#💬│general
1 messages · Page 166 of 1
Ah I thought you were talking about Gemini
The YouTube sources on Perplexity is primo
Gemini is 5 per month. Grok gives like 5 per 2hrs or something like that.
the quality is completely different though
grok deeper search is more like pplx DR while Gemini is on a whole other scale
The rate limits make up for it imo. If I kick it around a bit, it gives me what I need. Can't do the same with perplexity.
Yea honestly I only see me having about 5 earth core searches a month anyways…. And I have like 5 gmail accounts lol
Wait. I will explain with an actual example.
Historical topics are better on GPT but the current topics fare better with perplexity I find
I compared a deep search with perplexity on the resurrection of Jesus and they both said the same thing but GPT just elaborated on each topic more.
Grok sucks at one shot but multiple prompts does the job.
Based on bible or assumptions/guesses?
Grok + Perplexity UI > GPT
Based on historical evidence and apologetical arguments.
Wait so you fire up a convo, talk for a bit, then hit deep research when you have put together more context?
I give it a task. It misses out stuff. I cuss and abuse it a bit.
Eventually it gets me what I want.
Is it better than with gemini 2.5pro for general questions? Someone here told me gemini 2.5pro is the best but i dont understand these things reallly
This thread will explain.
I guess this will do for the next month…. Going to try perplexity pro and not resub for GPT.. we shall see how it goes.
Just wait till grok is on your desk lol
Depends on use case. If there's no math or engineering complexity then Gemini and grok are good. The grok on perplexity is the non-reasoning one.
I apologise too sometimes, lol
What is this app
Kortex ai I guess.
For no math or anything like that, just general regular questions, what is better gemini or grok? I just want to set the best one and forget about it
Gemini
Thanks!
Gemini is very elaborate in its approach to what it allows the user to tweak which poses changes for intuitive UI
Whats its gimmick?
I have no idea. Ask him.
So basically gemini is the best for perplexity general questions from what i understand thanks!
No it’s a note taking app… pretty new. It’s designed for writers, thinkers, etc. you can chat with all your docs. It’s new, so kind of still in beta.
Better than notion?
Oh! I’m sorry are you asking which model to use inside perplexity?
I’m so biased against notion… so yes.
Whys that
Too clunky for my tastes. Too complex and not intuitive. Especially for mobile. Which is important for me.
I like notion since it works well with notion calendar
But I don’t need database…. Definitely cannot do what notion can do. But it handles note taking and second brain loads better since you can use it as a knowledge source for your AI chats
You know if gemini 2.5pro is the best for general questions in perplexity?
I use craft docs for my planning and calendar… its daily note is unmatched.
I do need the database so that i can sync my notion tasks with my calendar
Craft also crushes the tasks in notion.
kortex ai?
Not available on android sadly
Yea
Awwwww man.. I was hoping you weren’t going to say that. Web version wouldn’t be fun to use but it works.
But its strength is the offline mode.
Yeah lol im not into using web versions of apps too annoying
But notion works great for my use case so its good
Then don’t try Kortex AI until it’s fully released for mobile lol
I was also talking about craft docs its also not on android
Both kortex and craft dont have an android app
Pretty dumb imo
Kortex most likely will. They haven’t even released iOS app yet..
They likely went for making native apps then
It’s only web app for mobile
And so the android team is nonexistent or behind...
The only thing that i want is an app that can automate and run specific prompts for me like an agent
I wasn't too huge of a fan of craft when I tried it
Like background tasks?
I’ve been having a lot of fun with Raycast to do things like this
How long ago was that?
oohh I love raycast 🙏
last summer or september, I forgot
No like having the same prompts in a preset and i like upload a file to the preset and then it runs the prompts until i get a final response after all the prompts were ran
GPT-1 will be coming out soon
oh hm
So like advanced templates, with multiple messages or options too
some form of workflows basically
Not on windows lol
Oh.. the new calendar and tasks is super nice…. And they have a huge update tomorrow for search features.
I just use it for planning, sermons, and projects. Being able to see every single task from all my docs in one place is a life saver.
Dang…… grim. But at least you can play Star Citizen lol
Can’t believe people let me use Alfred for so long
Something like that, not really advanced, like i upload a .csv and then prompt 1 is ran and then prompt 2 and then 3 and then 4 and then until the last prompt is ran
can't believe there are still people using alfred 💔
hey in the perlplexity hackathon, do we need to make a web app, or a proper app?
Not into p2w games lol
Lolllll barelyyyyy… they said sorry I think 🤣
Yeah its only 25k per premium package lol
Saw someone call it Store Citizen… hilarious.
It’s a beautiful game, though. Can’t doubt that.
One of its kind.
Beautiful it is, fun... cant say that imo
Too early to really enjoy it unless you have some friends to traverse with. I don’t even play it lol I just watch
I've got it on windows myself, a wider release was supposed to be in q2 but I doubt that
I am looking for some fun games to play but get bored of most quickly
Only really enjoy playing civ 6 and fifa lately
Yea… I think I’m phasing out of games but deep inside i don’t want to let go lol…. It comes in waves… usually through WoW which never ever lasts more than a month.
I just watch tv shows while working since games are too boring for me more and more as time passes
Legacy: (something) and sorcery looked fun
I am getting into vr games lately and been having fun
Still exploring new games luckily i get to work from home most of the time so more free time to have fun with it
Hey mat
So apparently Gemini got nerfed
Any info on that (and source)?
Difficult to link is it...?
I don't get it.
Isn't may version supposed to be better?
Atleast the benchmarks say so.
Vibe wise some people prefer the old one it seems
But again people take for granted the things they have, people complaining about the old one being gone when they could have just had nothing, since it was all for free in the first place (at least on the small scale developer API)
Yeah
not really
sacrificed everything else for coding
@distant aurora tried grok Dr again and it did it's research pretty thoroughly
Researching and comparing player engagement between 2 MMOs
Perplexity just looked at steam charts
Grok looked at twitch views, subreddit activity, official forums and steam charts
Yeah. Grok digs a lot.
Grok DR is kinda interesting in how it works
Like on the surface it feels like they just gave a brave search tool to the model and let it go wild but
There's surely something beneath that
Not sure...
Grok seems to go more in-depth with sources
It's capable but it holds back.
So sometimes it's not a whole lot more sources than ppxl
But it digs thoroughly through it
It eventually gives me what I need if I abuse it. Why it doesn't do it from the start is what pisses me off.
You can't do the same with perplexity. No matter what you prompt it, it's limited.
Probably a cost cutting measure.
Perplexity can’t create charts right ? https://x.com/testingcatalog/status/1924000725807366147?s=46
There's a high variant that can. But it hasn't been released to most people.
Someone shared a thread a while ago.
ill be real lads is ai really that traceable
Yes, but with the same AI you can make it untraceable
Ive written my own piece of work and written an ai version and software is telling me my version is almost certainly AI
@pliant axle
Not all of them have the same accuracy or quality
how essentially, since its ai generating the responses?
TurnItIn
Free version? It sucks, try any other detector
GPTzero
also did the same
Try originality
Yeah
free?
You also have to take into account something, how AI is to human eyes
Yeah
It will force you to login after some time so use in incognito.
That way they can't track your cookies.
But if you want you can let me track your cookies to me to help you
meh says i have to sign up

I used it without signup multiple times.
yeah its not letting me
ive signed up anyway
neko ban @tame current
wait what?
