the question. But neither copilot or internet searching access are available through the api apart from the PPLX-online models.
Good morning. Guys, is Gemini Ultra already available on Perplexity Pro?
Will Copilot be availabe for external services like TypingMind in the future to be integrated as search engine?
Wouldn't that be something to request on TypingMind? Thanks for the introduction, btw. It looks like a cool UI. And TypingMind leverages APIs, so it seems just a matter of implementing Perplexity's APIs...
Hey @worn monolith!
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.
Oh! Actually, you can already do that: https://docs.typingmind.com/chat-models-settings/use-with-perplexity-ai
Itās easy to setup Typing Mind for using with Perplexity AI (https://www.perplexity.ai/) Here is a quick guide. Go to https://www.perplexity.ai/ and sign up for an account. Then go to https://www.perplexity.ai/settings/api to get your API key.
mmmm
Appreciate the clarity! Perhaps itās just not āproduction readyā yet
perplexity ai or bing ai
PERPLEXITY mate
there is no real comparision fam
Copilot is neat but temperamental lol
which copilot? I think every AI company in existence has product called "copilot". The word has become synonymous with "AI Assistant"
Bing Chat
Is anyone familar with an AI tool for parsing videos to counter clickbait? For example, if a video title is "Top warning signs you cannot ignore for brain health" I would like to be get a summary with the specific points and timestamps. I have experimented with notegpt youtube summarizer, but very mixed results. Perplexity/copilot does not seem to be able to parse the video itself.
Gemini advanced canāt render latex lmao
Maybe try gemini
Using the YouTube extension
gemini is probably the fastest but you could also just use any old ai transcriber and put it into gpt
Thanks, sometimes the obvious solution is right under my nose.
Hey @gaunt shore!
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.
Gemini is being the most frustrating. I'll take a look at the youtube extension later. NoteGPT does a good job of providing transcripts. Perplexity and copilot both seem to be effective at parsing the transcript and providing the summary with timestamps.
Anyone else have the issue where they type something to perplexity and then it just doesnāt respond? The I refresh the page and only half of the time is the answer actually there
I have this issue, I have to go to my threads and collections and find my message there to get it back
Yeah. Super annoying. Iāve filed a bug report a few times and hope it gets resolved soon
Not sure where to ask this, but how do you keep previous responses to a prompt when switching between GPT4 to Claudia 2.1 and other models. I donāt want to lose the already generated responses but just to read and compare answers using different models to get different nuances.
anyone has a prompt to be able to write 2000 words at least ?
Strongly prefer perplexity
open 2 separate tabs, ask same question to diffrent models. only way to compare atm
Perplexity unless I want images, then microsoftener coparent
Absolutely love Perplexity, but the iPad app is my only gripe. It has the floating window for current thread, when holding the sides of my iPad I constantly tap outside of the window which exits the thread. It seems overly sensitive.
hello
hi, do we have any channel for developers using API of perplexity?
Same feeling for me.
Almost one month in, and I'm still perplexed by how we can't copy a part of the response but have to copy the whole payload
this is a known bug in ios and will be fixed
We have add Perplexity to our site. Welcome to see how perplexity performs against other AIs answers. https://www.aiswers.com/
Also, Perplexity page is here https://www.aiswers.com/bot/perplexity
No new iOS beta testers are currently being accepted.
:(
Anyone know what to do or how to correct this error when using the API? I'm using it through App script.
Error: HTTP status 429, Response: {"error":{"message":"Request rate limit exceeded, please try again later.","type":"request_rate_limit_exceeded","code":429}}
how do you get in touch with support?
Are you running it on a bunch of cells? Probably hitting the rate limit as it says
Based on your roles it looks like youāre a Pro subscriber. If you log into Perplexity (either via the web or app) you can head to your account settings and youāll see āPro Supportā
The iOS app will provide you with the Pro Support email while on the web itāll open a page to send a message/open a ticket.
If none of the options above are available you can message support@perplexity.ai
aah thanks
Hey @tame current!
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.
Hey guys! My name is Parv and I am a High School Senior interested in AI and Tech! I would love to work with and learn from the talented engineers at perplexity. Is there any way I could do this while in High School? Are there any roles available for internships for high schoolers?
