#š¬āgeneral
1 messages Ā· Page 175 of 1
Sue them
Malfunctioned.
And i am sure this was a laser stuff
Have you heard of the brand "noise" ?
I know some watches have a material that some peoples skins donāt like, can cause irritation etc. the ālasersā shouldnāt be the cause since theyāre just light basically
I just have the detection turned off unless i exercise
I have seen another person using a xiaomi watch different model and same issue
Or maybe the whole watch was overheating.
I think so too
This is also likely but they'd know right
wearing watches for a long time too without taking them off and for example moisturizing can cause problems too
Did you take it off to let your skin breathe every night?
no the watch is not overheating, i think is the laser color he emits green light
and red light
can we change the color emition for blue ?
If anything the skin under where i wear my watch is less tanned
Thatās not how it works
Nope
@short narwhal thats why im asking š
Even if we were to be scientific, blue would burn more.
Lol.
You might just need to ditch the watch
The lights arenāt causing it
@short narwhal so what is your bet for this
I'd personally say go visit a dermatologist
disapeared
We are in 2025.
We should have a floating assistant by now.
Instead of watches.
Google gLaSs
Iāll do the test which image do you want to test it with
I'll dm you.
Also, thereās a new model
Bet
but yeah if someone knows what could cause that on the skin i appreciate. And maybe brands have problems with this
links of apps made by labs are permanent ?
There should be a suggestions box if there isn't š
#šāfeedback-general #1111765634267742408 @misty sandal
Should be
We already have rayban meta glasses
Oh thanks
Done.
They're audio based, no visuals atleast currently out for consumers
discord is a great mess @misty sandal and i honestly i dont like the tabs but im going to try do my best sharing at the correct places
True true
i think that's US only
Yeah, I found the feedback and posted my thoughts on why Claude Opus 4 should be added to the list of current A.I models Perplexity has. I really hope that they consider my letter.
Hi! Is this unlimited use Perplexity Labs for Pro user? I feel that kinda too good to be true š¤£
Isnt it 50 a month?
Unlimited use, you say? I use Complexity (Perplexity's browser based extension) and I've found that it only says 50 uses.

Who said unlimited? š
This dude.
Nah bro. It's 50 per month.
Cause I use both original desktop apps and web-based and I dont't see the limit?
I just guess
Even pro searches are 300 per day including deepsearch.
Oh okay thanks for ur info
I wonder if enterprise subscribers get higher limits on labs
I havent used that massive amout of work intraday so i'm not aware of the limit
its more then enough . I hope the future energetic caos with AI gonna down but i dont think so
yeah true
what is the partnership with Lewis Hamilton about? Seems to me either something with Perplexity finally getting faster or maybe his voice?
Why not? He dmāed me and gave me a code after a short convo about my situation. So I think it was a very good idea.
Just about the potential risk
After the blackout at Portugal and spain, by today we did not know what happened. Only that we trade energy with the EU market.... Strange days
Not every admin is going to do that for you, you know. Everyone has different personalities and may not take kindly to that
Yeah we need more varied renewable energy sources
And like renewable storage
Similar happened in Chile
Portugal have a big renewable energy sources very big for the scale of the country
At least it's good problems to have, hopefully they don't repeat lol
Sometimes we have wind turbines stopped xd imagine watching 25 wind turbines but only 3 works
it will repeat, it wasn't a problem of being auto sustainable at energy and that's kind of scary at europe
Mods are relatively chill here with respect to typical discord mod behaviour.
However, that doesn't guarantee that things will always stay that way.
Nuclear
Need to stop the fear mongering around nuclear energy.
Also, Chernobyl was a man made error.
That's the main one for sure
And a rather deliberate one.
We have crazy people arround the world provoking wars thats why the world can't moves to nuclear sources. Specially when nuclear is near of a lot of wars
At europe we have nuclear at spain , france and more places....
i hope not š
In fact nuclear energy is not required highly enriched Uranium then we can somewhat avoid nuclear tension.