@tame current
kiwi browser in 2025 is crazy
Last update was in may 2025
appreciated brother, my dumbass was doing some next stupud shit
Lol.
was clicking in the wrong place
That was a mistake, the last update was in 2023
hahahaha
bruv i just used 100 percent ai text from perplexity, shoved it in and it said it was human
like straight up
Nobody told you? Perplexity doesn't use any AI, they're all Chinese typing from their homes
nah g be serious
these things keep telling my own work is AI but when I go onto perplexity and paste it says not AI
Hmm, well, it really comes to have a natural language within everything, and if I leave sources, it will say more that it is a human text
You are the AI...
im a frickin bot
i just wanted to ask how good these ai checkers are because if i get flagged on turn it in
I just realised where I heard belvedere
ayo
valverde
baller
They are not really that accurate, but they can help...
this is for my uni btw
fuck it im not worrying god knows i havent cheated
Teachers should not use detectors, after all they are not all accurate and there may be errors or confusion
They're not paid enough for that much scrutiny, lol.
hahaha
TurnItIn does it for them
wym fonts
Tell deepseek to make you a human text, and follow the process by telling it what AI marks you, etc., until it marks 100% human
on perplexity?
If you want
cheers guys
@pliant axle
What's apa?
I'm pretty sure you can bypass this if you use homographs.
It would be all gibberish to ai. But reader will read in English.
can u use deepseek on perplexity
APA format is a standardized writing style used in academic and scientific writing, featuring 12-pt Times New Roman font, double-spacing, 1-inch margins, and a title page with the paper title, author, and institution. It includes in-text citations (Author, Year) and a reference list in alphabetical order with hanging indents. Headings organize content, and language should be clear, concise, and unbiased. (Based on the 7th edition, 2020.)