Hello
may i know why perplexity is so struggling with long response that generated from the model (gpt 4 or claude)
(it gave something went wrong after wrting, then gone, then you cannot view this thread)
just try gpt 4 and claude on Poe
their site is able to deliver that very long response i ask
1000 word split into 10 part is easy for the 32k Gpt4 or claude for their specs
500/1000 word for 1 part
make it 10 part
it should be like that,
on perpllexity, the site is broke when generate long answer as i mentioned above
poe able to delver it properly
hey guys my name is Dano and i feel scammed as hell, so i buyed pro version of perplexity ai and choose "chatgpt4" as model, i asked the chatbot in writing mode if he is chatgpt 3.5 or 4 and he replied that chatgpt 3.5 is what he is. I asked them who was the german chancellor and he told me the old one that left 3 years ago like wtf, if you dont have chatgpt 4 than dont say it
it's still gpt4 the model itself just doesn't know it
Is there unlimited in gpt 4?
im sorry but this is peek manipulation here is why: The Chatbot itself says that his data last until september 2021 which is right because if you ask him who is the current German chancellor he replies with "Angela merkel" who is not anymore. Now here is my question: Where is the diffrence between "Chatgpt4" with data reaching only to 2021 and chatgpt 3.5 with also data reaching until 2021,
gpt4 is smarter basically
but its not chatgpt 4, if i where to go to openai and buy chatgpt 4 there it would have new data, it would know who the current chancellor is lmao
Chatgpt uses GPT-4 Turbo 32k cutoff 2023, Perplexity uses GPT-4 8/32k cutoff 2021
GPT-4-32k model has a knowledge cutoff in October 2021
"As an AI developed by Perplexity with a knowledge cutoff in September 2021"
*September
i just used copilot to confirm it
I don't quite understand what you're trying to say š¤
perplixity AI has copilot integrated, to be honest this is the reason i wont cancel my subscription. "Copilot" functions as a high quality research assistance, which means if you dont wanna read trough 1000 of articles just for knowing one thing, you can just ask copilot and it will do everything for you. i asked it if he can look up the data of GPT-4-32k and he told me that it last until octoboer 2021. he also gave me sources so yeah i dont think it is lying
look my problem is not the data i mean perplexity allows the chatbot to connect with the internet so it doesnt make any diffrence but i can not believe that this is chatgpt 4 with the smartness of chatgpt 4. to be honest your "prompt manipulation" treat got me more suspicious because you basicly are saying "Yeah its chatgpt3 but you can tell it to be chatgpt 4, you can even tell it to be john cena" yeah but than its not REALLY chatgpt 4 or john cena is it?
I didn't say that, please read the post again, it's all explained very clearly and in detail š
yeah i am sorry to have bothered you so much its just that i am a really sceptical person. But thanks you managed to blow my worries away
Hey @rigid hull!
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.
So, are there any news on whether or not the Perplexity API can be used for localized translations, as long as perplexity.ai is cited as the source?
It's been almost a week now with no response in mail after I asked about this from Perplexity.
I'd be very happy to start developing but it's good to know for sure what is allowed in terms of the API (i.e. machine translations) and what's not...
Hey! š hope everyone is doing just fine! here to wish a great day, thats all lol.(and to explore this discord server + I am a Perplexity addict. I guess that's one thing everyone share, am I right? š
Hi all, has anyone been able to find information about when perplexity will switch to use the latest OpenAI model (turbo)?
Will depend when the company will release the model api.
LLMs are pretty darn cool
When is perplexity moving to its homegrown LLM?
And any plans to integrate a maps feature in perplexity search?
what is the context length with perplexity pro for gpt-4?
still in beta, only for android users atm
is it gpt4 or gpt4 turbo
so does pro get gpt4 and non pro get gpt 3.5?
more importantly, you get Experimental with PRO #šāpro-lounge message
How to use this map feature sir?
It says no access sir
What does it look like?
That is what i want to know sir. I did not know there is such kind of feature
Never seen it
still in beta, for android beta testers only
how to be a beta tester sir
and how to use sir
i am an android user sir, can i be a beta tester sir?
#š¬āgeneral message Android beta is also closed
Can you take a screen shot
Oh theyāre using mapbox
Hello, I did not get any response yet from my email, how can I directly contact anyone to get help? It is the second time I tried sending an email.
Hi everyone 
hello
You can also ask your question here
it's about data related to my company so I'd rather not, but thank you š
The FAQ can't help you?
No, I would not ask otherwise
Gemini adv sux at coding
why do people pick perplexity over phind pro? Phind uses internet also and gives better ans imo then perplexity
and phind can code too, i mean for me phind was the clear choice over perplexity
for me perplexity give better search result as search engine sir. I dont do coding that is why i prefer perplexity though.