I was surprised when i know it's only require 3-5% enrichment .
for the start of nuclear fusion but after it starts it can't be stoped
We need AGI tbf
We need nuclear for AGIšµāš«
Anyway, although there was a dark time, nuclear technology do contribute to finding new sustainable energy source.
It's a vicious cycle fr fr
its like with AI now. Though we may not aware of harms but we do see the potential
It's my opinion and it values what it values. We need to build peace first. Im not against or in favor of anything we are all learning to build something different and better. But we do not should have fixed decisions or minds.
Yes we all want peace but there's some ppl in the world don't so we have regional wars
Tbf the world has had nuclear for 70+ years and a bunch of countries, including some that hate each other, already have nuclear weapons
Trump's Golden Dome will be the game changer
I like discuss this theme's not to say everything is cathasthrophy. And history is not linear....
If history of the past was good probably we continue only with spades and Portugal have detained the entire world
and that was not good
These are the best topics imo
it was a time of discoverys
Yeah just like Economic cycle, stock prices
And it's ended, now maybe could be the time of improve all of what we researched and dicovered for years before going in a rush
Now? this was old š
All of A.I was very old, now its available for users. Ultimate discovery ( I thiked you call the AI as ultimate) š
xd
All we need now for AI to take the next step is sustainable energy.
I know he's not Portuguese but if Colombus had asked Perplexity how long it took to get to India things would be way different
not a dev, but I now can build a application and i love that š¤£
Real, like I need it to do this
No, not that way...ok perfect š
What suggestions for improvements do you have š
if you have a big trouble in your apps like cybersecurity how you gonna fix it ? This tools are great , but theory and the basics should not be replaced by this
Implement please
Can't wait until it automatically create motion graphic or 3D design with a prompt
you can do it now with everything you have xd
Tbf mine are personal apps, not to sell to market, those would get some better backend work in
It would be max a few prompts tho, 2025 is wild
what im trying to say is, do not replace knowledge by tools
tools are great and accelerate the way we work if we build a plan
True
Also use your labs tokens by tomorrow as they will expire end of month
I think they introduced it this late in the month so users can splurge
i lived in a age where technology was more limited then now, but now its available for more people. The problem is knowledge of years and years will be not replaced by tools, if we use this in correct way balancing the theory with pratical stuff we gonna get different results in less time. Does not mean we should stop learning
Lol in fact I showed off to my friends about my idea's prototype built by perplexity labs š¤£
Donāt get the hype around labs. Canāt even do iterative improvements. Basically restricted to one-shotting. I can do all that in gemini canvas.
Good stuff
What'd you build
kinda financial analysis tool
Yes it has issue with follow ups
still thinking which AI models to plug in
Main attraction is ability tobuild something quickly with minimal instructions so far and hosting it
since they release new update everyday
They really need to fix the issue with follow ups
As labs seems to not integrate them well with the existing provided solution/app
It's relevant to input token limit i think
hmm I lost labs in perplexty, anyone?
Yeah it seems to be some other issue
Input token limit is 32k that's set
kinda modest since we have 1m token limit in API
1m hallucination?
Im using desktop app/ web version work just fine
hmm weird, i lost it, also on mobile
For me it wasn't showing until I disabled complexity.
Try force stopping your app.
i using web
Then refresh.
On mobile I had to enable a vpn for it to show up, now itās there always even without the vpn
done it many times, still use it 2 hours ago
I remember reading somewhere, 1m input & ~65k output
That's for 2.5pro api iirc
How many lab prompts did you run?
This is old ui.
Still showing for me.
Try refreshing after clearing cache
I respect you refusing the pro role even tho you've a subscription
Oh you're right. Then I assume best to use 2.5 pro to build from scratch
Likely
Yeah if you've got the subscription
lol, I have disable my Adblock list Hazegi ultimate and it shows up again
Canvas is pretty neat w gemini
Hope they are able to launch comet soon and get out of this slop show
Mods
No one's falling for that.
I would love to have permission for deleting server messages.
Mod check
Iām not seeing the labs choice in selection
I'mma ask grok.
Yeah. Apparently multiple people are facing this issue. Some guy complained about this earlier.
Yeah not cool being a āproā sub without adding functionality
Figured it out needed to update app
does anzone got that json file to read the rate limits for perplexitz pro account
w
Thanks
so they use an interval counter for that as well
I OWN NOTHING! ALL RIGHTS BELONG TO WARNER BROTHERS, DC VERTIGO, AND NETFLIX
With a cameo from Michael. #lucifer #lucifermorningstar #tomellis #chloe #chloedecker #deckerstar #laurengerman
my theory is that Perplexity doesn't use hard resets for rate limits
This opens my current limts?
Could be dynamic.
instead it slowly replenishes over time
that was the case for regular searches, you could use a few and go down to 597 or so, and some time later it's back to 600
That would probably make it 80 per month.
You mean regular pro searches.
It's reduced to 300+ btw.
In their claim atleast.
actually I would assume they calculate the exact interval for 50 per month
yes, you can do some searches and watch the 600 go down, and then go back up after a while
use the link, it's still 600
Thanks that's helpful
Think you found the weirdest AI response?
Prove itāand score a hoodie!
Full announcement & prizes:
#š°āserver-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
I bought Perplexity Pro, I wonder if I can use API?
Yep
"pplx_beta_limit":50
Must be for labs
they gave you 5$ usage of API, further than that you have to pay for it
it is
Gonna burn it all today and wait like a villian for it to regenerate on the 1st june
how to create api key
Settings > under workspace> api
I wouldn't be so sure, considering mine went up by 1
doesn't make sense to both replenish and hard reset
plot twist its not the christian calendar
How to get this btw?
Hijri calender users rise up
i think that would be good because it has less days
Looking at the moon, waiting for quota to replenish
it's just a link, use a browser
I think he means how to generate it
idk, probably not like this
No, I mean, how do I get a json file that reveals rate limits.
Yeah.
that's what the link does?
So you are saying it revealed my limits for me?
I didn't find it, been circulating for a while on discord
yes
My beta is still at 49.
Nice
But my gpt 4 says 600 even though I used a few.
Maybe the limit json is not working
I told you, it replenishes quickly
Huh.
Makes no sense
Perplexity being generous.
Because it's supposed to be daily replenish
Surprisingly
Maybe potato is right.
what is create limit
It replenishes slowly instead of at the beginning or end of a time period.
Potat is right likely
Pages I think.
images actually
hmmm thats good and bad
Doesnāt pages have their own limit
if i had listened to hero and used all my quotes i would have been cooked
I was wrong.
so it replanish over a month not at the end of the month
Imma just burn through it anyway
And contact sam afterwards
Saying they scammed me
quick delete that message
Noo donāt do that to sam
Someone has to stand on ground
Meta ray bands are 350 oh god
The ones that came out way back?
Yeah
I don't see their point atleast for now
It's just like a gopro glass
@cobalt vale my gpt4 limit reset already
I used it refreshed json, it goes down
I just refreshed again
Back to 600
Maybe pplx isn't as bad as they say it is
Depends on use case.
So howās everyone using projects now?
Claude is pretty good on perplexity. O4 mini misses sometimes.
Standard search is choosing sonar right?
Nah. I'm talking about search using reasoning models.
Free search is okayish but I choose reasoning for multistep.
Since i started using claude reasoning it feels stupid to go to anything else now
That way it can do multiple searches to question your answer.
It literally can't make a website that can solve spelling and grammatical errors from input
It's produced good and apt results
Thatās a good feedback point if that is what youāre experiencing
Did you manage to report it
Gemini works great on perplexity too.