Yeah that probably won't work.
ˢᵉᵗᵗⁱⁿᵍ ᵗʰᵉ ˢᵗᵃᵍᵉ ʸᵒᵘ’ᵛᵉ ᵖⁱᵗᵗᵉᵈ ¹⁰⁰ ᵃᵛᵉʳᵃᵍᵉ, ᵘⁿᵃʳᵐᵉᵈ ʰᵘᵐᵃⁿ ᵐᵃˡᵉˢ ᵃᵍᵃⁱⁿˢᵗ ᵒⁿᵉ ᵃᵈᵘˡᵗ ᵐᵃˡᵉ ˢⁱˡᵛᵉʳᵇᵃᶜᵏ ᵍᵒʳⁱˡˡᵃ ⁱⁿ ᵃ ʰʸᵖᵒᵗʰᵉᵗⁱᶜᵃˡ, ᵒᵖᵉⁿ-ᶠⁱᵉˡᵈ ᶠⁱᵍʰᵗ ᵗᵒ ᵗʰᵉ ᶠⁱⁿⁱˢʰ. ᴸᵉᵗ’ˢ ᵇʳᵉᵃᵏ ᵈᵒʷⁿ ᵗʰᵉ ᵏᵉʸ ᵛᵃʳⁱᵃᵇˡᵉˢ ᵇᵉᶠᵒʳᵉ ᵖʳᵉᵈⁱᶜᵗⁱⁿᵍ ᵃⁿʸ ᵒᵘᵗᶜᵒᵐᵉ. --- #### ¹. ᴾʰʸˢⁱᶜᵃˡ ˢᵖᵉᶜˢ | | ᴬᵛᵉʳᵃᵍᵉ ᴬᵈᵘˡᵗ ᴹᵃˡᵉ (×¹⁰⁰) | ˢⁱˡᵛᵉʳᵇᵃᶜᵏ ᴳᵒʳⁱˡˡᵃ | |----------------|---------------------------|--------------------| | ᵂᵉⁱᵍʰᵗ | ~⁸² ᵏᵍ ×¹⁰⁰ ≈ ⁸ ²⁰⁰ ᵏᵍ ᵗᵒᵗᵃˡ | ¹⁶⁰–²⁰⁰ ᵏᵍ | | ᴿᵉᵃᶜʰ | ~⁰.⁷⁵ ᵐ | ¹ ᵐ+ (ˡᵒⁿᵍᵉʳ ᵃʳᵐˢ) | | ᴿᵃʷ ˢᵗʳᵉⁿᵍᵗʰ | ᴮᵃˢᵉˡⁱⁿᵉ “¹×” | ⁴–⁹× ᵗʰᵉ ˢᵗʳᵒⁿᵍᵉˢᵗ ʰᵘᵐᵃⁿ | | ᴮⁱᵗᵉ ᶠᵒʳᶜᵉ | ~⁶⁰⁰ ᴺ | ¹ ³⁰⁰–² ⁴⁰⁰ ᴺ | | ᴰᵘʳᵃᵇⁱˡⁱᵗʸ | ᴴᵘᵐᵃⁿ ˢᵏⁱⁿ & ᵇᵒⁿᵉˢ | ᵀʰⁱᶜᵏᵉʳ ˢᵏⁱⁿ, ᵈᵉⁿˢᵉʳ ᵐᵘˢᶜˡᵉ & ᵇᵒⁿᵉ | ᴷᵉʸ ᵗᵃᵏᵉᵃʷᵃʸ: ᴼⁿᵉ ᵍᵒʳⁱˡˡᵃ ᵐᵒᵛᵉˢ, ʰⁱᵗˢ, ᵃⁿᵈ ᶜˡⁱᵐᵇˢ ˡⁱᵏᵉ ᵃ ⁴⁰⁰-ˡᵇ (¹⁸⁰ ᵏᵍ) ᵖᵃʳᵏᵒᵘʳ ᵉˣᵖᵉʳᵗ ʷⁱᵗʰ ᵗʷⁱᶜᵉ ᵃⁿ ᴼˡʸᵐᵖⁱᵃⁿ’ˢ ˢᵗʳᵉⁿᵍᵗʰ. --- #### ². ˢᵖᵃᶜᵉ & ᴱⁿᵍᵃᵍᵉᵐᵉⁿᵗ ᴰʸⁿᵃᵐⁱᶜˢ • ᴵⁿ ᵃ ¹⁰-ᵐᵉᵗᵉʳ ʳᵃᵈⁱᵘˢ ʳⁱⁿᵍ, ᵒⁿˡʸ ¹⁰–¹⁵ ᵐᵉⁿ ᶜᵃⁿ ᵖʰʸˢⁱᶜᵃˡˡʸ ˢᵘʳʳᵒᵘⁿᵈ ᵗʰᵉ ᵍᵒʳⁱˡˡᵃ ᵃᵗ ᵒⁿᶜᵉ. • ᵀʰᵉ ᵒᵗʰᵉʳˢ ᵐᵘˢᵗ ʷᵃⁱᵗ, ᶜᵃᵘˢⁱⁿᵍ ᵃ “ᶠᵘⁿⁿᵉˡ ᵉᶠᶠᵉᶜᵗ” ᵗʰᵃᵗ ⁿᵉᵘᵗʳᵃˡⁱᶻᵉˢ ᵗʰᵉ ʳᵃʷ ¹⁰⁰-ᵗᵒ-¹ ᵃᵈᵛᵃⁿᵗᵃᵍᵉ. • ᴬ ᵖᵃⁿⁱᶜᵏᵉᵈ ᵐᵒᵇ ⁱˢ ᵘⁿᶜᵒᵒʳᵈⁱⁿᵃᵗᵉᵈ; ᵗʰᵉ ᶠⁱʳˢᵗ ʷᵃᵛᵉ ᵐᵘˢᵗ ʷⁱˡˡⁱⁿᵍˡʸ ˢᵃᶜʳⁱᶠⁱᶜᵉ ᵗʰᵉᵐˢᵉˡᵛᵉˢ ᵗᵒ ⁱᵐᵐᵒᵇⁱˡⁱᶻᵉ ˡⁱᵐᵇˢ ʷʰⁱˡᵉ ˡᵃᵗᵉʳ ʷᵃᵛᵉˢ ᵖⁱˡᵉ ᵒⁿ. --- #### ³. ᴾˢʸᶜʰᵒˡᵒᵍⁱᶜᵃˡ ᶠᵃᶜᵗᵒʳˢ • ᴴᵘᵐᵃⁿˢ ʰᵃᵛᵉ ᶠᵉᵃʳ; ᵃ ᵍᵒʳⁱˡˡᵃ ˢʰᵒʷˢ ⁱⁿᵗⁱᵐⁱᵈᵃᵗⁱᵒⁿ ᵈⁱˢᵖˡᵃʸˢ (ᶜʰᵉˢᵗ-ᵇᵉᵃᵗⁱⁿᵍ, ᵇˡᵘᶠᶠ ᶜʰᵃʳᵍᵉˢ). • ᴱᵛᵉⁿ ᵗʳᵃⁱⁿᵉᵈ ᵐᵃʳᵗⁱᵃˡ ᵃʳᵗⁱˢᵗˢ ˢᵗʳᵘᵍᵍˡᵉ ᵗᵒ ˢᵗᵃʸ ᶜᵃˡᵐ ʷʰᵉⁿ ᵃ ¹⁸⁰ ᵏᵍ ᵖʳⁱᵐᵃᵗᵉ ʳᵘˢʰᵉˢ ᵗʰᵉᵐ ᵃᵗ ³⁰ ᵏᵐ/ʰ. • ᴹᵒʳᵃˡᵉ ᶜᵒˡˡᵃᵖˢᵉˢ qᵘⁱᶜᵏˡʸ ᵒⁿᶜᵉ ᵖᵉᵒᵖˡᵉ ʷⁱᵗⁿᵉˢˢ ˡⁱᵐᵇˢ ᵗᵒʳⁿ ᵒʳ ˢᵏᵘˡˡˢ ᶜʳᵃᶜᵏᵉᵈ. --- #### ⁴. ᴾˡᵃᵘˢⁱᵇˡᵉ ᶜᵒᵐᵇᵃᵗ ᵀⁱᵐᵉˡⁱⁿᵉ ¹. ᴵⁿⁱᵗⁱᵃˡ ᴿᵘˢʰ (⁰–¹⁰ ˢ) ᵀʰᵉ ᵍᵒʳⁱˡˡᵃ ᶜʰᵃʳᵍᵉˢ, ˢʷᵃᵗˢ, ᵒʳ ᵇⁱᵗᵉˢ ᵗʰᵉ ᶜˡᵒˢᵉˢᵗ ᵐᵉⁿ. ᴱˣᵖᵉᶜᵗ ⁵–¹⁰ ˢᵉʳⁱᵒᵘˢ ᶜᵃˢᵘᵃˡᵗⁱᵉˢ ᵃˡᵐᵒˢᵗ ⁱⁿˢᵗᵃⁿᵗˡʸ. ². ᴰᵒᵍ-ᴾⁱˡᵉ ᴾʰᵃˢᵉ (¹⁰–³⁰ ˢ) • ᴵᶠ ᵗʰᵉ ᶠʳᵒⁿᵗ ˡⁱⁿᵉ ʳᵉˢⁱˢᵗˢ ᶠᵉᵃʳ ᵃⁿᵈ ᵐᵘˡᵗⁱᵖˡᵉ ᵐᵉⁿ ᵍʳᵃᵇ ᵉᵃᶜʰ ˡⁱᵐᵇ, ʷᵉⁱᵍʰᵗ ˢᵗᵃʳᵗˢ ᵗᵒ ᵐᵃᵗᵗᵉʳ. • ᵀʰᵉ ᵍᵒʳⁱˡˡᵃ ˢᵗⁱˡˡ ᵈʳᵃᵍˢ ᵇᵒᵈⁱᵉˢ, ᶠˡⁱⁿᵍˢ ˢᵒᵐᵉ ᵒᶠᶠ, ᵃⁿᵈ ᵇⁱᵗᵉˢ. ᶜᵃˢᵘᵃˡᵗʸ ᶜᵒᵘⁿᵗ ᶜˡⁱᵐᵇˢ ᵗᵒʷᵃʳᵈ ²⁰–³⁰. ³. ˢᵘᶠᶠᵒᶜᵃᵗⁱᵒⁿ ᵒʳ ᴱˣʰᵃᵘˢᵗⁱᵒⁿ (³⁰ ˢ–² ᵐⁱⁿ) • ᴼⁿᶜᵉ ⁶⁰⁰–⁸⁰⁰ ᵏᵍ ᵒᶠ ʰᵘᵐᵃⁿˢ ᵃʳᵉ ᵖʰʸˢⁱᶜᵃˡˡʸ ˡʸⁱⁿᵍ ᵒⁿ ᵗᵒᵖ, ʳᵉˢᵖⁱʳᵃᵗⁱᵒⁿ ᵇᵉᶜᵒᵐᵉˢ ᵈⁱᶠᶠⁱᶜᵘˡᵗ ᶠᵒʳ ᵗʰᵉ ᵍᵒʳⁱˡˡᵃ. • ᴴᵒʷᵉᵛᵉʳ, ᵗʰᵃᵗ ᵐᵃˢˢ ᵃˡˢᵒ ᶜʳᵘˢʰᵉˢ ᵖᵉᵒᵖˡᵉ ᵘⁿᵈᵉʳⁿᵉᵃᵗʰ ᵃⁿᵈ ᵗʰᵉ ᵍᵒʳⁱˡˡᵃ’ˢ ᵗʰʳᵃˢʰⁱⁿᵍ ᶠʳᵃᶜᵗᵘʳᵉˢ ʳⁱᵇˢ, ᵃʳᵐˢ, ᵃⁿᵈ ⁿᵉᶜᵏˢ. --- #### ⁵. ᴸⁱᵏᵉˡʸ ᴼᵘᵗᶜᵒᵐᵉˢ ¹. ᵁⁿᵗʳᵃⁱⁿᵉᵈ ᶜʳᵒʷᵈ (ᵐᵒˢᵗ ʳᵉᵃˡⁱˢᵗⁱᶜ) • ᴾᵃⁿⁱᶜ ˢᶜᵃᵗᵗᵉʳˢ ᵗʰᵉ ᵍʳᵒᵘᵖ; ᵗʰᵉ ᵍᵒʳⁱˡˡᵃ ᵈᵒᵉˢⁿ’ᵗ ᵉᵛᵉⁿ ⁿᵉᵉᵈ ᵗᵒ ᵏⁱˡˡ ᵉᵛᵉʳʸᵒⁿᵉ—ʲᵘˢᵗ ᵐᵃⁱᵐˢ ²⁰–⁴⁰, ᵗʰᵉ ʳᵉˢᵗ ᶠˡᵉᵉ. • ᴳᵒʳⁱˡˡᵃ “ʷⁱⁿˢ.” ². ᴴⁱᵍʰˡʸ ᶜᵒᵒʳᵈⁱⁿᵃᵗᵉᵈ, ˢᵃᶜʳⁱᶠⁱᶜⁱᵃˡ ᵀᵉᵃᵐ • ᵀʰⁱⁿᵏ ʳᵘᵍᵇʸ ᶠᵒʳʷᵃʳᵈˢ ʷⁱᵗʰ ᴺᵃᵛʸ-ˢᴱᴬᴸ ᵈⁱˢᶜⁱᵖˡⁱⁿᵉ. • ᶠⁱʳˢᵗ ²⁰ ᵐᵉⁿ ᶠᵒʳᵐ ᵃ ˢᵃᶜʳⁱᶠⁱᶜⁱᵃˡ ʷᵃˡˡ, ⁿᵉˣᵗ ²⁰ ˡᵃᵗᶜʰ ᵒⁿᵗᵒ ᵃʳᵐˢ/ˡᵉᵍˢ, ʳᵉᵐᵃⁱⁿⁱⁿᵍ ⁶⁰ ᵖⁱˡᵉ ᵒⁿ ᵘⁿᵗⁱˡ ᵗʰᵉ ᵍᵒʳⁱˡˡᵃ ˢᵘᶠᶠᵒᶜᵃᵗᵉˢ. • ᶜᵃˢᵘᵃˡᵗⁱᵉˢ: ³⁰–⁵⁰ ᵈᵉᵃᵈ/ⁱⁿᶜᵃᵖᵃᶜⁱᵗᵃᵗᵉᵈ. • ᴴᵘᵐᵃⁿˢ “ʷⁱⁿ,” ᵇᵘᵗ ᵃᵗ ᵃⁿ ᵉⁿᵒʳᵐᵒᵘˢ ᶜᵒˢᵗ. ³. ᵂⁱᵗʰ ᴹⁱⁿⁱᵐᵃˡ ᴵᵐᵖʳᵒᵛⁱˢᵉᵈ ᵂᵉᵃᵖᵒⁿˢ (ᵇᵉˡᵗˢ, ʲᵃᶜᵏᵉᵗˢ, ˢᵗⁱᶜᵏˢ) • ᶜʰᵒᵏⁱⁿᵍ ᵗᵒᵒˡˢ ᵃⁿᵈ ʳᵉᵃᶜʰ ᵉˣᵗᵉⁿᵈᵉʳˢ ᵇᵒᵒˢᵗ ʰᵘᵐᵃⁿ ᵒᵈᵈˢ ˢˡⁱᵍʰᵗˡʸ—ˢᵗⁱˡˡ ¹⁵–²⁵ ᵈᵉᵃᵗʰˢ ᵉˣᵖᵉᶜᵗᵉᵈ. • ᴴᵘᵐᵃⁿˢ ᵉᵈᵍᵉ ᵃ ᵛⁱᶜᵗᵒʳʸ ⁱᶠ—ᵃⁿᵈ ᵒⁿˡʸ ⁱᶠ—ᶜᵒᵒʳᵈⁱⁿᵃᵗⁱᵒⁿ ʰᵒˡᵈˢ. --- #### ⁶. ᴱᵗʰⁱᶜᵃˡ & ᴾʳᵃᶜᵗⁱᶜᵃˡ ᴺᵒᵗᵉ ᴵⁿ ᵗʰᵉ ʳᵉᵃˡ ʷᵒʳˡᵈ, ˢᵘᶜʰ ᵃ ᵐᵃᵗᶜʰᵘᵖ ʷᵒᵘˡᵈ ᵇᵉ ᵘⁿᶜᵒⁿˢᶜⁱᵒⁿᵃᵇˡʸ ᶜʳᵘᵉˡ ᵗᵒ ᵇᵒᵗʰ ʰᵘᵐᵃⁿˢ ᵃⁿᵈ ᵗʰᵉ ᵉⁿᵈᵃⁿᵍᵉʳᵉᵈ ᵍᵒʳⁱˡˡᵃ ˢᵖᵉᶜⁱᵉˢ. ᶜᵒⁿˢᵉʳᵛᵃᵗⁱᵒⁿⁱˢᵗˢ ᵉˢᵗⁱᵐᵃᵗᵉ ᵒⁿˡʸ ᵃ ᶠᵉʷ ᵗʰᵒᵘˢᵃⁿᵈ ᵐᵒᵘⁿᵗᵃⁱⁿ ᵍᵒʳⁱˡˡᵃˢ ʳᵉᵐᵃⁱⁿ; ᵃⁿʸ ˢᶜᵉⁿᵃʳⁱᵒ ᵗʰᵃᵗ ʰᵃʳᵐˢ ᵒⁿᵉ ⁱˢ ᶜᵒᵘⁿᵗᵉʳᵖʳᵒᵈᵘᶜᵗⁱᵛᵉ ᵗᵒ ʷⁱˡᵈˡⁱᶠᵉ ᵖʳᵒᵗᵉᶜᵗⁱᵒⁿ. --- #### ᴮᵒᵗᵗᵒᵐ ᴸⁱⁿᵉ ᴬ ˢⁱˡᵛᵉʳᵇᵃᶜᵏ ᵍᵒʳⁱˡˡᵃ ᶜᵃⁿ ᵃᵇˢᵒˡᵘᵗᵉˡʸ ᵈᵉˢᵗʳᵒʸ ˢᵐᵃˡˡ ᵍʳᵒᵘᵖˢ ᵒᶠ ᵘⁿᵃʳᵐᵉᵈ ᵐᵉⁿ. ᴬᵍᵃⁱⁿˢᵗ ¹⁰⁰, ʳᵃʷ ⁿᵘᵐᵇᵉʳˢ ᵉᵛᵉⁿᵗᵘᵃˡˡʸ ᵗᵉˡˡ—ᵇᵘᵗ ᵒⁿˡʸ ⁱᶠ ᵗʰᵒˢᵉ ʰᵘᵐᵃⁿˢ ᵃᶜᶜᵉᵖᵗ ʰᵒʳʳⁱᶠⁱᶜ ˡᵒˢˢᵉˢ ᵃⁿᵈ ʳᵉᵐᵃⁱⁿ ᵉˣᵗʳᵃᵒʳᵈⁱⁿᵃʳⁱˡʸ ᵈⁱˢᶜⁱᵖˡⁱⁿᵉᵈ. ᴵⁿ ᵐᵒˢᵗ ᵇᵉˡⁱᵉᵛᵃᵇˡᵉ ˢᶜᵉⁿᵃʳⁱᵒˢ, ᵖᵃⁿⁱᶜ ᵐᵃᵏᵉˢ ᵗʰᵉ ¹⁰⁰-ᵗᵒ-¹ ᵃᵈᵛᵃⁿᵗᵃᵍᵉ ᵉᵛᵃᵖᵒʳᵃᵗᵉ, ᵃⁿᵈ ᵗʰᵉ ᵍᵒʳⁱˡˡᵃ ʷᵃˡᵏˢ (ᵒʳ ˡⁱᵐᵖˢ) ᵃʷᵃʸ ᵛⁱᶜᵗᵒʳⁱᵒᵘˢ.
This worked. Lol.
i need to publish in times new roman LOL
lollollollol
sup yall
I use only the reasoning ones.