Don't know about coding either (though think that's meant to be Phind's niche, so wouldn't be surprised if it outperformed there), but for general web search.. apples and oranges
phind gives great ans on normal non coding quesitons also is why I chose it over perplexity
but maybe i'm missing something i learn by being corrected lol
Mamba大čÆčØęØ”åå¹¶äøęÆäøäøŖē¹å®ēć广为人ē„ēčÆčØęØ”ååē§°ļ¼č³å°åØęēęåę“ę°ę°ę®ļ¼2021幓9ęļ¼ę¶ęÆčæę ·ćåØäŗŗå·„ęŗč½åčŖē¶čÆčØå¤ēé¢åļ¼ę许å¤äøåēčÆčØęØ”åļ¼å®ä»¬åęē¹č²ļ¼ęØåØč§£å³åē§åę ·ēé®é¢ćä¾å¦ļ¼GPTļ¼Generative Pre-trained Transformerļ¼ē³»åćBERTļ¼Bidirectional Encoder Representations from Transformersļ¼ćXLNetēé½ęÆéåøøčåēčÆčØęØ”åļ¼å®ä»¬åØå¤ē§čŖē¶čÆčØå¤ēä»»å”äøč”Øē°åŗč²ć
čÆčØęØ”åēåŗę¬ē®ę
can we
can we update GPT-4 model to newer ? it said it use old model.
the newest model has knowledge till 2024.01
Hello, I'm new here and a newbie Perplexity user. I was hesitating whether to subscribe to the pro version, for some aspects I find Perplexity very interesting but on the other hand its free version is still very imprecise or incomplete... I was wondering if the Pro version guarantees a significant improvement (today, for example, the free version was not aware of the existence of the pro version!).
I havenāt found any incomplete or wrong answers in the free version
Except once when it relied on a government page that itself was out of date
Oh! For me, half of answers are not useful... an example, today I asked the same question about musicians date tour: Perplexity said there is no info, Gemini give me all dates around the world
(At the moment I am quite comfortable with the free version of Gemini)
I'm perplexed š§ š¤
Feel free to share the relevant threads here āŗļø
I understand now: I have to write in English: https://www.perplexity.ai/search/c49f98e2-f332-4402-86d7-b4eb5388e83b
is there something wrong with the webapp at the moment? I'm logged in and asking perplexity questions just on the default prompt entry and everything just goes to loading answer and nothing comes.
I get that sometimes. Not sure if it's a browser thing. I'm using Edge at the moment. Refresh the page or open a new tab and go back to your library list, it should be on top. This usually fixes it for me.
It just seems to not rerender the page with your query sometimes.
Hey everyone, just wanted to share that today I needed to debug a R code and I tried GPT-4 (openai's website), Gemini Ultra 1.0 and perplexity. GPT and Gemini weren't solving.. the one that solved it on the second prompt was perplexity with Claude 2.1 model.
Honestly, I think perplexity have chances to be the best A.I. on a global and market scale plus get the proper recognition. (My opinion is that it actually is the best, hands down).
I'm having a hard time with model accuracy, I just tried out pplx-7b-online and pplx-70b-online, which I'd assume are the online models that are pretty much realtime, or ..? they did get some facts right that have happened in the past few days, but very odd hits and misses
I'm trying to supplement a helper bot so that it could query perplexity via the API whenever it needs to be up to date...
Has anyone found the Claude 2.1 has been really slow the past week? I like the model but it's unbearably slow compared to the others.
I personally would use GPT-4/Claude 2.1 for accuracy. API wise neither have access to the internet though so ehhh hard
I have written my own front end for the perplexity labs models, i use codellama-34b-instruc for code questions and it seems pretty good. granted i have also had decent results with GPT4
been using tic-tac-toe as a logical benchmark for LLM's. Unfortunately the new PPLX MoE model can't seem to play tic-tac-toe. Edit mistral medium can easily play tic-tac-toe though it sucks at it the board is always perfect.
Whatās that?
a type of search on google or other search engines
This is quite 'illuminating' - https://www.techtarget.com/whatis/definition/Google-dork-query
nice. if you feel comfortable sharing it with me, I would appreciate it. educational purposes only, of course
Ive dm'ed you
Same here
Hello Im trying to subscribe to Pro, but it seems I dont have the 7 days free trial (never tested before)
the 7 days trial is only for whitelisted email domains.
can you explain sorry
for example @outlook.com @gmail.com this is done to prevent abuse
im using gmail
there might be some other things that also play a role here that I dont know about. please contact someone from the perplexity team they will help able to help you with this issue.
how can I contact someone? Directly here on discord?
I think they will see it
they'll catch it once they're online. in the meantime, you can scope out referral program team has set up for everyone
Neither the 7b-online nor the 70b-online versions? Well, I guess the names were misleading ...
They did have some data on stuff that happened a couple of days ago, but nowhere nearly as accurate as perplexity.ai on the website
down?
503 service unavailable
yes
503 serive unavailbel
@wheat granite @toxic moss ?