Yeah i just use gemini natively on their app since i have a sub
I am still telling it do that fix that
Way more context window
Fair. That's the main advantage of Gemini.
how come the perplexity bot is not responding when i add it to my server
discord bot
That bot is not in service for now
ahhh ok
#š¬āgeneral message Might be my fav use case
#š°āserver-news message the bot has been disabled
Definitely cost cutting
ohhhh
the whatsapp bot is definetly costing more
Did you buy a domain for this or something?
It Self hosts
Oh.
Didn't know that's a thing
Hi guys. I have an issue where some of my questions I ask are not being searched online. Anyone know of a fix for this ?
Like it says it has no internet access and can't search online
You might have disabled it accidentally.
Got a cool feature where it sends you news related to your country daily too
I checked this first. It's definitely enabled. Within a chat, 1 or 2 questions will be searched for online and thereafter none of them will be searched for online
That's definitely useful, but no way to sign in to use pro features through whatsapp i guess
Ask it verbally to search the internet.
Do Standard free searches use web search?
If not maybe that's the issue for you
I am not on a free plan
Ok then definitely might be a bug
Let me test this
That's a pplx user alright
@tough forge https://discord.com/channels/1047197230748151888/1376715465056976896 This seems like the same issue you're facing. Might be a good idea to chime in there and let them know you're experiencing it too (+ sharing a link to the thread where it happened helps)
h
this new "labs create projects from scratch" functionality uses google ai studio?
Two different products
no
hello
I saw that Pplx labs uses gpt api
Where is the evidence?
Umm, i saw it when it was thinking
@all please send best thing you created using perplexity labs
Trying to ping all? Straight to jail
You ain't slick bro
How to use perplexity to find me a girlfriend?
Skill issues, ai can't fix that
Don't be so certain before asking Perplexity Labs to create a guide
Maybe labs can make him a robot girlfriend
Only issue is she'd be limited by tokens
short term relationship forcefully
Nice fling
You ask perplexity how to be a better man.
Think you found the weirdest AI response?
Prove itāand score a hoodie!
Full announcement & prizes:
#š°āserver-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
Then you ask perplexity how to dodge awful girls.
Rest is waiting.
Culture makes a difference of day and night.
Western dating tactics would lead to women accusing you of sa in India.
Truee
is there any way to have perplexity give timestamps for a youtube video?
Go to some transcript generator.
Paste transcript into perplexity.
Or simply download and paste subtitles.
good looks bro, i thought there was a quicker way like in gemini where u just copy and paste the link
and also does anyone know what model is used on pplx whatsapp?
created a react website and a ppt for that website it was good describing the way you want your website it will create accordingly the most interesting case is it will send you the all the jsx files and the styling files in a zip folder with the instructions how to install the vita react js from scratch and implementing or deploying in the server
What exactly
Can it do
pretty good
i asked it to make an infografic based on specific heat capacity and its pretty accurate
Wild
Didn't expect hydrogen to be more than water.
i see this being used for articles and research
Idk the dif between labs and research
water has one of the highest heat capacities among non metalloifd
Nice.
@distant aurora Is Labs limited to 50 uses per month??
is it unlimited for pro users?
Should be fr
- Personal Search & Memory
I don't have it š
Dayummmmmmmmm
Not sure. Potato was saying it recharges slowly. But someone here got a message from support saying 50 per month.
I came here for this. What is that
Gay
Perplexity seems to only generate python code
Despite asking for typescript..
Send prompt.
"Rolling out over the next couple weeks" š hope that includes all of the world and not just the US area
like ChatGPT memory, not fully rolled out in Perplexity despite what the announcement says
Yeahhh... true!!
"What we partially shipped this week"
Yeah one of the oldest un rolled features out there. I don't have it BTW
can you make improvements to this repository https://github.com/skullcrushercmd/AnyGPT-experimental
support also said it was calendar month so allegedly it would reset on June 1st
Is this using labs?
Huh
Me ? Coding? š
Does Opus power Labs?
Labs makes use of artifacts but it's not meant for coding at all
It's for making finished products
What do you think? #š¬āgeneral message
Use gemini canvas /GitHub copilot/ cursor fpr coding
Code files be crazy
Does anyone have any tips regarding how to use ai models via api without draining my money?
unclear right now
openrouter is like the only free api
Very stable
I meant that I'm willing to recharge it.