I like 4.1 and sonnet thinking for stuff I think it would be better for
so you'd recommend those ones for academic ones
Absolutely
noted
@storm kindle is this new?
You can change in what order guns unlock in bp.
Is that battlepass reward order
Yeah

Well ig it wouldn't make much of a difference when all the rewards are shit
Which they have been lately
You'd still get the refund if you buy it at t1
The timing doesn't affect that
But ig getting it at t50 is assuring since you'd have already unlocked all the cp tiers
Yeah. I don't trust myself or my life with consistency.
@storm kindle
I did a quadra in a round.
The map is shit though.
Is that the purplish scifi map
Snd is good with team
The people who invited me left the lobby.
You use vc?
God no
I never use vc.
Snd in shipment would be wild.
I throw gas in one corner and thermite/molotov in the center.
Maybe I should play br more.
Mostly it'd just go to waste so I just stuck to my primary guns
Bad idea, it's all bots nowadays
Unless you plan to hit legendary rank
@storm kindle pubg would have been good if they had scope reflection like codm.
That's when you start sharing the lobby with real players
Snipers are too op.
Never played the game
Is br that shit?
It is now
In mp I start getting actual players at master/grandmaster.
Hm
Maybe if you team up with high ranking players you'll get enemies who aren't bots
That's a little more exciting than killing dummies
That's one of the reasons why I left codm
Depends on time of the season too.
What do you play now?
It only takes a day for the lb grinders to get the high ranks of br
Nthn
Wordle, if that counts
One game a day, by newyork times
How I can expedite memory feature beta testing on my pro account please?
Hello everyone, I am wondering if the reply I got from Perplexity regarding the deep research function is accurate? I see that perplexity ranks itself the lowest among other competitors.
Is this accurate? I did not have the chance to test the ARI, but I have tested the Grok3 Deeper Research and I feel indeed slightly better than Perplexity research.
Did any of you had the chance to test the ARI and let me know if this claim is accurate?
Thanks to all the community
Replace ari with chatgpt and it'll become accurate.
Ari yaps a lot but it's reasoning is dull.
I submitted a simple question several times today and knew the correct answer. Over and over I got the wrong answer from Perplexity Pro. What's up with this? How can I have any confidence in a system that can't get the simple things correct?
What was the question?
what zodiac sign is the moon currently located
after getting the wrong answer, I asked for advice about how to ask the question to ensure the correct answer and validate it in the future, but the sequence just went in circles
??
Huge…
This means they can use perplexity instead. Maybe the US will go that way.
Ban
Hope everyone follows after
Wrong server, apologies 😭
I can't believe how bad perplexity is.
Used it with 2.5pro and can't even give me a simple answer so disappointed
stop spreading fake screenshots
I know I was complaining a bit about using PPLX for codiing, but TBH, just wrong tool for the wrong job. VS Studio + CoPilot has everything built in for coding
What did he send?
Typical crypto scam nonsense lol
elon musk giving free bitcoin
Where? I want to
Its not real silly
is perplexity down for anyone else or is my wifi trash 
Your wifi trash
working for me, ur wifi is trash
You're welcome pretty
If you need more help you can write to me privately

atm its fair to put ppxl research at the bottom
loool
Never waste an opportunity, nephew
It doesn't matter if you're trans, if you're a man, if you're a woman
There's always a gap

😐
with the Web Search on?
The system prompt has a big role to play, turning off web search would allow you to use pure Gemini that has a minimal system prompt which is negligible
Yes, the answer I wanted required web search
It's working great for me so far, not sure why it's unable to answer your question
My use case is mostly programming related 
Promt?
yo guys im building an app and i got an idea
its like a personal stylist and you take selfies give send it to an AI vision model that processes it ect gives feedback. you can give it your taste your wardrobe ect and all and ye
do u think this will even take off?
i need some uhm what do u call it
feedback
There will be at least 100 of these around, it will be a first to market, first to adopt kind of thing
i havent seen much like it
I have
after searching a bit on perplexity
oh damn
Impractical jokers season 11 release date
in one of my business networking groups a guy was working on this app
oh so not released
doesn't mean don't do it
yet
I don't know, don't keep tabs on the guy
ah
Accidentally wrote here 12 I meant 11
because i havent seen a mobile app like this
how would one monetize this anyways
have subscription plans?
And it just told me that season 11 has not been released yet which is just wrong
og impractical jokers was the best
and there is only 11 seasons btw
what do you mean? It would actually be the easiest to monetize
subscription plans or paid app
this is my first time making an app instead of website
so i have no idea
free, and affiliate commissions
how would affiliate commissions work
ask perplexity
ojhhhHH
Yeah I know I asked it the release date of the 11th season and it told me it did not come out yet
you mean like that
ohh
aight ty
I agree the show became worse and worse since joe left
I remember I saw something like this before
just like top gear
after jeremy clarkson and co left
Gear
Gear 5
Is that a one piece reference
IS IT
moe
anyone else vibe coding with Rust?
The Perplexity research model is basically built on an ensemble Model of DeepSeek R1 (Ensemble Reasoning - At least I believe so if not wrong! ). I know it makes it much faster and affordable, but I wonder
why Perplexity pro users, Could not expect at least a higher and more accurate results than the Deep Research of the ChatGPT?
I am honest here, I am not a big fan of OpenAI ( OK . OK. only the O3 Model, I admit that I am impressed, but O4-mini makes hallucinations.) I wish Perplexity gave the option to its users to use O3 for reasoning instead of O4-Mini.
BTW. did you try the ARI for a real case scenario?
I think Perplexity needs to trailblaze a little here. They need to go all in on agentic frameworks
I just ditched Perplexity for vscode+git-copilot and damn, life is good
at minimum, for coding they need to make a perplexity extension for some other application like notepad++ or even a native one for vs code
Is that even generated by PPLX?
gpt
of course
Bro
I can officially say o3 is laywer approved
My public defender saw me like a monkey when I rejected the plea deal
Then when I gave all the reasoning from o3, then a process went where the public defender gave me the investigator to meet up because she was convinced. It payed off because I won the case and have the opportunity to sue for APS taking a case and their report was garbage. APS can’t be sued because they’re monkeys that can do reports without solid evidence on the spot but you can at least sue the person who was directly responsible for this
I won the case
They can smd
2 felonies prevented
On my mama
Celebrating rn
no shit
Do you have access to ari?
you got it?
looks like chrome with baked in perplexity ?
why no one is showing how it works
not a single video on Yt
Is it just me or is perplexity being slow?
It is being slow
Perplexity is fake AI
It's the worst AI tool i tested yet
When you choose research it's take about half an hour then gives you very short details without any charts or tables
If you try same search in grok it's faster more details
I don't know why Perplexity is like this
Maybe cuz theyre trying to cut costs by limiting the max output tokens & context length
May be
But it's give bad quality responses
Lmao? Half an hour? Can't relate.
Damn
Hi guys, I have a question. Perplexity wrote the plugin to me in Minecraft, in some kind of chat, but how to find out in which chat I created the plugin with him, I have a lot of these chats. Is there any kind of search?
hi
I had a chat in space 😅
You could do the search
But perplexity is bad at coding
You better switch to Gemini ( free ) or any other llm to get better and bigger code
I understand. 😔
You can still search it if you remember the words inside it. Doesn't matter whether or not the thread was in a space.
thanks bro helped out.
whats happening to perplexity sir? it is really searching bad for academic references sir?