Same here
Thanks everyone for the notification. We are fixing this issue
seems to be back
Thanks for confirmation! Let us know if there are any issues.
I've got 2 Questions:
-
Where do I input my Rabbit R1 Redeem Code?
-
I noticed there is now the option to select en Image generation option, but when I ask the prompt to generate me an image it says it cannot help me. Only when I press the GENERATE IMAGE option after the first prompt returns a failed ressponse do I get an image...
is there any other way to do it?
Hello, I wondered if you planned to implement a "Code Interpreter" that executes Code (the same type as we have in ChatGPT Plus) ?
Hello! I have definitely verified that inputs in Spanish in most cases return insufficient or wrong answers. Is there something I'm missing?
Hey, @proper plover! Please share some threads you found incorrect, with all the relevant details in the #1140622008086970420, we'll check.
Thank you Alex šš»
Hey @proper plover!
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.
Hello, can someone tell me when do perplexity tokens refresh. It says that I have 567 tokens for today left. But when does it refresh?
each use comes back 24 hours after use
I mean, perplexity's reply stops in the middle when the reply includes "<|im_end|>"
I love perplexity š„°
Me too
hello muchachos. just making sure i understood this correctly, i can use the API solely as pay-per-usage, without paying any monthly expense, yes ?
Does it refresh to 600 again?
24hours after your last use it will be 600
so if you send a prompt at 7am youāll get that use back the next day at 7am
Thanks)
Hey @spring vapor!
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.
Two things jump out at me in the Gemini FAQ page. First, in answer to the question "What are Google's principles for AI Innovation?" they say nothing directly about achieving a high degree of factual accuracy. One may generously infer it as implicit in their stated goals, but if they don't care enough to state it as a core part of their mission, they clearly don't care about it very much. Second, in answer to "Is Gemini able to explain how it works?" they go to extremes to urge people to "pay no attention to that man behind the curtain." Personally, if they urge me to use an information source that they disavow when it comes to their own self-interest, I don't want to use that platform for anything of importance to me.
https://gemini.google.com/faq?gad_source=1&gclid=Cj0KCQiAw6yuBhDrARIsACf94RXwwoXpktZalDwo6OO8RsVYvKAaDpxT1Cr_XIek-8kBnPaZa7Jb5bwaAvsQEALw_wcB
Their google one subscription seems like a good deal tho
And that's the only thing that picked up my attention
dropbox2tb = 10 usd, mega2tb = 10 usd, icedrive1tb = 6 usd (12 usd for 3tb)
also it has implementation for gmail and docs
It is a pretty good deal if you are into the google ecosystem
I wouldn't say I'm impressed by what they have right now, they seem to imply that more implementation is coming soon, because they also have a extension tab that is quite empty
They have some small features that are nice like drafts and modify response option but it's nothing mind-blowing
accepted, bro
gemini can't do basic assistant stuff like set calendar stuff
another failed product launch
talking about advanced btw
which model is better, Perplexity or Experimental?
experimental is pplx-70b but it doesn't say how many parameters "Perplexity" is
guessing it's 7b but that's disappointing
LLMs are just unwieldy ha
do you mean the default model?
the default model is gpt3.5
I think Perplexity is my new favorite chat interface.
Are you also coming from ChatGPT?
yes
I've been using GPT since 2020, so I feel like I've just kinda stuck with them.
Got the Rabbit pre order and the partnership with pplx and rabbit brought me here.
best tool ive used.
https://thenewstack.io/more-than-an-openai-wrapper-perplexity-pivots-to-open-source/
depends on your needs and goals
it would be nice if it said that instead of just "Perplexity" model
so is that the same that the free tier uses?
Can Perplexity completely substitute ChatGPT subscription?
Iād say ya
Although Iāve never had a ChatGPT sub lol
deep dive ā¤#š¬āgeneral message
im curious on perplexity as a code assistant. ChatGPT is pretty good. I assume using the GPT 4 model would be similar but with additional searched context?
interested on this as well!
when set by default does it provide an answer based on the best model to the question?
My Pro subscription just went away but I donāt think it has been a year yet unless I am wrong.
Is Perplexity down? Or is it just me
The bot is stuck in a loading screen, even when I restart and disable/enable copilot and switch models
Is Perplexity is Down?
I really like the voice interface of the mobile chatgpt. I dont think that is here yet. That and the possibility of Dall-e kept me on it. Finally cancelled subscription after installing Librechat. It gives me Dall-e access. And with the frequency at which I am using it, the pay as you go model might turn out cheaper for me.
Anybody with some perplexity knowledge that can answer a very specific question?
It was for a bit for me too
I donāt have pro though
Whatās the question though?
If you don't have a pro, I doubt you can answer
Its about pplx-8x7b