But what parameters I should keep to minimise token usage/wastage.
I got sent this but not sure how it works yet #šāpro-lounge message
No access.
what
how do you not have the role yet
I swear I've seen you ping Kesku to give someone else pro
Why don't you have the pro role when when you have the pro
Gawk twah
yes bro i just want the memory feature
If you just pay the 5 dollars for the api, that setting shows up
Gives all the data to openAI though for training
Fair deal
But free tokens!!!!
but what regions is it available in
@scarlet oxide For a moment I thought that's esdeath in your pfp.
My friend in ukraine got it, so I doubt it's region locked
Do i have to add a card tho
Anygpt better fr
Insane
Imagine paying for ai
I need a spare card.
Couldn't be me
But after that can i remove the card
Nope
it's paying for free AI though
Insane
But I get free ai anyways
How do I code with labs?
it's paying for very smart free AI though
gives free 4.5 tokens which is nice
I'm
Talkin every available model on the market
Why can't i remove the card after paying the 5usd credits
Don't know š¤·
In case i go over the limit it bills me?
Nope, unless you specifically enable it
just ask it questions and it will code when necessary. don't use it for dedicated coding, it doesn't really help with that
You get 1m tokens of 4.5 per day as per @cobalt vale ss
or o3
Why
Free?
Yes
Once you have the option
You'll just have to share your project data
With Sammy
Do I get free O3 tokens daily too?
I think only 4.5 iirc
Or o4 mini high?
why would I share my project data with Perplexity's support guy?
Sam Altman
One of my providers
Whut
Sus
Free
here it is
š
What discord is that?
Are those api key dumps
Mine
How do I apply for this?
Yup
What are you scraping
Is that the server on your bio
I haven't tried yet, but it looks like every account who pays for at least $5 USD in credits is eligible for the offer
10 million o4 mini
5 usd every month or just once?
just once
Unlimited key since I partner wit him
Once
lol the labs rate limits resets kind of
Keys
I think the end date for the offer isn't decided yet
What's thr best frontend to use it with on android?
Where
Blud..
From anywhere It could be leaked at
Not telling u the method tho
Is that your server in your bio
The whole internet?
Yeah but that inv link expo
Sure..
Have fun š
Appreciate it.
Gay
Bro thinks he can exhaust few million in a day.
nobody said it was unlimited
Lol.
My api fr
Self promo max
His guy is a bot.
Lol
Bruh it's free not everyone scrapes everything like you
Bruh
We do we what we make do
Ai generated pfp
Think you found the weirdest AI response?
Prove itāand score a hoodie!
Full announcement & prizes:
#š°āserver-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
Tch
I give free
And I get free
So it's free
š
Get real
Give browser based front end if available.
Straight up terminal and curl kek
Want some sort of ui ykno
Gimme some examples
Bettergpt
I'm asking š
Self hosted openwebui
Just made my first lab request on your word chief
what, you don't trust Sam?
Im conflicted
A true dire dilemma
Also btw what front end do you use for consuming api on android and windows @cobalt vale
none right now
I just discovered that to use the Comet, we need macOS with the new M series chip. Now I'm sad :/
Life is not perfect :/
there is no Mac with A series chip (A series is for iPhones)
Oh comet is currently only for macs
will be supported on more platforms in the future
Wtf
don't worry, most of us mac users haven't received the email either
They don't ask device details so makes sense
I'm already know that they will add Windows like in 2 years
So I'm not even sad anymore
Arc died because of that shit thinking... at least Perplexity has other things that a browser lol
I loved the idea of Arc but the ignoring of Windows Users made it a "Hype Browser for the cool Apple People"
And look who is "reinventing themselves after a not so well result in the market"
(Know as "We failed but we still have some money")
wdym 3.7 never works on you.com, also they said they were working on bringing image support to claude 4
I never had any issue with image inputs
Keeps giving error.