Like what is happening sir? anyone noticing this sir?
Even basic statements to find for citations it cannot give anymore sir? try doing it sir. I think they nerfed something sir.
deep research mode? I used it today and it wwas fine
no sir, the normal search sir, like before when i use Gemini 2.5 pro, sir, it accurately find me a supporting reference for a statement sir, but since last night sir, it cant until today sir.
You want me to laser you homelander sir?
im so confused
maybe try a differnt model
idk i never use gemini, if i do i go to the gemini website cuz they have a better interface lol. i prefer using grok and gpt for daily use
yeah sir, i am trying deep search now sir, this never happened before sir, maybe their cost cutting scheme again sir.
🤨
I don’t want to be all gassy
Yeah. The doc will take care of that while the ambulance vrooms.
My ahh asking chatgpt how to make eye drops last longer than a month after opening.
Bro is being risky.
💀
Ngl, the mods are pretty chill in this server.
R1
You can change the model in perplexity
I use Claude 3.7 for coding
coding through perplexity?
interesting
https://perplexity.supply/products/perplexity-subscription @heavy horizon is back online
Give the gift of knowledge with a 1, 3, 6 or 12 month Perplexity Pro subscription. Perplexity Pro users can search 3x as many sources, get access to the latest AI models, search through their own files, and more. It's the perfect gift for a curious friend or loved one. Recipient will receive the gift subscription via p
Tf did you do?
Is the music from Sherlock?
Now why are you emo?
He didn't get mother's milk.
worst decision ever 😩
#tlou #thelastofusseries #thelastofusseason2 #tlou2 #thelastofuspart2 #lastofus #hbo #elliewilliams #caileespaeny #bellaramsey #gamingreels #games
I would never believe
Too good to be true
"When they try every possible Ellie casting except the right one" 🥲
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
#ellie #elliememes #thelastofus #thelastofushbo
#thelastofusmemes #tloucontent #bellaramsey
Your search has been upgraded to pro… STOP DOING THAT PERP!!
I only get 2 every 24 hours

Don't worry you're not missing anything
Pro users barely get pro searches

Use grok free if you are poor.
Ppxl dr with 17 sources...
@distant aurora apparently copilot already rolled out dr
But it's only visible to subs atm...
I see.
Weird is that I don't see videos and reviews yet
If it's the same as full or light deep research from OpenAI but without limits or high limits then it would've hit the news by now
There's so little info out about it
Does that girl have Down's Syndrome
@green crystal perplexity tripping fr
And no, I didn't use the 300 per day in case anyone's about to say that.
toast
Its a bug dear
Same shows with file attachment
Sometimes
Your chicken her rice. Want my sauce?
Lol.
For people who aren't getting context.
Hello guys
Free Gemini dr + free grok dr + chatgpt plus seems like it's gonna do it for me
I'm unsubbing lol
Giving another $20 to ppxl feels like a waste atm
Does mobile app support image generation?
yea it's already happening that these $20 a month subscriptions are getting blown out by free models
i just cancelled my $20 a month cursor subscription as well after getting constantly rate limited
Google just went ahead and did it 😂
Subs are fine if you get the value out of it and if the service is on par or better than others
Ppxl is getting further and further away from the others
https://youtu.be/6DmXlJwaP0I?si=2o8yD6gSwF9BsPJG&t=609
can someone tell me the context of this scene
steak steak steak steak
I don't own the rights to any of these clips. All rights belong to Fox.
Hello, everyone, I'm finding a good friend.
I like making friends.
Please contact me to be a good friend.
Please contact me.
Because people aren't normally so obvious lol
I have tried. It's not worth it.
So it’s pretty much broken after last update anybody know when this will be fixed
If you ask a audio question it completely freaks out and repeats itself over and over
/create picture animation of a merman prince {ts:90} flying through a magical sky. He has a muscular upper body with shimmering scales covering his lower fish tail. His hair flows majestically in the wind as he soars through fluffy clouds. The merman wears a golden crown and royal accessories. The scene is set during sunset with vibrant orange and purple hues in the background. Cinematic lighting, ultra-detailed, 4K resolution, muted colors to maintain consistent contrast.
Pretty
yup
i still think openai is better though in some ways
gpt make mostly accurate images but with less visuals
imagen is less accurate than gpt but visuals are better than gpt
its trying to make him look like a fishman
like in cartoons
oh yea i see that
yea imagen needs to learn the difference between a fishman and merman
we can expect that from imagen 4
supposedly tomorrow? or am i being lied to
imagen 3 just generated a topless mermaid just from the prompt "mermaid" google i want a refund
🥲 not even kidding
from where are you using imagen
The home for AI experiments at Google
lets see
F61NTF9J
Invalid
This one is valid
Thanks
in pplx ?
Things I do in my free time:
1- Save people from fires
2- Research to fight cancer
3- Search for a cure for menstrual pain
4- Cook for my mom
5- Give clothes and money to the poor
6- Read the Bible
Hm: Talk about how strong and brave women are
Idk why its bein like that though
Women? Brave? Debatable.
Is a bait
I remember.
There's some award that Peter and his friends win together.
Some trophy.
But it goes missing.
So they check which one of the friends stole it.
Did you know that cancer has a natural cure?, even if the criminal pharmaceutical industry says otherwise.
Sea water is healing against cancer, sunbathing for 30 is healing, grounding is healing, and so on.
Hmmm, okay?
The thing is, the criminal pharmaceutical industry wants to keep people sick.
Ok bro
Nova?
You changed your name
Well, now you know. Do your research deeply. Good afternoon.
La elite satánica que gobierna ha creado un negocio con la salud de las personas. No es rentable curar sino que es rentable enfermar, porque una persona enferma supone un ingreso a la industria de la salud por décadas y décadas.
Biblical Health ☠️
Brother, if you see that I don't care, just leave it
is pplx down again?
No
@green crystal new record
LOL
Man so much nerfing and cost cutting everywhere
Maybe they should change the sub price from $20 to 10 or 5
Yeah I started switching over to search with o4 mini on chatgpt
It's amazing even for free tier
I didn't include Gemini cuz it isn't as good as them 2 at complex reasoning.
Does o4 use more sources with chatgpt plus?
They have the high variant.
Normal o4 mini is worse than O3 mini high in problem solving.
So eh more reasoning and searches then
Honestly, I don't feel much difference between O3 mini high and o4 mini high in practical use.
Latter just feels faster.
We saw the new variant of deep research sent by someone some time ago.
They told us they're working on deep search but no known deadline yet.
According to the AMA it's "very soon"

Almost 2 weeks ago
Well you can't even really use the current deep research while waiting for it because it barely uses any sources and reasoning..
I wonder how apple thought about working with perplexity for search
Sonar? I'm sure apple could figure something out that's on par or even beat it
Apple won't do it for cheap.
Perplexity probably will.
Guess maybe that's why there's so much talk about speed and latency
And barely any about model capabilities
It's for when someone using an iphone or something just wants to quickly lookup something using the default search bar widget
Look at the examples in moto
Wait
Reviewing the Motorola Edge 60 Pro, a flagship smartphone for £600 UK, boasting a triple lens camera, excellent battery life and some decent Moto AI features. It's up against the Poco F7 Ultra and other more budget-friendly flagship phones, but the Motorola Edge 60 Pro is still worth considering. Here's my verdict after using it as my phone for...
No I don’t
It's good with loooong texts but the scraping/reasoning isn't as good as they say.
5 Ways to Know That I Am Coming:
- The sky turns black
- Women start running and looking at me with disgust
- The plants begin to die
- Mirrors start breaking
- The sun stops shining
Girl what
Notebooklm releases on Android today
why do we only have grok 3 beta and not the full version?
Someone will interest you.
i got the 2024 moto edge last month lol
maybe shoulda waited
You need polycarbonate mirrors.
What's your budget?
it was 300-350 ish USD but im not buying anything anymore lol
ill keep this for another 2 years it does everything i need
iOS too!
best model for studying?
Ayyee I only sold the announcement for Android
Check the app above
I like sonnet reasoning for harder stuff but 4.1 can manage most
Honestly, in my experience I like using Gemini because it doesn't hallucinate as much
alright
4.1 is fast but Gemini seems to be more verbatim
ive only used 4.1 and claude 3.7 rlly
ahh ok
I myself like 4.1's answers and behavior more than other models, but it really depends on you
What I would do is make a space with all your information for the topic and figure out what model represents your notes the best
ok
But I still recommend giving notebook LM a try
Especially because it can make a podcast on your notes and make it easier to understand because you can listen to it
This isn't trying to be smart-assness but what you can do is use sonnet reasoning with web turned off and ask it to make a prompt to better study using perplexity spaces. That's what I did for writing emails
alright ty
oh yeah this is actually really good
so far perplexity has rlly helped with studying
@hushed lantern awesome I'm glad to hear it. Make sure if you wanna focus on your notes that you turn off web so not only are the responses quicker but are more likely to reflect your notes in citation
What happen in android app?
Web worked well and showing answers, which not shown in the app. Error occured and then thinking starts again, and error again
Try killing the app and reloading
It could also be a VPN issue
Pplx required vpn?
The new update is too good
Thanks, yes, most likely vpn
If you want more consistency on mobile, I recommend using Firefox with the complexity add-on. There is a plugin that refreshes cloudflare automatically which is the issue you are having
What about Chrome?
Even without vpn this error happen again : 😭
Whats this
Chrome on mobile doesn't have extension support. That's why I suggested Firefox
how did you guys add the gradients to the role?
I have no idea.
Didn't work as expected. Lol.
so there is nothing new?
regarding this?
how will I enjoy the Pro features
but as a consumer I have the right to express my opinion
about the implementation of my language.
Okay.
dyou guys actually use deep research?
when I do it blunders a bit, mixing names and small details, I think its because they use their sonar model idk
I just dont see the edge deep research has, when I prompt gemini 2.5 pro extensively, I can get it to use 200 links
They have been making adjustments to DR in PPLX.
Most users report experiences just like yours
They have accouned a new DR version coming soon. But they have also acknoledged it is behind schedule
yh pplx gotta watch out cause gpts DR, man that thing is solid
There is no more information about that at the moment
thanks
It has greatly improved this past few weeks
man we dont even need agi or better agents for how all this tech is already gonna flip society upside down imo
I had a hard question that needed a short answer. For the first time I got a great answer on GPT DR a few sentences long. I was greatly impressed!!!
Only 10 sources???????
just skip DR bro, gem 2.5 competes with gpt DR id even say
How many dank memes per day can i make with the gpt image generator?
Infinite if you have plus and keep it down to 100 or so per 15 minutes
What's missing? Arms or legs?
I notice the other "reasoning" models on perplexity are horrible to use in comparison.
-always cutting short on explanations
-less insightful than gemini
o4 mini pisses me off so much
I have perplexity pro. How do I know if its using the best openai image model
Ahhh I thought you meant on ChatGPT
Hmmm I'm not sure about PPLX
Go to settings and choose GPT Image 1
Seem like you already have it looking at your image
What are my image gen limits with gpt1
dyou try gem 2.5 too, if so what r your impressions
It's good. But reasoning isn't as strong as grok 3 thinking imo.
interesting
It's weird actually. Cuz like Gemini answers a complex question correctly in a chat.
yeah, love the elaborations
Then I ask it the same question a few more times in new chats and it never gives that correct answer again.
oh
Feels like intentionally induced error.
maybe the temperature on gem 2.5 pro is higher, meaning its less deterministic and more creative, hence sometimes very smart
kinda cheating
Also, Gemini is bad at remembering instructions.
Is this gpt 1
hm, honestly I doubt that, ill need to try grok first, cause I basically use gemini to make scripts for me, and edit them.
its a new project too
Unlimited
Is that bush?
J powell
Yeah
actually, check this, maybe all you need to do is elaborate more.
when I use gemini for making me a script or do edits, I literally upload two text files, that recap things completely.
so when you add context it shines
Money
I ain't that dumb. I put custom instructions in ai studio.
Seems elaborate enough to both me and Gemini:
Custom Instructions for Generating Responses:
Direct Opening: Start the response immediately with the main subject or answer to the user's query. Do not use any introductory phrases that announce the response, such as "Here's a note about...", "Based on the text...", "Okay, let's look at...", or similar conversational openings.
Natural Structure: Present the information with a natural flow. If the request implies distinct parts (like a definition, a list of points, and a concluding thought), structure the response logically with clear paragraphs or bullet points as appropriate, but without explicit section titles unless requested.
Avoid Explicit Summary Markers: Do not begin concluding paragraphs or summary statements with phrases like "In essence," "Essentially," "Overall," "To summarize," "Ultimately," or any similar words or phrases that solely serve to explicitly mark the start of a summary or conclusion. Allow the final thoughts to flow naturally from the preceding points.
Human-like Phrasing: Use natural-sounding language throughout the response. Avoid overly formal, robotic, or formulaic phrasing where possible.
No Source Mention: Absolutely do not refer to or mention the source document used to generate the information.
Clean Closing: The response should end after the requested information has been provided. Do not add conversational closing remarks, questions, or feedback requests within the response content.
Check that every rule was followed before giving the final answer.
oh ai studio, maybe how it handles context data is very diferent.
within pplx gem is super reliable I find, because pplx maybe indexes messages, so not everything gets fed to gemini.
and I think pplx uses multiple gemini requests within one prompt.
so the experience is surely much different to ai studio
Pplx?
on perplexity, I only use gemini inside it
even for code, cause it needs to rely on api references a lot
ah yes actual deep research
Erection?
Goon
bro
Finally
I’m off from work
@distant aurora bro finally chatgpt will be used to the max potential for me
I’m finally starting cybersec so it’s gonna make studying even better now
Perplexity is gonna be useful asf for me now
Also they added codex to mobile @distant aurora
So far it’s impressive asf ngl codex is definitely above o3 fs, it ranks above deep research specialized o3 in coding fs
The new project sam altman teased is gonna be something to look forward to tbh
But dang I gotta catch up my ai knowledge is dying little by little from my lack of use since work and college
For sure
Above o3
Fs means for sure lol
Ngl It’s been extremely helpful when I have gpt 4.5 on one tab and o3 on the other tab
Both are op asf especially when o3 is on projects and full limits unlocked
why does perplexity add a citation but then when I press check sources it shows no sources
What's with all this "it seems I'm having technical issues" live chat messages? Does anyone know? I keep getting this for the most basic questions and internet searches.
Yes
I always got it
This is your best chance because I won’t be available for a week again
In this video, I'll be telling you about how Microsoft has opensourced Github Copilot and is now fully committed to making VS Code the best AI editor that is Opensource.