Everytime.
I don't face this issue with sonnet 4.
do you guys know when the macos app gets updated? It always get the new features with some delay :(
Explain this electricity bill in simple language. What was the electricity usage for last 6 months?
Put this on you.com
Any open ai model.
Then see if the readings and the months match.
I cant read that language so i have no idea if the model is correct or not
Just see the numbers beside the bar graph.
That's normal numbers.
Even that doesn't match with the months.
Did it just rn.
your image is larger than 5mb....
...
go ahead
mine is working fine
email their support if you continue having this issue
It's working on web for me.
Not on the app though.
Readings are still inaccurate.
Somehow only Gemini answers it correctly on both you.com and perplexity.
image input works fine on the app as well
more of anthropics issue
Same with open ai models too.
you are never going to get the same experience on models served by API compared to the native app
hello
T3 chat did it correctly somehow.
With o4 mini.
nothing much...
i needed some guidance in ML actually
K
so
im learning ML
im done with the ML but now am stuck and cant decide what to do
NLP,DL ?
i tried them both but for some reason i keep getting overwhelmed
what should i do
Apparently image selector in app doesn't stop me even if image is larger than 5mb.
I uploaded it via select image.
Sub
Nice
You.com is useless without sub.
Even with sub, you can't do quality web search with it.
only OpenAI models can get this one LOL
Think you found the weirdest AI response?
Prove itāand score a hoodie!
Full announcement & prizes:
#š°āserver-news message
- Drop a screenshot + thread link in
https://discord.com/channels/1047197230748151888/1373806851358986320 - Vote for your faves with

Submit by June 1, vote by June 8.
Only advantage is access to multiple models.
This always shows up at the worst time
With no said limits.
Ikr.
Pops up in the middle of the conversation.
submit a bug report then
why on earth did they think it was a good idea to text wall every 5 minutes
Send image
20 usd
Too many services and subs to track
Use ai studio if you can't buy anything.
Our lord and savior.
What can you leverage there?
I think generally Gemini is providing much of the stuff available for free users
What is the laboratory?
Did it hallucinate?
yes
are you sure you want spoilers?
Yes ofc
I have no idea how Minecraft works. Is copper there?
is veri intellilence š„
turns out perplexity just routes images to another model if the selected model doesn't have vision
So today I discovered a weird thing.
Chatgpt app can read text that even Google lens can't.
Crazy.
Quality might be the issue
Hallucinating me thinks
Nope. The numbers, months, and order, match.
Wow
I think chatgpt zoomed and cropped manually.
That's actually well trained model then
try gemini
Clear
Disappointment.
- 2.5 Pro for free with seemingly no hard limit
- Full 1m context window with every model, for free which is unheard of
- Control over model parameters such as temperature, and even customizable system instructions
- Native audio AND video input, no other decent model is capable of this at all
- Native realtime voice chats and video sharing with 2.5 Flash, first speech to speech model I've seen that supports thinking (also supports search grounding at the same time)
- Limited Veo 2 access (free!!!)
- Agentic app builder (vibe coding), (probably not quite as good as v0 or bolt but completely free)
Hot Damn
Didn't know about the veo 2 part.
With gemini advanced sub it's only gonna get better
@cobalt vale o4 mini couldn't https://www.perplexity.ai/search/explain-this-electricity-bill-JloAGwZ9QPu0k8agLT5OTw
Thanks potat!
@cobalt vale Gemini did it. https://www.perplexity.ai/search/explain-this-electricity-bill-ee0RxsJeSuSZ856StXGoaA#0
Yeah if it thinks that something is too blurry for it it'll try to zoom in and check again 
on the contrary, this kills Gemini sub for me. most of the killer features are free on AI Studio, and the Gemini app's feature set is quite disappointing.
Fair play
Veo3 makes me dope videos so that justfies me for now getting that student offer thing
also, half of those free AI Studio features are not even available on the gemini app
native audio and video file uploads are a big one
Gonna have to sit down on this one later and explore the limits
Gemini works better on ai studio than the website š
Meanwhile I've just been refreshing my pplx limits API constantly to see it my labs limit gets back to 50

Sweet
Gotta mention the downsides:
- UI is very functional but not visually polished
- Google will train on all your messages
Sometimes veo3 can make pure ass too
veo 3 is 50/50 for me
They steal your data regardless
So win win anyway
when it makes good stuff it's leagues above the rest, but its also 50% horrible outputs
Yeah it's either perfect of just hallucinating artifacts
YES
The UI is so bad, I don't even care how good the other stuff they offer is it finally killed my gemini sub
Nah those "strange and unknown" lands beyond the US border are last priority /s
@cobalt vale chatgpt killing it.
I meant AI Studio but yes, the Gemini app is surprisingly unusable
Lmao the window
oh, well yeah. both
It can do basic stuff well
some how google cant figure out how to make a nice UI
But if you wanna do something more you'll just have to start using the site
Atleast it's not buggy like pplx
Pplx just straight gives up half the time
1m context
I'm happy with AI Studio
Until i refresh
65k response.
Same. š»
Oil = freedom
it's so functional that the clutter doesn't matter
They added pre market data...
What's that?
Like work on deep research, context size and fully rolling out memory
Vc money>
Ppxl doesn't even provide candle sticks
Who can take that seriously for trading

Guess that's for another update instead of working on AI
We gonna use labs for that, lol.
Ah yes they're going to use image gen to make those charts lol
Definitely precise info
Perplexity is good.
Next up labs will probably reinvent SMA indicator everytime you prompt it
no copper at all, the allay is holding a tiny gold ore block but I would let it slide if not mentioned since that's not exactly "clearly visible"
Yeah with core search
It's their main thing
All this backlash is cuz of marketing not meeting experience.
Yep
We wouldn't be ranting if deep research wasn't so heavily promoted.
Allot of fancy advertising
Not much in reality
probably did the fancy zoom thing
Or if we got teased lol
We never would've known it existed
Well people would still be disappointed
try 4.1 maybe
You probably didn't see the message yet where it zoomed, cropped and rotated the image like 10 times.
Source count went from 100+ to 12 and back constantly
Does anyone here have an access code??
that has always been something ChatGPT has blown it's competitors out of the water with. it is really good when it comes to interacting with images
Gemini is OK but has always left me feeling like it could do better
Is this a separate app or inside Opera browser?
Any information about pricing?
A new browser it seems
they are being as mysterious as comet was
If installer is available then bingo.
@cobalt vale spam mails are compulsory now.
That's crazy
"hey Opera Neon agent, please go to my gmail and unsubscribe from Opera marketing updates"
that is WILD
is this just manus? but free??
did they remove the model selector on the chat site...
I'm guessing the the agent never had one
yes, and they lied
It made me parts for an android app.
I have no idea how effective it will be.
I'll share link.
wait maybe they didn't lie
@dry kestrel if grok 1m context works only with deepsearch, then where tf is grok 3 mini even using it?
omg minimax came through
Or is grok 3 mini used for deepsearch?
it was wrong, but at least it answered
That would explain the generous rate limits.
Is that a book?
I'm not even getting what your question means there tbh.
encyclopedia
Ah...
Honestly, there's no way to tell.
Minimax agent seems free.
No referral code or waitlist bullshit.
I don't see any rate limits either.
how is it at actually understanding that 4m context tho
scott says 10m but it be forgetting shit
@cobalt vale https://agent.minimax.io/share/274700069716075
š
Had to push it a bit to get correct answer.
Okay. Time to test smartphone battery prompt.
Weeeee
I am sooooo impressed with Perplexity Labs.
I love how good Perplexity is at helping me with NHL 25 ultimate team lol
to me this shows that manus is merely bundling as many MCP features as possible to Claude and selling it at a profit
they are so not going to survive
Has anyone tried integrating their MCP into an IDE? I am in the middle of that but am not sure what to expect there.