Rare footage of you. https://youtube.com/shorts/213pJGKLtM4
What happens when an intruder breaks into a Hollywood stuntman's house... but his dog only listens to the owner? 😂 Watch this hilarious yet intense moment from Once Upon a Time in Hollywood where Cliff Booth's (Brad Pitt) dog becomes the ultimate guard pup—attacking on command! � Who needs a security system when you've got a dog this loya...
Which android skin has best battery life per unit capacity. Refer to actual tests. I want to see actual tests in your response. I want you to compare multiple smartphones across various price and hardware brackets across all android skins. Take atleast 10 phones of each brand for comparison, prefer the newer ones. Add motorola and asus to this comparison too. And don't miss realme oppo vivo nothing os
Anyone knows if perplexity plans on integrating the perplexity chat with the perplexity assistant?
Like for example I asked it when is the Google io conference and it tells me the answer then I can tell it to set a reminder for 15 minutes before the conference and then it does that like it does in chatgpt for example
Moto has this I think. Not sure if it's available on all phones though.
Hopefully it gets released on Samsung
bro @distant aurora share me the group link
@distant aurora
This data is not of much use if it shows numbers from 1996
Wdym?
Oh
idk how often it updates, or why, but its no bueno.
been weeks
Is it that way on the website too or just the app?
@storm kindle 3 shot in all ranges is diabolical.
Never saw the save the chart feature before.
nothing on web
hah they're specializing into finance now

Hey guys. Would anyone of yall be interested in answering my AI enquete for my school project?
https://docs.google.com/forms/d/e/1FAIpQLSf8Hyre6xBT5U6GFyF7xRm9LGWD533w841AWRzgwspMGyliTQ/viewform?usp=dialog
hii can yall plsplspls like my video and share its a competition for my school https://youtu.be/SFEo64dAMDA?si=Yh6TeTijSQivygN4
done!
done
@raw harness @placid warren I’d suggest sharing these in #😎│cool-projects instead
Would probably find more people interested in them as well
u should enable multi selection for choosing ai model
But I did answer it as well 🫡
Are they dead
half
thanks for the advice!
alr thx! i will work onit
I’ve noticed the Space option to ‘Save as a file to Space’ has changed to ‘Share thread with Space’. Is it just a cosmetic change in wording, or does it now work differently? Checked the Changelogs but can’t find any mention of it?
I had no idea that I could export an entire thread as a pdf instead of one response. Great.
No sorry - it’s an option on the threads within a Space
Wait
What does save as a file to space do?
Uploads the thread as if it were a file in the Space, so that further threads (not answers within threads, but entire new threads within a space) can use the context from the saved threads
I don't remember seeing that option there. Maybe I didn't notice.
I thought just sharing the thread with the space shares the context too.
Yeah been there for a little while (maybe a couple/few months?) but it’s changed now anyway to ‘Share thread with Space’
Yeah. It likely does the same thing.
So you can add a thread you created outside of a Space to a Space, but then the threads that are in that Space can be saved as files so the new threads can use them for context. Honestly though, use it with caution
This is important: any threads you save as files to the Space are seen by the AI as being authoritative (in my experience). If you received an inaccurate answer to one of your questions in a thread and you then save that entire thread as a file in the Space, any new threads will see that answer and repeat it and even elaborate on it and strive to justify it and generally just ‘make stuff up’ to explain why it’s the case, even though it’s wrong.
And that’s even when the thread is allowed to use web searches too. Seems to give the information inside saved files in the Space more weight than answers found online (including threads you’ve saved as a file in the Space).
When you start a new thread in the Space, you can add “Don’t use previous threads as sources” (or words to that effect), but I tend to find that it just ends up not using ANY of the files in the Space when you do.
This hasn’t changed.(At least on my end)
In your Space:
1- Either you view all the threads in the Space, and you see this: "Share thread with Space"
2- Or you open a specific thread in the Space, and you see this: "Save as file to Space"
what is the best ai upscaler for artwork generated by chatgpt
something that preserves the finer details like brushstrokes and grain
Yeah a pdf with 1 half or quarter filled page
@left stream @tame current
daym
its unavailable for now
lets see when the mail will come
Cute vibe certainly
is sonar api available free of cost for the hackathon?👴
You should have received credits if you registered for it
oh okay
Impressive I wonder how much will it cost
Then there's jules
Perplexity should atm be in the shallow department lol
@distant aurora market research preset prompt uses 19 sources hah
Wonder when it'll go below 10
Wait. Research has profiles/presets in perplexity? Where?
No it's just when you click on the empty search bar then you see a bunch of prompts
Oh
But they're all the same if you don't type in anything
I see.
The longest time it took for perplexity research to provide an answer for a single query you made. Go
Average conversation with grok.
Funnily enough, I got access to someone's grok api key because they left it exposed as the app would make requests straight to the API.
Of course I reported it to them but their app has a whole bunch of flaws waiting to be used.
Grok app?
Nah, they used grok for synthesizing a personalized email for sending out
So are you saying that the website is secure?
And how exactly did you get api key from that?
In requests, they contacted the xAI completions endpoint and left their Authorization details when they could've had it contact their own endpoint which could contact the grok endpoint.
Probably is, just has a huge slew of vulnerabilities concerning external providers
Ah. So you are saying third party grok ain't secure.
Nothing of the sort, the app creators just love exposing their private keys and losing money from that
So the flaw was in their app.
Yes
GitHub and similar are full of leaked API keys but I would personally never use one. If they leak a key, they likely also haven't configured spending limits. Then others spend up this poor person's money before they realize what's going on.
If a company leaks the key, I'd feel differently.
Lmao what it defends cuckolding?
You wanna read the chat?
Nah I need my brain cells alive
I think that @distant aurora can cause issues directly over discord 😉
Here in Atlanta this is the third day where pplx isn't working. What's the deal
I'm gonna bomb that guy who put call buttons everywhere.
Like traps/landmines.
anyone else have issues with generating images? it seems very inconsistent. do you need pro? is there a limit? there's no direct generate image button and it seems to be based on your prompt. any help appreciated
Image generation counts as an enhanced query, so without pro you'll get 3 of them a day
Asking it to generate something should work just fine
ah i see, thank you
Copilot just got the image generation of gpt 4o
And there doesn't really seem to be a limit
Censors ruin it.
Yeah it's definitely heavier than chatgpt
Like if I ask it to edit an image of my hand it trips lol
lovely
It won't make copyrighted characters 😭
Where is the pro chat or pro discord for users is it this one??
@serene sand
How does Perplexity know whether a Discord user has a Pro subscription or not?
not complaining, just curious
Pro members get a different invite link afaik.
oh that's the trick
It's been disabled #📰│server-news message
A man was seriously practicing his presentation, but Sage the cat had other plans—loud meow meow commentary all the way! When the human finally gets mad, Sage flips the switch and acts like the cutest, most innocent creature ever. Classic cat move!
#catnamedsage #funnycatmoment #catinterruptions #meowtime #catantics #presentationfail #talkat...

It's time to I/O! Tune in to learn the latest news, announcements, and AI updates from Google.
To watch this keynote interpreted in American Sign Language (ASL), please click here:
https://youtube.com/live/0BLsd7HfU9w
For all things I/O, visit https://goo.gle/4jlp80v
#GoogleIO
Subscribe to our Channel: https://www.youtube.com/google
Find u...

idk what that is