What's funny about Perplexity is that I thought it was borderline survival when I first tried it. Then I tried a full day of using it and comparing it to Gemini Pro, and I couldn't get away from it anymore.
It kind of does a lot better job on coding than Gemini Canvas for me too.
ai studio, qwen, minimax, deepseek. Our saviours.
Hopefully qwen will fix it's deepsearch soon.
@distant aurora scira added accounts
big imo
Finally
also Claude Sonnet 4
In case you didn't know, scira was named mini perplx earlier.
just noticed that when reading the github
What the crap? What does minimax do exactly?
I guess I could ask it...
looks like it's a meh model with some MCP agentic features, for free
scira extreme?
Anyone know why I have the ābug sniperā role?
you submitted a #1140622008086970420
?
"meh" as in "meh, it's okay" or "meh" as an acronym for something?
UI is my passion
Former
my experience was fine...
Dafuq
I haven't used it much though
My android looks lovely.
@cobalt vale Also have you seen #šāpro-lounge message yet
it's nothing special
Hasn't Perplexity transformed into that with Perplexity Labs? I had it create a beautiful first rough draft about "Agentic AI Software Engineering" last night and placed it on my website. I'll be spending the weekend with Perplexity to enhance it and create a demo too.
Lovely
I did, haven't been selected though (yet š )
Oh. That's scira. I thought that's grok.
Mine looks neat.
lol yeah
Labs is a massive improvement to the report, but it appears that the research remains identical
@vague pagoda can I have pro role?
finally
What is it about?
Sorry, new to this lingo. Whose report?
Kinda annoying to see "no access" pop up.
I guess I'm confused on the pros/cons of what you were saying
š
to the report (answer) section, but not the source gathering stage
Good builder. Not as good researcher.
Or should I say scraper.
You mean PPLX is a good builder and not as good at researching as minimax, or did you mean vice-versa?
Labs specifically is an excellent builder.
Oh, okay. Yeah, the way I used it last night was that I entered research mode, did all of my research to build up PPLX's context, and then I switched to Labs and asked it to take that research and turn it into an SPA.
Not as good as digging the internet for information compared to agentic search on minimax I think. It's my first time trying minimax so it's too soon to say something firmly.
I've found that when I'm working with LLMs on bigger things, it's better to "grow" their context much like you would a person. On my local machine's Cursor rules, I tell it to keep longer-term notes in a scratchpad file that doesn't get checked into the repo.
That way they understand things, and THEN you make your real ask.
Ah. Yes. I have read about this.
That's why people use Gemini/ai studio to support it.
To support "it"? Sorry, I'm not as smart as LLMs; you'll have to give me the context lol.
waited 2 months for you to say that
LOL
LOL
I didn't have pro before the easter egg thing.
ik lol, glad you won!
Yeah. Thanks. š»
This is great. This lines up with my thoughts. I've also started to think that Cursor is nerfed. I guess I should try AI studio and some other stuff.
Maybe I should try Windsurf
is windsurf what google named their thing?
Nah
Google has firebase studio or Jules
Yeah
Has that been announced publicly though yet??
If I use AI studio, can it run locally and use models other than Gemini too? If it's unnerfed, I'm in.
Don't think so
I just need to get access to my WSL2 terminal so I can use git and other stuffs.
Awwwww
Google app so likely limited to google suite tools
nop
Yeah, that's why I just ignored it "by default". But given that info in @distant aurora 's image, I am tempted.
Opus added yet?
nop
@left stream did you see the o3-pro "leak" yet?
What you gonna test on it first?
@cobalt vale https://agent.minimax.io/share/274700069716082
Did that smartphone battery life prompt.
huh it doesnt have 1m context
what are you talking about
They dont provide that 1m context anywhere
Not on grok.com, not via API
Except deepsearch I think.
No its capped at 128k
is that good?
now try Scira Extreme
